- developers - lists.mariadb.org

[Maria-developers] MDEV-4667 DATE('string') incompability between mysql and mariadb
by Alexander Barkov 03 Jul '13

03 Jul '13

Hi Serg, my review comments for http://bazaar.launchpad.net/~maria-captains/maria/5.3-serg/revision/3665 The patch looks ok. Just a couple of small comments: > === modified file 'include/my_time.h' > --- include/my_time.h 2011-12-11 09:34:44 +0000 > +++ include/my_time.h 2013-07-03 07:46:20 +0000 > @@ -68,14 +68,17 @@ > #endif > > /* Flags to str_to_datetime */ > -#define TIME_FUZZY_DATE 1 > + > +/* > + TIME_FUZZY_DATES is used for the result will only be used for comparison > + purposes. Conversion is as relaxed as possible. > +*/ > +#define TIME_FUZZY_DATES 1 > #define TIME_DATETIME_ONLY 2 > #define TIME_TIME_ONLY 4 > -/* Must be same as MODE_NO_ZERO_IN_DATE */ > -#define TIME_NO_ZERO_IN_DATE (65536L*2*2*2*2*2*2*2) > -/* Must be same as MODE_NO_ZERO_DATE */ > -#define TIME_NO_ZERO_DATE (TIME_NO_ZERO_IN_DATE*2) > -#define TIME_INVALID_DATES (TIME_NO_ZERO_DATE*2) > +#define TIME_NO_ZERO_IN_DATE (1UL << 23) /* == MODE_NO_ZERO_IN_DATE */ > +#define TIME_NO_ZERO_DATE (1UL << 24) /* == MODE_NO_ZERO_DATE */ > +#define TIME_INVALID_DATES (1UL << 25) /* == MODE_INVALID_DATES */ > > #define MYSQL_TIME_WARN_TRUNCATED 1 > #define MYSQL_TIME_WARN_OUT_OF_RANGE 2 > > === modified file 'libmysql/libmysql.c' > --- libmysql/libmysql.c 2013-01-10 14:40:21 +0000 > +++ libmysql/libmysql.c 2013-07-03 07:46:20 +0000 > @@ -3549,7 +3549,7 @@ > case MYSQL_TYPE_TIME: > { > MYSQL_TIME *tm= (MYSQL_TIME *)buffer; > - str_to_time(value, length, tm, TIME_FUZZY_DATE, &err); > + str_to_time(value, length, tm, 0, &err); > *param->error= test(err); > break; > } > @@ -3558,7 +3558,7 @@ > case MYSQL_TYPE_TIMESTAMP: > { > MYSQL_TIME *tm= (MYSQL_TIME *)buffer; > - (void) str_to_datetime(value, length, tm, TIME_FUZZY_DATE, &err); > + (void) str_to_datetime(value, length, tm, 0, &err); > *param->error= test(err) && (param->buffer_type == MYSQL_TYPE_DATE && > tm->time_type != MYSQL_TIMESTAMP_DATE); > break; > @@ -3681,9 +3681,7 @@ > case MYSQL_TYPE_DATETIME: > { > int error; > - value= number_to_datetime(value, 0, > - (MYSQL_TIME *) buffer, TIME_FUZZY_DATE, > - &error); > + value= number_to_datetime(value, 0, (MYSQL_TIME *) buffer, 0, &error); > *param->error= test(error); > break; > } > > === modified file 'mysql-test/r/adddate_454.result' > --- mysql-test/r/adddate_454.result 2012-08-29 15:55:59 +0000 > +++ mysql-test/r/adddate_454.result 2013-07-03 07:46:20 +0000 > @@ -4,6 +4,8 @@ > d > 2012-00-00 > update t1 set d = adddate(d, interval 1 day); > +Warnings: > +Warning 1292 Incorrect datetime value: '2012-00-00' > select * from t1; > d > NULL > > === modified file 'mysql-test/r/cast.result' > --- mysql-test/r/cast.result 2013-03-17 06:41:22 +0000 > +++ mysql-test/r/cast.result 2013-07-03 07:46:20 +0000 > @@ -268,37 +268,37 @@ > 0001-02-03 10:11:12 > select cast(120010203101112.121314 as datetime); > cast(120010203101112.121314 as datetime) > -0000-00-00 00:00:00 > +NULL > Warnings: > Warning 1292 Incorrect datetime value: '120010203101112.121314' > select cast(cast(1.1 as decimal) as datetime); > cast(cast(1.1 as decimal) as datetime) > -0000-00-00 00:00:00 > +NULL > Warnings: > Warning 1292 Incorrect datetime value: '1' > select cast(cast(-1.1 as decimal) as datetime); > cast(cast(-1.1 as decimal) as datetime) > -0000-00-00 00:00:00 > +NULL > Warnings: > Warning 1292 Incorrect datetime value: '-1' > select cast('0' as date); > cast('0' as date) > -0000-00-00 > +NULL > Warnings: > Warning 1292 Incorrect datetime value: '0' > select cast('' as date); > cast('' as date) > -0000-00-00 > +NULL > Warnings: > Warning 1292 Incorrect datetime value: '' > select cast('0' as datetime); > cast('0' as datetime) > -0000-00-00 00:00:00 > +NULL > Warnings: > Warning 1292 Incorrect datetime value: '0' > select cast('' as datetime); > cast('' as datetime) > -0000-00-00 00:00:00 > +NULL > Warnings: > Warning 1292 Incorrect datetime value: '' > select cast('0' as time); > @@ -306,7 +306,7 @@ > 00:00:00 > select cast('' as time); > cast('' as time) > -00:00:00 > +NULL > Warnings: > Warning 1292 Truncated incorrect time value: '' > select cast(NULL as DATE); > @@ -323,13 +323,13 @@ > NULL > select cast(cast(120010203101112.121314 as double) as datetime); > cast(cast(120010203101112.121314 as double) as datetime) > -0000-00-00 00:00:00 > +NULL > select cast(cast(1.1 as double) as datetime); > cast(cast(1.1 as double) as datetime) > 0000-00-00 00:00:01 > select cast(cast(-1.1 as double) as datetime); > cast(cast(-1.1 as double) as datetime) > -0000-00-00 00:00:00 > +NULL > explain extended select cast(10 as double(5,2)); > id select_type table type possible_keys key key_len ref rows filtered Extra > 1 SIMPLE NULL NULL NULL NULL NULL NULL NULL NULL No tables used > @@ -764,4 +764,10 @@ > CAST(TIME('10:20:30') AS DATE) + INTERVAL 1 DAY > NULL > Warnings: > -Warning 1292 Truncated incorrect date value: '0000-00-00' > +Warning 1292 Incorrect datetime value: '0000-00-00' > +SET SQL_MODE=ALLOW_INVALID_DATES; > +SELECT DATE("foo"); > +DATE("foo") > +NULL > +Warnings: > +Warning 1292 Incorrect datetime value: 'foo' > > === modified file 'mysql-test/r/date_formats.result' > --- mysql-test/r/date_formats.result 2013-03-17 06:41:22 +0000 > +++ mysql-test/r/date_formats.result 2013-07-03 07:46:20 +0000 > @@ -586,7 +586,7 @@ > 1 AM > SELECT DATE_FORMAT('%Y-%m-%d %H:%i:%s', 1151414896); > DATE_FORMAT('%Y-%m-%d %H:%i:%s', 1151414896) > -1151414896 > +NULL > Warnings: > Warning 1292 Incorrect datetime value: '%Y-%m-%d %H:%i:%s' > select str_to_date('04 /30/2004', '%m /%d/%Y'); > > === modified file 'mysql-test/r/func_sapdb.result' > --- mysql-test/r/func_sapdb.result 2013-03-17 06:41:22 +0000 > +++ mysql-test/r/func_sapdb.result 2013-07-03 07:46:20 +0000 > @@ -168,7 +168,7 @@ > 1997-12-31 > select date("1997-13-31 23:59:59.000001"); > date("1997-13-31 23:59:59.000001") > -0000-00-00 > +NULL > Warnings: > Warning 1292 Incorrect datetime value: '1997-13-31 23:59:59.000001' > select time("1997-12-31 23:59:59.000001"); > @@ -176,7 +176,7 @@ > 23:59:59.000001 > select time("1997-12-31 25:59:59.000001"); > time("1997-12-31 25:59:59.000001") > -00:00:00 > +NULL > Warnings: > Warning 1292 Truncated incorrect time value: '1997-12-31 25:59:59.000001' > select microsecond("1997-12-31 23:59:59.000001"); > @@ -250,8 +250,6 @@ > select microsecond(19971231235959.01) as a; > a > 10000 > -Warnings: > -Warning 1292 Truncated incorrect time value: '19971231235959.01' > select date_add("1997-12-31",INTERVAL "10.09" SECOND_MICROSECOND) as a; > a > 1997-12-31 00:00:10.090000 > > === modified file 'mysql-test/r/func_time.result' > --- mysql-test/r/func_time.result 2013-06-17 15:25:55 +0000 > +++ mysql-test/r/func_time.result 2013-07-03 07:46:20 +0000 > @@ -1239,14 +1239,13 @@ > set time_zone= @@global.time_zone; > select str_to_date('10:00 PM', '%h:%i %p') + INTERVAL 10 MINUTE; > str_to_date('10:00 PM', '%h:%i %p') + INTERVAL 10 MINUTE > -NULL > -Warnings: > -Error 1411 Incorrect datetime value: '10:00 PM' for function str_to_date > +22:10:00 > select str_to_date("1997-00-04 22:23:00","%Y-%m-%D") + interval 10 minute; > str_to_date("1997-00-04 22:23:00","%Y-%m-%D") + interval 10 minute > NULL > Warnings: > -Error 1411 Incorrect datetime value: '1997-00-04 22:23:00' for function str_to_date > +Warning 1292 Truncated incorrect date value: '1997-00-04 22:23:00' > +Warning 1292 Incorrect datetime value: '1997-00-04' > create table t1 (field DATE); > insert into t1 values ('2006-11-06'); > select * from t1 where field < '2006-11-06 04:08:36.0'; > @@ -1452,13 +1451,15 @@ > NULL > SELECT WEEK(DATE_ADD(FROM_DAYS(1),INTERVAL 1 MONTH), 1); > WEEK(DATE_ADD(FROM_DAYS(1),INTERVAL 1 MONTH), 1) > -0 > +NULL > +Warnings: > +Warning 1292 Incorrect datetime value: '0000-00-00' > # > # Bug#12584302 AFTER FIX FOR #12403504: ASSERTION FAILED: DELSUM+(INT) Y/4-TEMP > 0, > # > DO WEEK((DATE_ADD((CAST(0 AS DATE)), INTERVAL 1 YEAR_MONTH)), 5); > Warnings: > -Warning 1292 Incorrect datetime value: '0' > +Warning 1292 Incorrect datetime value: '0000-00-00' > # > # BUG#13458237 INCONSISTENT HANDLING OF INVALIDE DATES WITH ZERO DAY > # SIMILAR TO '2009-10-00' > @@ -1755,7 +1756,7 @@ > Warning 1441 Datetime function: time field overflow > select cast('131415.123e0' as time); > cast('131415.123e0' as time) > -00:00:00 > +NULL > Warnings: > Warning 1292 Truncated incorrect time value: '131415.123e0' > select cast('2010-01-02 03:04:05' as datetime) between null and '2010-01-02 03:04:04'; > @@ -1775,12 +1776,12 @@ > NULL > select truncate(date('2010-40-10'), 6); > truncate(date('2010-40-10'), 6) > -0.000000 > +NULL > Warnings: > Warning 1292 Incorrect datetime value: '2010-40-10' > select extract(month from '2010-40-50'); > extract(month from '2010-40-50') > -0 > +NULL > Warnings: > Warning 1292 Incorrect datetime value: '2010-40-50' > select subtime('0000-00-10 10:10:10', '30 10:00:00'); > @@ -1858,6 +1859,8 @@ > select timestampadd(week, 1, f1) from t1; > timestampadd(week, 1, f1) > NULL > +Warnings: > +Warning 1292 Incorrect datetime value: '0000-00-00' > select timestampadd(week, 1, date("0000-00-00")); > timestampadd(week, 1, date("0000-00-00")) > NULL > > === modified file 'mysql-test/r/parser.result' > --- mysql-test/r/parser.result 2011-03-01 12:24:36 +0000 > +++ mysql-test/r/parser.result 2013-07-03 07:46:20 +0000 > @@ -555,14 +555,10 @@ > DROP TABLE IF EXISTS t1; > SELECT STR_TO_DATE('10:00 PM', '%h:%i %p') + INTERVAL 10 MINUTE; > STR_TO_DATE('10:00 PM', '%h:%i %p') + INTERVAL 10 MINUTE > -NULL > -Warnings: > -Error 1411 Incorrect datetime value: '10:00 PM' for function str_to_date > +22:10:00 > SELECT STR_TO_DATE('10:00 PM', '%h:%i %p') + INTERVAL (INTERVAL(1,2,3) + 1) MINUTE; > STR_TO_DATE('10:00 PM', '%h:%i %p') + INTERVAL (INTERVAL(1,2,3) + 1) MINUTE > -NULL > -Warnings: > -Error 1411 Incorrect datetime value: '10:00 PM' for function str_to_date > +22:01:00 > SELECT "1997-12-31 23:59:59" + INTERVAL 1 SECOND; > "1997-12-31 23:59:59" + INTERVAL 1 SECOND > 1998-01-01 00:00:00 > > === modified file 'mysql-test/r/partition_pruning.result' > --- mysql-test/r/partition_pruning.result 2013-03-17 06:41:22 +0000 > +++ mysql-test/r/partition_pruning.result 2013-07-03 07:46:20 +0000 > @@ -1906,10 +1906,9 @@ > # test with an invalid date, which lead to item->null_value is set. > EXPLAIN PARTITIONS SELECT * FROM t1 WHERE b < CAST('2009-04-99' AS DATETIME); > id select_type table partitions type possible_keys key key_len ref rows Extra > -1 SIMPLE t1 p20090401 ALL NULL NULL NULL NULL 2 Using where > +1 SIMPLE NULL NULL NULL NULL NULL NULL NULL NULL Impossible WHERE noticed after reading const tables > Warnings: > Warning 1292 Incorrect datetime value: '2009-04-99' > -Warning 1292 Incorrect datetime value: '2009-04-99' > DROP TABLE t1; > CREATE TABLE t1 > (a INT NOT NULL AUTO_INCREMENT, > > === modified file 'mysql-test/r/type_date.result' > --- mysql-test/r/type_date.result 2013-06-28 08:00:25 +0000 > +++ mysql-test/r/type_date.result 2013-07-03 07:46:20 +0000 > @@ -136,7 +136,7 @@ > 1311 > select year(@d), month(@d), day(@d), cast(@d as date); > year(@d) month(@d) day(@d) cast(@d as date) > -0 0 0 0000-00-00 > +NULL NULL NULL NULL > Warnings: > Warning 1292 Incorrect datetime value: '1311' > Warning 1292 Incorrect datetime value: '1311' > > === modified file 'mysql-test/r/type_datetime.result' > --- mysql-test/r/type_datetime.result 2013-06-28 08:00:25 +0000 > +++ mysql-test/r/type_datetime.result 2013-07-03 07:46:20 +0000 > @@ -657,8 +657,8 @@ > insert into t1 values ('2000-12-03','22:55:23'),('2008-05-03','10:19:31'); > select case when d = '2012-12-12' then d else t end as cond, group_concat( d ) from t1 group by cond; > cond group_concat( d ) > -0000-00-00 00:00:00 2000-12-03 > -0000-00-00 00:00:00 2008-05-03 > +NULL 2000-12-03 > +NULL 2008-05-03 > Warnings: > Warning 1292 Incorrect datetime value: '22:55:23' > Warning 1292 Incorrect datetime value: '10:19:31' > > === modified file 'mysql-test/suite/vcol/r/vcol_misc.result' > --- mysql-test/suite/vcol/r/vcol_misc.result 2013-01-22 05:29:19 +0000 > +++ mysql-test/suite/vcol/r/vcol_misc.result 2013-07-03 07:46:20 +0000 > @@ -187,7 +187,11 @@ > tsv TIMESTAMP AS (ADDDATE(ts, INTERVAL 1 DAY)) VIRTUAL > ) ENGINE=MyISAM; > INSERT INTO t1 (tsv) VALUES (DEFAULT); > +Warnings: > +Warning 1292 Incorrect datetime value: '0000-00-00' > INSERT DELAYED INTO t1 (tsv) VALUES (DEFAULT); > +Warnings: > +Warning 1292 Incorrect datetime value: '0000-00-00' > FLUSH TABLES; > SELECT COUNT(*) FROM t1; > COUNT(*) > > === modified file 'mysql-test/t/cast.test' > --- mysql-test/t/cast.test 2012-05-20 12:57:29 +0000 > +++ mysql-test/t/cast.test 2013-07-03 07:46:20 +0000 > @@ -439,3 +439,6 @@ > # > SELECT CAST(TIME('10:20:30') AS DATE) + INTERVAL 1 DAY; > > +SET SQL_MODE=ALLOW_INVALID_DATES; > +SELECT DATE("foo"); > + > > === modified file 'sql-common/my_time.c' > --- sql-common/my_time.c 2012-06-08 17:15:01 +0000 > +++ sql-common/my_time.c 2013-07-03 07:46:20 +0000 > @@ -83,7 +83,7 @@ > { > if (not_zero_date) > { > - if ((((flags & TIME_NO_ZERO_IN_DATE) || !(flags & TIME_FUZZY_DATE)) && > + if (((flags & TIME_NO_ZERO_IN_DATE) && > (ltime->month == 0 || ltime->day == 0)) || ltime->neg || > (!(flags & TIME_INVALID_DATES) && > ltime->month && ltime->day > days_in_month[ltime->month-1] && > @@ -115,7 +115,7 @@ > length Length of string > l_time Date is stored here > flags Bitmap of following items > - TIME_FUZZY_DATE Set if we should allow partial dates > + TIME_FUZZY_DATE > TIME_DATETIME_ONLY Set if we only allow full datetimes. > TIME_NO_ZERO_IN_DATE Don't allow partial dates > TIME_NO_ZERO_DATE Don't allow 0000-00-00 date > @@ -1324,7 +1324,7 @@ > if (nr > 9999999 && neg == 0) > { > if (number_to_datetime(nr, sec_part, ltime, > - TIME_FUZZY_DATE | TIME_INVALID_DATES, was_cut) < 0) > + TIME_INVALID_DATES, was_cut) < 0) > return -1; > > ltime->year= ltime->month= ltime->day= 0; > > === modified file 'sql/field.cc' > --- sql/field.cc 2013-02-28 20:48:47 +0000 > +++ sql/field.cc 2013-07-03 07:46:20 +0000 > @@ -5127,10 +5127,9 @@ > Lazy_string_str str(from, len); > > func_res= str_to_datetime(from, len, &ltime, > - (TIME_FUZZY_DATE | > - (thd->variables.sql_mode & > + (thd->variables.sql_mode & > (MODE_NO_ZERO_IN_DATE | MODE_NO_ZERO_DATE | > - MODE_INVALID_DATES))), > + MODE_INVALID_DATES)), Why not to add a new method in THD to return these sql_mode datetime flags? These flags repeat many times. Something similar to sql_mode_for_dates(), but always returning the "strict_date_checking" version that ORs all three flags: return thd->variables.sql_mode & (MODE_NO_ZERO_IN_DATE | MODE_NO_ZERO_DATE | MODE_INVALID_DATES); By the way, sql_mode_for_dates() is used in str_to_datetime_with_warn() and Item::send(). Do you know why? Perhaps they also should follow all these three sql_mode date flags... > &error); > return store_TIME_with_warning(&ltime, &str, error, func_res > MYSQL_TIMESTAMP_ERROR); > } > @@ -5144,11 +5143,10 @@ > Lazy_string_double str(nr); > > longlong tmp= double_to_datetime(nr, &ltime, > - (TIME_FUZZY_DATE | > - (thd->variables.sql_mode & > + (thd->variables.sql_mode & > (MODE_NO_ZERO_IN_DATE | > MODE_NO_ZERO_DATE | > - MODE_INVALID_DATES))), &error); > + MODE_INVALID_DATES)), &error); > return store_TIME_with_warning(&ltime, &str, error, tmp != -1); > } > > @@ -5161,11 +5159,10 @@ > THD *thd= table->in_use; > Lazy_string_num str(nr); > > - tmp= number_to_datetime(nr, 0, &ltime, (TIME_FUZZY_DATE | > - (thd->variables.sql_mode & > + tmp= number_to_datetime(nr, 0, &ltime, (thd->variables.sql_mode & > (MODE_NO_ZERO_IN_DATE | > MODE_NO_ZERO_DATE | > - MODE_INVALID_DATES))), &error); > + MODE_INVALID_DATES)), &error); > > return store_TIME_with_warning(&ltime, &str, error, tmp != -1); > } > @@ -5181,17 +5178,16 @@ > structure always fit into DATETIME range. > */ > have_smth_to_conv= !check_date(&l_time, pack_time(&l_time) != 0, > - (TIME_FUZZY_DATE | > - (current_thd->variables.sql_mode & > + (current_thd->variables.sql_mode & > (MODE_NO_ZERO_IN_DATE | MODE_NO_ZERO_DATE | > - MODE_INVALID_DATES))), &error); > + MODE_INVALID_DATES)), &error); > return store_TIME_with_warning(&l_time, &str, error, have_smth_to_conv); > } > > my_decimal *Field_temporal::val_decimal(my_decimal *d) > { > MYSQL_TIME ltime; > - if (get_date(&ltime, TIME_FUZZY_DATE)) > + if (get_date(&ltime, 0)) Now get_date() is called in a lot of places with 0 flags. Why not to add a method: bool get_date(MYSQL_TIME *ltime) { return get_date(ltime, 0); } And a similar wrapper for get_arg0_date(), with flags==0... ? > { > bzero(&ltime, sizeof(ltime)); > ltime.time_type= mysql_type_to_time_type(type()); > @@ -5330,7 +5326,8 @@ > bool Field_time::get_date(MYSQL_TIME *ltime, uint fuzzydate) > { > THD *thd= table->in_use; > - if (!(fuzzydate & (TIME_FUZZY_DATE|TIME_TIME_ONLY))) > + if (!(fuzzydate & TIME_TIME_ONLY) && > + (fuzzydate & TIME_NO_ZERO_IN_DATE)) > { > push_warning_printf(thd, MYSQL_ERROR::WARN_LEVEL_WARN, > ER_WARN_DATA_OUT_OF_RANGE, > @@ -5459,7 +5456,7 @@ > ltime->time_type= MYSQL_TIMESTAMP_TIME; > ltime->hour+= (ltime->month*32+ltime->day)*24; > ltime->month= ltime->day= 0; > - return fuzzydate & (TIME_FUZZY_DATE | TIME_TIME_ONLY) ? 0 : 1; > + return !(fuzzydate & TIME_TIME_ONLY) && (fuzzydate & TIME_NO_ZERO_IN_DATE); > } > > > @@ -5848,7 +5845,7 @@ > bool Field_datetime::send_binary(Protocol *protocol) > { > MYSQL_TIME tm; > - Field_datetime::get_date(&tm, TIME_FUZZY_DATE); > + Field_datetime::get_date(&tm, 0); > return protocol->store(&tm, 0); > } > > @@ -5931,7 +5928,7 @@ > if (!tmp) > return (fuzzydate & TIME_NO_ZERO_DATE) != 0; > if (!ltime->month || !ltime->day) > - return !(fuzzydate & TIME_FUZZY_DATE); > + return fuzzydate & TIME_NO_ZERO_IN_DATE; > return 0; > } > > @@ -5984,11 +5981,10 @@ > error= 2; > } > else > - tmp= number_to_datetime(nr, sec_part, &ltime, (TIME_FUZZY_DATE | > - (thd->variables.sql_mode & > + tmp= number_to_datetime(nr, sec_part, &ltime, (thd->variables.sql_mode & > (MODE_NO_ZERO_IN_DATE | > MODE_NO_ZERO_DATE | > - MODE_INVALID_DATES))), &error); > + MODE_INVALID_DATES)), &error); > > return store_TIME_with_warning(&ltime, &str, error, tmp != -1); > } > @@ -5996,7 +5992,7 @@ > bool Field_datetime_hires::send_binary(Protocol *protocol) > { > MYSQL_TIME ltime; > - Field_datetime_hires::get_date(&ltime, TIME_FUZZY_DATE); > + Field_datetime_hires::get_date(&ltime, 0); > return protocol->store(&ltime, dec); > } > > @@ -6004,14 +6000,14 @@ > double Field_datetime_hires::val_real(void) > { > MYSQL_TIME ltime; > - Field_datetime_hires::get_date(&ltime, TIME_FUZZY_DATE); > + Field_datetime_hires::get_date(&ltime, 0); > return TIME_to_double(&ltime); > } > > longlong Field_datetime_hires::val_int(void) > { > MYSQL_TIME ltime; > - Field_datetime_hires::get_date(&ltime, TIME_FUZZY_DATE); > + Field_datetime_hires::get_date(&ltime, 0); > return TIME_to_ulonglong_datetime(&ltime); > } > > @@ -6020,7 +6016,7 @@ > String *unused __attribute__((unused))) > { > MYSQL_TIME ltime; > - Field_datetime_hires::get_date(&ltime, TIME_FUZZY_DATE); > + Field_datetime_hires::get_date(&ltime, 0); > str->alloc(field_length+1); > str->length(field_length); > my_datetime_to_str(&ltime, (char*) str->ptr(), dec); > @@ -6035,7 +6031,7 @@ > if (!packed) > return fuzzydate & TIME_NO_ZERO_DATE; > if (!ltime->month || !ltime->day) > - return !(fuzzydate & TIME_FUZZY_DATE); > + return fuzzydate & TIME_NO_ZERO_IN_DATE; > return 0; > } > > > === modified file 'sql/field_conv.cc' > --- sql/field_conv.cc 2012-04-05 21:07:18 +0000 > +++ sql/field_conv.cc 2013-07-03 07:46:20 +0000 > @@ -413,7 +413,7 @@ > static void do_field_temporal(Copy_field *copy) > { > MYSQL_TIME ltime; > - copy->from_field->get_date(&ltime, TIME_FUZZY_DATE); > + copy->from_field->get_date(&ltime, 0); > copy->to_field->store_time_dec(&ltime, copy->from_field->decimals()); > } > > @@ -884,7 +884,7 @@ > if (from->cmp_type() == TIME_RESULT) > { > MYSQL_TIME ltime; > - if (from->get_date(&ltime, TIME_FUZZY_DATE)) > + if (from->get_date(&ltime, 0)) > return to->reset(); > else > return to->store_time_dec(&ltime, from->decimals()); > > === modified file 'sql/filesort.cc' > --- sql/filesort.cc 2013-03-17 06:41:22 +0000 > +++ sql/filesort.cc 2013-07-03 07:46:20 +0000 > @@ -868,7 +868,7 @@ > else > { > MYSQL_TIME buf; > - if (item->get_date_result(&buf, TIME_FUZZY_DATE | TIME_INVALID_DATES)) > + if (item->get_date_result(&buf, TIME_INVALID_DATES)) > { > DBUG_ASSERT(maybe_null); > DBUG_ASSERT(item->null_value); > > === modified file 'sql/item.cc' > --- sql/item.cc 2013-03-17 06:41:22 +0000 > +++ sql/item.cc 2013-07-03 07:46:20 +0000 > @@ -252,7 +252,7 @@ > String *Item::val_string_from_date(String *str) > { > MYSQL_TIME ltime; > - if (get_date(&ltime, TIME_FUZZY_DATE) || > + if (get_date(&ltime, 0) || > str->alloc(MAX_DATE_STRING_REP_LENGTH)) > { > null_value= 1; > @@ -308,7 +308,7 @@ > { > DBUG_ASSERT(fixed == 1); > MYSQL_TIME ltime; > - if (get_date(&ltime, TIME_FUZZY_DATE)) > + if (get_date(&ltime, 0)) > { > my_decimal_set_zero(decimal_value); > null_value= 1; // set NULL, stop processing > @@ -367,7 +367,7 @@ > int Item::save_date_in_field(Field *field) > { > MYSQL_TIME ltime; > - if (get_date(&ltime, TIME_FUZZY_DATE)) > + if (get_date(&ltime, 0)) > return set_field_to_null_with_conversions(field, 0); > field->set_notnull(); > return field->store_time_dec(&ltime, decimals); > @@ -1205,7 +1205,7 @@ > if allowed, otherwise - null. > */ > bzero((char*) ltime,sizeof(*ltime)); > - return null_value|= (fuzzydate & (TIME_NO_ZERO_DATE|TIME_NO_ZERO_IN_DATE)); > + return null_value|= !(fuzzydate & TIME_FUZZY_DATES); > } > > bool Item::get_seconds(ulonglong *sec, ulong *sec_part) > @@ -6032,7 +6032,7 @@ > case MYSQL_TYPE_TIMESTAMP: > { > MYSQL_TIME tm; > - get_date(&tm, TIME_FUZZY_DATE | sql_mode_for_dates()); > + get_date(&tm, sql_mode_for_dates()); > if (!null_value) > { > if (f_type == MYSQL_TYPE_DATE) > @@ -8161,8 +8161,8 @@ > } > else > { > - field->get_date(&field_time, TIME_FUZZY_DATE | TIME_INVALID_DATES); > - item->get_date(&item_time, TIME_FUZZY_DATE | TIME_INVALID_DATES); > + field->get_date(&field_time, TIME_INVALID_DATES); > + item->get_date(&item_time, TIME_INVALID_DATES); > } > return my_time_compare(&field_time, &item_time); > } > @@ -8343,7 +8343,7 @@ > value_cached= true; > > MYSQL_TIME ltime; > - if (example->get_date_result(&ltime, TIME_FUZZY_DATE)) > + if (example->get_date_result(&ltime, 0)) > value=0; > else > value= pack_time(&ltime); > > === modified file 'sql/item.h' > --- sql/item.h 2013-05-03 13:07:13 +0000 > +++ sql/item.h 2013-07-03 07:46:20 +0000 > @@ -936,7 +936,7 @@ > Item **ref, bool skip_registered); > virtual bool get_date(MYSQL_TIME *ltime,uint fuzzydate); > bool get_time(MYSQL_TIME *ltime) > - { return get_date(ltime, TIME_TIME_ONLY | TIME_FUZZY_DATE); } > + { return get_date(ltime, TIME_TIME_ONLY); } > bool get_seconds(ulonglong *sec, ulong *sec_part); > virtual bool get_date_result(MYSQL_TIME *ltime,uint fuzzydate) > { return get_date(ltime,fuzzydate); } > > === modified file 'sql/item_cmpfunc.cc' > --- sql/item_cmpfunc.cc 2013-02-25 03:16:11 +0000 > +++ sql/item_cmpfunc.cc 2013-07-03 07:46:20 +0000 > @@ -713,7 +713,7 @@ > bool value; > int error; > enum_mysql_timestamp_type timestamp_type; > - int flags= TIME_FUZZY_DATE | MODE_INVALID_DATES; > + int flags= TIME_FUZZY_DATES | MODE_INVALID_DATES; > > if (warn_type == MYSQL_TIMESTAMP_TIME) > flags|= TIME_TIME_ONLY; > @@ -885,7 +885,7 @@ > else > { > MYSQL_TIME ltime; > - uint fuzzydate= TIME_FUZZY_DATE | TIME_INVALID_DATES; > + uint fuzzydate= TIME_FUZZY_DATES | TIME_INVALID_DATES; > if (f_type == MYSQL_TYPE_TIME) > fuzzydate|= TIME_TIME_ONLY; > if (item->get_date(&ltime, fuzzydate)) > > === modified file 'sql/item_func.cc' > --- sql/item_func.cc 2013-06-28 08:00:25 +0000 > +++ sql/item_func.cc 2013-07-03 07:46:20 +0000 > @@ -2538,7 +2538,7 @@ > if (compare_as_dates) > { > MYSQL_TIME ltime; > - if (get_date(&ltime, TIME_FUZZY_DATE)) > + if (get_date(&ltime, 0)) > return 0; > > return TIME_to_double(&ltime); > @@ -2567,7 +2567,7 @@ > if (compare_as_dates) > { > MYSQL_TIME ltime; > - if (get_date(&ltime, TIME_FUZZY_DATE)) > + if (get_date(&ltime, 0)) > return 0; > > return TIME_to_ulonglong(&ltime); > @@ -2597,7 +2597,7 @@ > if (compare_as_dates) > { > MYSQL_TIME ltime; > - if (get_date(&ltime, TIME_FUZZY_DATE)) > + if (get_date(&ltime, 0)) > return 0; > > return date2my_decimal(&ltime, dec); > > === modified file 'sql/item_strfunc.cc' > --- sql/item_strfunc.cc 2013-03-26 18:09:47 +0000 > +++ sql/item_strfunc.cc 2013-07-03 07:46:20 +0000 > @@ -3618,10 +3618,10 @@ > } > break; > case DYN_COL_DATETIME: > - args[valpos]->get_date(&vals[i].x.time_value, TIME_FUZZY_DATE); > + args[valpos]->get_date(&vals[i].x.time_value, 0); > break; > case DYN_COL_DATE: > - args[valpos]->get_date(&vals[i].x.time_value, TIME_FUZZY_DATE); > + args[valpos]->get_date(&vals[i].x.time_value, 0); > break; > case DYN_COL_TIME: > args[valpos]->get_time(&vals[i].x.time_value); > > === modified file 'sql/item_timefunc.cc' > --- sql/item_timefunc.cc 2013-06-28 08:00:25 +0000 > +++ sql/item_timefunc.cc 2013-07-03 07:46:20 +0000 > @@ -791,7 +791,7 @@ > MYSQL_TIME ltime; > longlong res; > int dummy; /* unused */ > - if (get_arg0_date(&ltime, TIME_FUZZY_DATE)) > + if (get_arg0_date(&ltime, 0)) > { > /* got NULL, leave the incl_endp intact */ > return LONGLONG_MIN; > @@ -855,14 +855,14 @@ > { > DBUG_ASSERT(fixed == 1); > MYSQL_TIME ltime; > - return get_arg0_date(&ltime, TIME_FUZZY_DATE) ? 0 : (longlong) ltime.day; > + return get_arg0_date(&ltime, 0) ? 0 : (longlong) ltime.day; > } > > longlong Item_func_month::val_int() > { > DBUG_ASSERT(fixed == 1); > MYSQL_TIME ltime; > - return get_arg0_date(&ltime, TIME_FUZZY_DATE) ? 0 : (longlong) ltime.month; > + return get_arg0_date(&ltime, 0) ? 0 : (longlong) ltime.month; > } > > > @@ -907,7 +907,7 @@ > { > DBUG_ASSERT(fixed == 1); > MYSQL_TIME ltime; > - if (get_arg0_date(&ltime, TIME_FUZZY_DATE)) > + if (get_arg0_date(&ltime, 0)) > return 0; > return (longlong) ((ltime.month+2)/3); > } > @@ -981,7 +981,7 @@ > DBUG_ASSERT(fixed == 1); > uint year; > MYSQL_TIME ltime; > - if (get_arg0_date(&ltime, TIME_NO_ZERO_DATE)) > + if (get_arg0_date(&ltime, TIME_NO_ZERO_DATE | TIME_NO_ZERO_IN_DATE)) > return 0; > return (longlong) calc_week(&ltime, > week_mode((uint) args[1]->val_int()), > @@ -994,7 +994,7 @@ > DBUG_ASSERT(fixed == 1); > uint year,week; > MYSQL_TIME ltime; > - if (get_arg0_date(&ltime, TIME_NO_ZERO_DATE)) > + if (get_arg0_date(&ltime, TIME_NO_ZERO_DATE | TIME_NO_ZERO_IN_DATE)) > return 0; > week= calc_week(&ltime, > (week_mode((uint) args[1]->val_int()) | WEEK_YEAR), > @@ -1008,7 +1008,7 @@ > DBUG_ASSERT(fixed == 1); > MYSQL_TIME ltime; > > - if (get_arg0_date(&ltime, TIME_NO_ZERO_DATE)) > + if (get_arg0_date(&ltime, TIME_NO_ZERO_DATE | TIME_NO_ZERO_IN_DATE)) > return 0; > > return (longlong) calc_weekday(calc_daynr(ltime.year, ltime.month, > @@ -1050,7 +1050,7 @@ > { > DBUG_ASSERT(fixed == 1); > MYSQL_TIME ltime; > - return get_arg0_date(&ltime, TIME_FUZZY_DATE) ? 0 : (longlong) ltime.year; > + return get_arg0_date(&ltime, 0) ? 0 : (longlong) ltime.year; > } > > > @@ -1082,7 +1082,7 @@ > { > DBUG_ASSERT(fixed == 1); > MYSQL_TIME ltime; > - if (get_arg0_date(&ltime, TIME_FUZZY_DATE)) > + if (get_arg0_date(&ltime, 0)) > { > /* got NULL, leave the incl_endp intact */ > return LONGLONG_MIN; > @@ -1354,7 +1354,7 @@ > { > DBUG_ASSERT(fixed == 1); > MYSQL_TIME ltime; > - if (get_date(&ltime, TIME_FUZZY_DATE)) > + if (get_date(&ltime, 0)) > return 0; > longlong v= TIME_to_ulonglong(&ltime); > return ltime.neg ? -v : v; > @@ -1365,7 +1365,7 @@ > { > DBUG_ASSERT(fixed == 1); > MYSQL_TIME ltime; > - if (get_date(&ltime, TIME_FUZZY_DATE)) > + if (get_date(&ltime, 0)) > return 0; > return TIME_to_double(&ltime); > } > @@ -1744,7 +1744,7 @@ > int is_time_flag = is_time_format ? TIME_TIME_ONLY : 0; > DBUG_ASSERT(fixed == 1); > > - if (get_arg0_date(&l_time, TIME_FUZZY_DATE | is_time_flag)) > + if (get_arg0_date(&l_time, is_time_flag)) > return 0; Perhaps now it's easier to put the expression directly to get_arg0_date(), instead of having a variable. > > if (!(format = args[1]->val_str(str)) || !format->length()) > @@ -1923,10 +1923,15 @@ > { > INTERVAL interval; > > - if (args[0]->get_date(ltime, TIME_NO_ZERO_DATE | TIME_FUZZY_DATE | TIME_NO_ZERO_IN_DATE) || > + if (args[0]->get_date(ltime, 0) || > get_interval_value(args[1], int_type, &value, &interval)) > return (null_value=1); > > + if (ltime->time_type != MYSQL_TIMESTAMP_TIME && > + check_date_with_warn(ltime, TIME_NO_ZERO_DATE | TIME_NO_ZERO_IN_DATE, > + MYSQL_TIMESTAMP_ERROR)) > + return (null_value=1); > + > if (date_sub_interval) > interval.neg = !interval.neg; > > @@ -2019,7 +2024,7 @@ > long neg; > int is_time_flag = date_value ? 0 : TIME_TIME_ONLY; One more time (see the previous comment) > - if (get_arg0_date(&ltime, TIME_FUZZY_DATE | is_time_flag)) > + if (get_arg0_date(&ltime, is_time_flag)) > return 0; > neg= ltime.neg ? -1 : 1; > > @@ -2430,7 +2435,7 @@ > > if (is_date) // TIMESTAMP function > { > - if (get_arg0_date(&l_time1, TIME_FUZZY_DATE) || > + if (get_arg0_date(&l_time1, 0) || > args[1]->get_time(&l_time2) || > l_time1.time_type == MYSQL_TIMESTAMP_TIME || > l_time2.time_type != MYSQL_TIMESTAMP_TIME) > > === modified file 'sql/time.cc' > --- sql/time.cc 2013-06-28 12:25:06 +0000 > +++ sql/time.cc 2013-07-03 07:46:20 +0000 > @@ -263,7 +263,7 @@ > @param nr integer part of the number to convert > @param sec_part microsecond part of the number > @param ltime converted value will be written here > - @param fuzzydate conversion flags (TIME_FUZZY_DATE, etc) > + @param fuzzydate conversion flags (TIME_INVALID_DATE, etc) > @param str original number, as a Lazy_string. For the warning > @param field_name field name or NULL if not a field. For the warning > > @@ -280,6 +280,7 @@ > > if (fuzzydate & TIME_TIME_ONLY) > { > + fuzzydate= TIME_TIME_ONLY; // clear other flags > f_type= MYSQL_TYPE_TIME; > res= number_to_time(neg, nr, sec_part, ltime, &was_cut); > } > @@ -289,7 +290,7 @@ > res= neg ? -1 : number_to_datetime(nr, sec_part, ltime, fuzzydate, &was_cut); > } > > - if (res < 0 || (was_cut && !(fuzzydate & TIME_FUZZY_DATE))) > + if (res < 0 || (was_cut && (fuzzydate & TIME_NO_ZERO_IN_DATE))) > { > make_truncated_value_warning(current_thd, > MYSQL_ERROR::WARN_LEVEL_WARN, str,

2 1

[Maria-developers] MDEV-4472 (audit plugin)
by Alexey Botchkov 03 Jul '13

03 Jul '13

Hi, Sergei. Here is the plugin code for you to check out. It constists of a single new file, so i send you the link to it. http://myoffice.izhnet.ru/~alex/server_audit/server_audit.c What's missing in it as i'm not sure how it can be done: - No server host/ip in the log. How get that information? - SYSLOG logging. The SysLog i knew was declared in the syslog.h. That's the interface talking to the local syslogd instance. If it's what we need to use, then i have no idea how to specify the Syslog server IP for that daemon (which is required by the spec). I see nothing in the API about it and i thought it should be configured with the daemon's cfg file or something like that. Best regards. HF

2 6

[Maria-developers] On the question of whether select_id allow to reconstruct the query tree
by Sergei Petrunia 02 Jul '13

02 Jul '13

1 0

[Maria-developers] Dynamic Columns Test Error
by River wubx 02 Jul '13

02 Jul '13

MariaDB [test]> show create table assets ; +--------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+ | Table | Create Table | +--------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+ | assets | CREATE TABLE `assets` ( `item_name` varchar(32) NOT NULL, `dynamic_cols` blob, PRIMARY KEY (`item_name`) ) ENGINE=InnoDB DEFAULT CHARSET=utf8 PAGE_CHECKSUM=1 | +--------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+ 1 row in set (0.00 sec) MariaDB [test]> insert into assets values ('MariaDB T-shirt', COLUMN_CREATE('color', 'blue', 'size', 'XL')); ERROR 1921 (22007): Illegal value used as argument of dynamic column function MariaDB [test]> select version(); +--------------------+ | version() | +--------------------+ | 5.5.31-MariaDB-log | +--------------------+ 1 row in set (0.00 sec) Why? -- 姓名： river/吴炳锡手机：13501245755 http://www.mysqlsupport.cn 创业公司MySQL 支持，优化，架构设计

2 1

[Maria-developers] Doubt at query cache resize
by Roberto Spadim 29 Jun '13

29 Jun '13

Hi guys I was trying to compile in windows and warning was displayed at mariadb 10:0:3 in sysvars.cc at line 2235 Ulong new cache size = ulonglong function Could we lost the query cache size? Like a overflow

2 2

Re: [Maria-developers] Please review: EXPLAIN UPDATE patch + refactoring
by Sergei Petrunia 28 Jun '13

28 Jun '13

Now with attachment. On Fri, Jun 28, 2013 at 09:18:56PM +0400, Sergei Petrunia wrote: > Hello Igor, > > Please review the attached patch against current 10.0-base which contains > - EXPLAIN UPDATE/DELETE support > - "Query Plan Footprint" refactoring. > > The tree is at lp:~maria-captains/maria/10.0-base-explain-update-r2 > > BR Sergei -- Sergei Petrunia, Software Developer Monty Program AB, http://askmonty.org Blog: http://s.petrunia.net/blog

1 0

Re: [Maria-developers] Rev 3768: Changed acl_roles to be stored into a HASH
by Sergei 28 Jun '13

28 Jun '13

Hi, Vicențiu! A couple of small notes: On Jun 27, noreply(a)launchpad.net wrote: > ------------------------------------------------------------ > revno: 3768 > committer: Vicențiu Ciorbaru <cvicentiu(a)gmail.com> > branch nick: trunk > timestamp: Fri 2013-06-28 01:20:45 +0300 > message: > Changed acl_roles to be stored into a HASH. > > === modified file 'sql/sql_acl.cc' > --- sql/sql_acl.cc 2013-06-25 21:49:50 +0000 > +++ sql/sql_acl.cc 2013-06-27 22:20:45 +0000 > @@ -51,6 +51,7 @@ > #include "hostname.h" > #include "sql_db.h" > #include "sql_array.h" > +#include "sql_hset.h" You should probably include hash.h not sql_hset.h because you don't use Hash_set<> template. > @@ -520,6 +521,13 @@ > return (uchar*) entry->key; > } > > +uchar* acl_role_get_key(ACL_USER *entry, size_t *length, > + my_bool not_used __attribute__((unused))) > +{ > + *length=(uint) (entry->user ? strlen(entry->user) : 0); You know, we generally try to avoid strlens, especially in a code that's used often, like here. Please, replace char *user; in the ACL_USER with LEX_STRING user; > + return (uchar*) entry->user; > +} > + > #define IP_ADDR_STRLEN (3 + 1 + 3 + 1 + 3 + 1 + 3) > #define ACL_KEY_LENGTH (IP_ADDR_STRLEN + 1 + NAME_LEN + \ > 1 + USERNAME_LENGTH + 1) Regards, Sergei

1 0

Re: [Maria-developers] Review request: mdev-4634
by Alexander Barkov 27 Jun '13

27 Jun '13

Hi Sergei, Please review a small patch that fixes the exact problem reported in 4634 (and nothing else). Thanks. (the other problems from the original big patch are now reported separately, with some of them already fixed). On 06/13/2013 12:08 PM, Sergei Golubchik wrote: > Hi, Alexander! > > On Jun 13, Alexander Barkov wrote: >>> Hi, Alexander! >>> >>> This looks like a lot of changes for what you explained the reason of >>> the bug was. There will be lots of "whys" below... >> >> The fix itself was quite small. However, an attempt >> to write a reasonable test case revealed a few other bugs. >> I just tried to fixing everything in a single patch. > > I understood this and it was okay. > > But I still wanted to know what change fixes what bug and what test > case. > >> Okey. Let's do everything separately. >> These are separate reports for the problems >> found during writing a test case for 4634: >> https://mariadb.atlassian.net/browse/MDEV-4652 >> https://mariadb.atlassian.net/browse/MDEV-4653 >> https://mariadb.atlassian.net/browse/MDEV-4654 >> >> Here's a new one, found today: >> https://mariadb.atlassian.net/browse/MDEV-4651 > > Okay, as you prefer :) > > Regards, > Sergei >

2 1

[Maria-developers] query cache issue
by Roberto Spadim 27 Jun '13

27 Jun '13

guys, i sent some (many) MDEV issues to have a better control over query cache i want to comment here to don't stay with this idea in my mind only... i created some labels in JIRA: querycache - relative to query cache qc_info - relative to qc_info plugin querycache_client - relative to implement query cache at client side (not in server) querycache_control - relative to DELETE queries in cache, DELETE tables in cache, LOCK/UNLOCK QUERY CACHE querycache_insert - relative to prune of queries that should be or shouldn't be cacheable querycache_lowmen - realtive to what we could do if we get low men in query cache querycache_result - features relative to when should or shouldn't hit query cache when query is cached if they will be or not implemented is another history is labels ok? should i remove it? -- Roberto Spadim

2 5

Re: [Maria-developers] Rev 3766: Added separation between roles and users in the mysql.user table
by Sergei 26 Jun '13

26 Jun '13

Hi, Vicențiu! On Jun 25, noreply(a)launchpad.net wrote: > ------------------------------------------------------------ > revno: 3766 > committer: Vicențiu Ciorbaru <cvicentiu(a)gmail.com> > branch nick: trunk > timestamp: Tue 2013-06-25 23:52:30 +0300 > message: > Added separation between roles and users in the mysql.user table > modified: > sql/sql_acl.cc > > === modified file 'sql/sql_acl.cc' > --- sql/sql_acl.cc 2013-06-23 12:25:04 +0000 > +++ sql/sql_acl.cc 2013-06-25 20:52:30 +0000 > @@ -542,7 +542,16 @@ > #endif /* HAVE_OPENSSL && !EMBEDDED_LIBRARY */ > #define NORMAL_HANDSHAKE_SIZE 6 > > -static DYNAMIC_ARRAY acl_hosts, acl_users, acl_dbs, acl_proxy_users; > +static DYNAMIC_ARRAY acl_hosts, acl_users, acl_dbs, acl_proxy_users, acl_roles; > +/* XXX > + ***** Potential optimization ***** > + role_grants could potentially be a HASH with keys as usernames and values > + as a DYNAMIC_ARRAY of pointers to granted roles > + XXX > + ***** Implementation choice for now ***** > + An array representing mappings between acl_users and acl_roles; > +*/ Agree about a hash. Hosts, users and dbs are stored in an array, because they support wildcards in MariaDB, not only exact matches. And very specific sort order, from most specific to least specific. But we only need exact matches for role names, so you can put all roles in a hash (see HASH or Hash_set). > +static DYNAMIC_ARRAY role_grants; Same. Do you need an array here? Is the order of entries important? I don't think it is. In fact, you probably may store role<->user relationships directly in the ACL_USER structure. For a user - list of granted roles. For a role - list of users it is granted to. Then it should be easy to check if a given user can SET a given role. > static MEM_ROOT mem, memex; > static bool initialized=0; > static bool allow_all_hosts=1; > @@ -876,11 +891,20 @@ > while (!(read_record_info.read_record(&read_record_info))) > { > ACL_USER user; > + bool is_role= FALSE; > bzero(&user, sizeof(user)); > update_hostname(&user.host, get_field(&mem, table->field[0])); > user.user= get_field(&mem, table->field[1]); > - DBUG_PRINT("info", ("user_entry %s", user.user)); > - if (check_no_resolve && hostname_requires_resolving(user.host.hostname)) > + > + /* If the user entry is a role, skip password and hostname checks > + A user can not log in with a role so some checks are not necessary > + */ > + if (!user.host.hostname) { > + is_role= TRUE; I'm sorry, but it wouldn't do. Apparently it is possible to create a user with the empty host name: MariaDB [test]> create user foo@''; Query OK, 0 rows affected (0.00 sec) MariaDB [test]> select user,host from mysql.user; +------+-----------+ | user | host | +------+-----------+ | foo | | Looking at compare_hostname() function, one can see that empty hostname means "any hostname is ok", it's a wildcard. It means, you cannot use empty hostnames to distinguish between roles and users. Which is a pity, I'd also prefer to use an empty hostname for that. You can use a NULL hostname (you'd need ALTER TABLE to make the column nullable) or a separate column, like, Is_role ENUM('N','Y'). > + } > + > + if (!is_role && check_no_resolve && > + hostname_requires_resolving(user.host.hostname)) Regards, Sergei

1 0

Re: [Maria-developers] Rev 3517: MySQL 5.6.10 performance schema: merge of host_cache table
by Sergei Golubchik 25 Jun '13

25 Jun '13

Hi, Timour! Just few comments On Jun 25, timour(a)askmonty.org wrote: > ------------------------------------------------------------ > revno: 3517 > revision-id: timour(a)askmonty.org-20130621115451-tyxgeeroqyxwz7gn > parent: monty(a)askmonty.org-20130623091543-zqu5dvhry6okna6g > committer: timour(a)askmonty.org > branch nick: 10.0-perf-schema-merge > timestamp: Fri 2013-06-21 14:54:51 +0300 > message: > MySQL 5.6.10 performance schema: merge of host_cache table > === modified file 'sql/hostname.cc' > --- a/sql/hostname.cc 2013-01-15 18:13:32 +0000 > +++ b/sql/hostname.cc 2013-06-21 11:54:51 +0000 > @@ -24,7 +24,7 @@ > Hostnames are checked with reverse name lookup and checked that they > doesn't resemble an IP address. > */ > - > +#include "my_global.h" > #include "sql_priv.h" > #include "hostname.h" > #include "my_global.h" you don't need to include my_global.h twice here > === modified file 'sql/sql_acl.cc' > --- a/sql/sql_acl.cc 2013-06-17 23:01:34 +0000 > +++ b/sql/sql_acl.cc 2013-06-21 11:54:51 +0000 > @@ -8730,7 +8737,6 @@ static int server_mpvio_read_packet(MYSQ > err: > if (mpvio->status == MPVIO_EXT::FAILURE) > { > - inc_host_errors(mpvio->thd->security_ctx->ip); > if (!mpvio->thd->is_error()) > { > if (mpvio->make_it_fail) > @@ -8890,6 +8896,10 @@ static int do_auth_once(THD *thd, const > else > { > /* Server cannot load the required plugin. */ > + Host_errors errors; > + errors.m_no_auth_plugin= 1; > + // merge-todo: struct MPVIO_EXT’ has no member named ‘ip’ > + // inc_host_errors(mpvio->ip, &errors); see above, the line you've removed in a previous chunk, used inc_host_errors(mpvio->thd->security_ctx->ip); > my_error(ER_PLUGIN_IS_NOT_LOADED, MYF(0), auth_plugin_name->str); > res= CR_ERROR; > } > @@ -8913,8 +8923,6 @@ static int do_auth_once(THD *thd, const > Perform the handshake, authorize the client and update thd sctx variables. > > @param thd thread handle > - @param connect_errors number of previous failed connect attemps > - from this host really? you've removed it from the comment, but not from the function, so now acl_authenticate() has one undocumented argument. > @param com_change_user_pkt_len size of the COM_CHANGE_USER packet > (without the first, command, byte) or 0 > if it's not a COM_CHANGE_USER (that is, if > @@ -9017,8 +9025,27 @@ bool acl_authenticate(THD *thd, uint con > > if (res > CR_OK && mpvio.status != MPVIO_EXT::SUCCESS) > { > + Host_errors errors; > DBUG_ASSERT(mpvio.status == MPVIO_EXT::FAILURE); > - > + switch (res) > + { > + case CR_AUTH_PLUGIN_ERROR: > + errors.m_auth_plugin= 1; > + break; > + case CR_AUTH_HANDSHAKE: > + errors.m_handshake= 1; > + break; > + case CR_AUTH_USER_CREDENTIALS: > + errors.m_authentication= 1; > + break; > + case CR_ERROR: > + default: > + /* Unknown of unspecified auth plugin error. */ > + errors.m_auth_plugin= 1; > + break; > + } > + // merge-todo: struct MPVIO_EXT’ has no member named ‘ip’ > + // inc_host_errors(mpvio.ip, &errors); same as above. And personally, I'd rather make a method errors.set_error(res); and move the switch inside. But you may decide to stay closer to 5.6 code. > if (!thd->is_error()) > login_failed_error(thd); > DBUG_RETURN(1); > @@ -9256,23 +9283,20 @@ static int native_password_authenticate( > #endif > > if (pkt_len == 0) /* no password */ > - DBUG_RETURN(info->auth_string[0] ? CR_ERROR : CR_OK); > + DBUG_RETURN(mpvio->acl_user->salt_len != 0 ? CR_AUTH_USER_CREDENTIALS : CR_OK); Why did you change the condition? Our pluggable auth code is not identical to MySQL's, so not all changes can be copied verbatim from there. I think the old condition was ok. > info->password_used= PASSWORD_USED_YES; > if (pkt_len == SCRAMBLE_LENGTH) > { > if (!mpvio->acl_user->salt_len) > - DBUG_RETURN(CR_ERROR); > + DBUG_RETURN(CR_AUTH_USER_CREDENTIALS); > > - if (check_scramble(pkt, thd->scramble, mpvio->acl_user->salt)) > - DBUG_RETURN(CR_ERROR); > - else > - DBUG_RETURN(CR_OK); > + DBUG_RETURN(check_scramble(pkt, thd->scramble, mpvio->acl_user->salt) ? > + CR_AUTH_USER_CREDENTIALS : CR_OK); could you change it back please and only replace the error code? if() is easier to debug, than a function call inside DBUG_RETURN(). I suppose that's why we have it that way. > } > > - inc_host_errors(mpvio->thd->security_ctx->ip); > my_error(ER_HANDSHAKE_ERROR, MYF(0)); > - DBUG_RETURN(CR_ERROR); > + DBUG_RETURN(CR_AUTH_HANDSHAKE); > } Regards, Sergei

1 0

[Maria-developers] Debug interface
by Roberto Spadim 25 Jun '13

25 Jun '13

hi guys, what ide could i use to debug easily in linux mariadb version? i'm using gdb via command line, but it's very complicate, i'm a visual studio user, but a console interface is nice, if i could put some break points and F5 (run),F8 (run one step) there's another place to ask about it? -- Roberto Spadim

3 2

[Maria-developers] RFC: Pausing a query thread
by Marian Marinov 25 Jun '13

25 Jun '13

Hi guys, I work with hosting companies and all of my clients have the same irritating problem, one query taking a lot of time to finish and consuming a lot of resources. So I have solved the same issue but with processes on the machine by simply pausing(kill -STOP) these processes when the load is high and continuing(kill -CONT) them when the server load goes down. I want to implement the same logic within MariaDB, so I can STOP/CONT the thread that is executing this query. My idea is to implement this in the 'KILL query_id' function, so it will accept one more argument, which will be either STOP or CONT. If the parameter is not present, the default behavior should remain. What do you think about my approach? Marian

6 18

Re: [Maria-developers] Help with MDEV-4419, ENUM optimization
by Roberto Spadim 24 Jun '13

24 Jun '13

hummm :) nice, i will study it =) it give me a lot of speed, since some system are closed source and i can't rewrite the query, just change the mysql server or put a mysql proxy

1 5

[Maria-developers] users table
by Roberto Spadim 24 Jun '13

24 Jun '13

hi guys, what's the better way to have a shared mysql schema? i'm talking about creating a user in one database and this user be created in many databases (in the same server) i'm trying with symbolic links in data directory, but sure i have some problems that i need same server version (ok no problem) and flush tables in all servers after add an user i don't know if i could do a two replications since it's like a master-master storage when one database add a user all databases need that information any idea? -- Roberto Spadim

4 4

[Maria-developers] Debug on windows
by Roberto Spadim 24 Jun '13

24 Jun '13

Hi guys i want to learn about optrange.cc, but i dont know how to debug I'm using visual studio 2008 on windows, could anyone help me I will place a break point at get_mm_leaf function and watch all variables Anyone could help me?

1 0

[Maria-developers] Build and/or support of MariaDB on AIX
by Frank Fegert 23 Jun '13

23 Jun '13

Hello all, i'm currently running MySQL 5.1.62 on IBM Power LPARs with AIX 6.1. I'll spare you the discussion of why AIX and not Linux and will just say "it is what it is ATM" ;-) Anyhow, with things proceeding as they are, i guess chances are that Oracle and IBM are not going to get any more friendly with each other, so i'm looking for an alternative way to move forward. That being said, i'd like to get an idea of the general feeling around the MariaDB development team on the support of non-(Linux|*BSD|Windows) platforms. Are there currently any plans? Would you be willing to invest some time in helping to get MariaDB running on AIX? Right now, i can successfully build stock MySQL 5.1.70 with some minor patches and stock MySQL 5.5.32 with some minor patches and the option "DWITHOUT_PERFSCHEMA_STORAGE_ENGINE=ON". With performance schema enabled the build fails at linking the unittests in "storage/perfschema/unittest/", see below for a full error output. The MariaDB 5.5.31 build does currently not even go that far, if fails earlier on the threadpool code, see below. Unfortunately my C/C++-fu is not strong enough to determine what might be considered a sound alternative to the epoll/kevent/port_event calls on AIX to do an alteration like the one in MDEV-4174. Can someone please help me out here? My build environment: Compiler: IBM XL C/C++ for AIX, V11.1 (5724-X13) Version: 11.01.0000.0009 AIX oslevel: 6100-08-01-1245 Thanks & best regards, Frank Fegert ## Begin: MariaDB 5.5.31 ############################################## ... [ 37%] Building CXX object sql/CMakeFiles/sql.dir/threadpool_unix.cc.o cd /home/user/rpmbuild/tmp/mariadb-5.5.31/build-mysql/sql && /usr/vacpp/bin/xlC_r -+ -DHAVE_CONFIG_H -DHAVE_EVENT_SCHEDULER -DHAVE_OPENSSL -DHAVE_POOL_OF_THREADS -DMYSQL_SERVER -ma -O2 -qlanglvl=extc99 -qstrict -qoptimize=2 -qmaxmem=16384 -DSYSV -D_AIX -D_AIX32 -D_AIX41 -D_AIX43 -D_AIX51 -D_AIX52 -D_AIX53 -D_AIX61 -D_ALL_SOURCE -DFUNCPROTO=15 -I/opt/freeware/include -qthreaded -qhalt=e -qstaticinline -g -DDBUG_OFF -I/home/user/rpmbuild/tmp/mariadb-5.5.31/build-mysql/include -I/home/user/rpmbuild/tmp/mariadb-5.5.31/include -I/home/user/rpmbuild/tmp/mariadb-5.5.31/sql -I/home/user/rpmbuild/tmp/mariadb-5.5.31/regex -I/home/user/rpmbuild/tmp/mariadb-5.5.31/build-mysql/sql -o CMakeFiles/sql.dir/threadpool_unix.cc.o -c /home/user/rpmbuild/tmp/mariadb-5.5.31/sql/threadpool_unix.cc "/home/user/rpmbuild/tmp/mariadb-5.5.31/sql/threadpool_unix.cc", line 39.2: 1540-0859 (S) #error directive: threadpool is not available on this platform. gmake[2]: *** [sql/CMakeFiles/sql.dir/threadpool_unix.cc.o] Error 1 gmake[2]: Leaving directory `/home/user/rpmbuild/tmp/mariadb-5.5.31/build-mysql' gmake[1]: *** [sql/CMakeFiles/sql.dir/all] Error 2 gmake[1]: Leaving directory `/home/user/rpmbuild/tmp/mariadb-5.5.31/build-mysql' gmake: *** [all] Error 2 ## End: MariaDB 5.5.31 ################################################ ## Begin: MySQL 5.5.32 ################################################ ... Linking CXX executable pfs-t cd /home/user/rpmbuild/tmp/mysql-5.5.32/build-mysql/storage/perfschema/unittest && /opt/freeware/bin/cmake -E cmake_link_script CMakeFiles/pfs-t.dir/lin k.txt --verbose=1 /usr/vacpp/bin/xlC_r -ma -O2 -qlanglvl=extc99 -qstrict -qoptimize=2 -qmaxmem=16384 -DSYSV -D_AIX -D_AIX32 -D_AIX41 -D_AIX43 -D_AIX51 -D_AIX52 -D_AIX53 -D_A IX61 -D_ALL_SOURCE -DFUNCPROTO=15 -I/opt/freeware/include -qthreaded -qhalt=e -qstaticinline -L/opt/freeware/lib/mysql -L/opt/freeware/lib -Wl,-bmaxdata:0x80000000 CMakeFiles/pfs-t.dir/pfs-t.cc.o -o pfs-t -Wl,-brtl,-bnoipath,-bexpfull ../../../unittest/mytap/libmytap.a ../libperfschema.a ../../../mysys/libmysys.a ../../../dbug/libdbug.a ../../../mysys/libmysys.a ../../../dbug/libdbug.a ../../../strings/libstrings.a -lz -lm -Wl,-blibpath:/opt/freeware/lib:/usr/lib:/lib ld: 0711-317 ERROR: Undefined symbol: .MDL_request::init(MDL_key::enum_mdl_namespace,const char*,const char*,enum_mdl_type,enum_mdl_duration) ld: 0711-317 ERROR: Undefined symbol: .LEX::LEX() ld: 0711-317 ERROR: Undefined symbol: .lex_start(THD*) ld: 0711-317 ERROR: Undefined symbol: Virtual table for class "DML_prelocking_strategy": Some possible causes are: first non-inline virtual function in "DML_prelocking_strategy" is not defined; or the class is a template instantiation and an explicit instantiation definition of the class is missing. ld: 0711-317 ERROR: Undefined symbol: .open_and_lock_tables(THD*,TABLE_LIST*,bool,unsigned int,Prelocking_strategy*) ld: 0711-317 ERROR: Undefined symbol: .Table_check_intact::check(TABLE*,const st_table_field_def*) ld: 0711-317 ERROR: Undefined symbol: .close_thread_tables(THD*) ld: 0711-317 ERROR: Undefined symbol: .lex_end(LEX*) ld: 0711-317 ERROR: Undefined symbol: .Query_tables_list::destroy_query_tables_list() ld: 0711-317 ERROR: Undefined symbol: .plugin_unlock_list(THD*,st_plugin_int**,unsigned int) ld: 0711-317 ERROR: Undefined symbol: THR_THD ld: 0711-317 ERROR: Undefined symbol: .sql_print_error(const char*,...) ld: 0711-317 ERROR: Undefined symbol: Virtual table for class "st_select_lex_node": Some possible causes are: first non-inline virtual function in "st_select_lex_node" is not defined; or the class is a template instantiation and an explicit instantiation definition of the class is missing. ld: 0711-317 ERROR: Undefined symbol: lower_case_table_names ld: 0711-317 ERROR: Undefined symbol: .Field_enum::store_type(unsigned long long) ld: 0711-317 ERROR: Undefined symbol: .ACL_internal_schema_registry::register_schema(const st_mysql_lex_string*,const ACL_internal_schema_access*) ld: 0711-317 ERROR: Undefined symbol: handler::clone(const char*,st_mem_root*) ld: 0711-317 ERROR: Undefined symbol: handler::get_error_message(int,String*) ld: 0711-317 ERROR: Undefined symbol: handler::column_bitmaps_signal() ld: 0711-317 ERROR: Undefined symbol: handler::index_read_idx_map(unsigned char*,unsigned int,const unsigned char*,unsigned long,ha_rkey_function) ld: 0711-317 ERROR: Undefined symbol: handler::index_next_same(unsigned char*,const unsigned char*,unsigned int) ld: 0711-317 ERROR: Undefined symbol: handler::read_multi_range_first(st_key_multi_range**,st_key_multi_range*,unsigned int,bool,st_handler_buffer*) ld: 0711-317 ERROR: Undefined symbol: handler::read_multi_range_next(st_key_multi_range**) ld: 0711-317 ERROR: Undefined symbol: handler::read_range_first(const st_key_range*,const st_key_range*,bool,bool) ld: 0711-317 ERROR: Undefined symbol: handler::read_range_next() ld: 0711-317 ERROR: Undefined symbol: handler::read_first_row(unsigned char*,unsigned int) ld: 0711-317 ERROR: Undefined symbol: handler::get_dynamic_partition_info(PARTITION_STATS*,unsigned int) ld: 0711-317 ERROR: Undefined symbol: handler::get_auto_increment(unsigned long long,unsigned long long,unsigned long long,unsigned long long*,unsigned long long*) ld: 0711-317 ERROR: Undefined symbol: handler::drop_table(const char*) ld: 0711-317 ERROR: Undefined symbol: Virtual table for class "handler": Some possible causes are: first non-inline virtual function in "handler" is not defined; or the class is a template instantiation and an explicit instantiation definition of the class is missing. ld: 0711-317 ERROR: Undefined symbol: .handler::ha_table_share_psi(const TABLE_SHARE*) const ld: 0711-317 ERROR: Undefined symbol: .handler::ha_statistic_increment(unsigned long system_status_var::*) const ld: 0711-317 ERROR: Undefined symbol: .handler::print_error(int,int) ld: 0711-317 ERROR: Undefined symbol: key_map_empty ld: 0711-317 ERROR: Undefined symbol: .calculate_key_len(TABLE*,unsigned int,const unsigned char*,unsigned long) ld: 0711-345 Use the -bloadmap or -bnoquiet option to obtain more information. gmake[2]: *** [storage/perfschema/unittest/pfs-t] Error 8 gmake[2]: Leaving directory `/home/user/rpmbuild/tmp/mysql-5.5.32/build-mysql' gmake[1]: *** [storage/perfschema/unittest/CMakeFiles/pfs-t.dir/all] Error 2 gmake[1]: Leaving directory `/home/user/rpmbuild/tmp/mysql-5.5.32/build-mysql' gmake: *** [all] Error 2 ###End: MySQL 5.5.32 ##################################################

3 3

Re: [Maria-developers] Build and/or support of MariaDB on AIX
by Frank Fegert 23 Jun '13

23 Jun '13

Hello Harald, On Sun, Jun 23, 2013 at 12:37:24PM +0200, Reindl Harald wrote: > your outoput contains "#error directive: threadpool is not available > on this platform" which clearly states you must not configure > "-DHAVE_POOL_OF_THREADS" on your platform thanks for your fast reply! Yes, to disable threadpool for AIX would be an option, but unfortunately the /dev/human attached to my computer seems to be defective ;-) All fun aside, i went through the configure option list, but nothing related to threadpools caught my eye. Manually patching the above option out of "sql/CMakeLists.txt" didn't help either. I'm missing something here, aren't i? Thanks & best regards, Frank

1 0

[Maria-developers] Uninstalling MariaDB 5.2 does not remove everything.
by Peter Laursen 22 Jun '13

22 Jun '13

For @wlad I guess: After uninstalling MariaDB 5.2.on Win7/64 some (don't remember which minor version), it remains in the list of installed programs in Control Panel .. Programs. Trying to uninstall displays this dialog (translates to "This action is only valid for programs currently installed"). I did not try if rebooting system would remove the item from the list. It may also have been fixed in the installer since the 5.2 release I installed. -- Peter

2 5

Re: [Maria-developers] [GSoC] Kerberize MariaDB -- some unclear point about the project
by QIU Shuang 22 Jun '13

22 Jun '13

Hi wlad, Thank you for your concern. > Create user > 'foo@bar'@localhost creates user foo@bar, on localhost. > create user that is identified with name and domain and can connect > from any computer I admit that I mis-understood the usage of User@Host in MariaDB. I thought the User and Host fields in MariaDB are in the same place as those in a Kerberos principal. i.e. if my MariaDB login name is qiush(a)sjtu.edu.cn, then my Kerberos principal will be qiush(a)sjtu.edu.cn/CHINA, where MariaDB login name is part of Kerberos principal. (if that case, the realm part is omitted in MariaDB, and we should find another way to figure it out. That's what I argued in my previous email.) From your reply, it seems qiush@sjtu.edu.cn/CHINA@sjtu.edu.cn, the bold part is MariaDB User and italic part Host, can be a valid login name in our project. Suddenly realise the Host in MariaDB login name will constraint the user login place. It's much clear now. > Re realm, I do not know this much but 'shuang(a)uni.shanghai.edu/REALM' also > does not look too weird to me. To me, either :). > Or perhaps I miss something still? Can you elaborate? No, you're right. I confused these two names. Thank you for you hints! Sincerely, Shuang On Jun 20, 2013, at 2:22 AM, Vladislav Vaintroub <wlad(a)montyprogram.com> wrote: > > > From: QIU Shuang [mailto:qiush.summer@gmail.com] > Sent: Mittwoch, 19. Juni 2013 19:52 > To: Vladislav Vaintroub > Subject: Re: [Maria-developers] [GSoC] Kerberize MariaDB -- some unclear > point about the project > > > Hi Shuang, > >>> Trying to make a nicer name, for example by removing domain part could > introduce some ambiguity here and different Kerberos users to login as the > same. >> I think so. >> But per my knowledge, the fully qualified name in MariaDB is > username@hostname. >> What about the realm/domain part? >> I think this may be a gap between MariaDB and Kerberos. > > Maybe I oversee something, but I do not really see any contradiction here. > Do you mean that @ is special character should not be used in usernames? It > actually can, it just must be properly escaped. Create user > 'foo@bar'@localhost creates user foo@bar, on localhost. > Hypothetical CREATE USER 'shuang(a)uni.shanghai.edu' @'%' IDENTIFIED WITH > 'Kerberos' > > will create user that is identified with name and domain and can connect > from any computer (due to use of wildcard for computername part, this > wildcard can be omitted). > Re realm, I do not know this much but 'shuang(a)uni.shanghai.edu/REALM' also > does not look too weird to me. > > Or perhaps I miss something still? Can you elaborate? > > Wlad >

3 4

[Maria-developers] help with parser
by Roberto Spadim 20 Jun '13

20 Jun '13

hi guys i'm trying to implement MDEV-4579<https://mariadb.atlassian.net/browse/MDEV-4579>, could anyone help me with parser + warning generation? MDEV-4579 title: SQL PARSER - Add Ignore to SET SQL command example: SET @@some_global_variable_that_dont_exists=1 *this cause an error* SET @@some_global_variable_that_dont_exists=1,@@some_variable_that_exists_and_should_be_changed=3 *this cause an error and don't change the second variable* *could we implement:* SET *IGNORE * @@some_global_variable_that_dont_exists=1,@@some_variable_that_exists_and_should_be_changed=3 *this cause an warning (first variable don't exists) and change the second variable* thanks -- Roberto Spadim

2 1

[Maria-developers] [GSoC] Kerberize MariaDB -- some unclear point about the project
by QIU Shuang 19 Jun '13

19 Jun '13

Hi Wlad and the community, I've developed a Kerberos plugin prototype with both raw API and GSSAPI supported on my own Linux box. The codes are pushed to my private repo already. It is just a prototype, which can handle only the normal case, without any tests in deployed environment. And I know there is a long way to go before accepted I now have a clear knowledge about the control flow of authentication plugin and Kerberos protocol. The working use case is like this: 1. login my computer as, for example, qiush; 2. create a user in MariaDB named qiush; 3. get the initial credential from Kerberos Key Distribution Center (KDC), i.e. tgt; 4. connect mysqld as qiush and now I can login DBMS passwordlessly before tgt expires. Correctly me if I miss something in this case. The exception case is how to get/renew tgt when the original expired. I have created a private repo based on 5.5 release version, which is available at https://launchpad.net/~qiush-summer. For those that concern about our project, you can track my progress. (The reason why I choose 5.5 release version instead of the latest development branch 10.x is that I cannot clone the repo lp:~maria-captains/maria/10.0-base from launchpad and I don't know why. I'll figure it out later if necessary). Some problems (three, exactly) when implementing the toy plugin. The first one is about the private repo and bazaar (I'm a git guy and have little experience working with bazaar). In order to make some demonstrated progress, I push my local update to the private repo each time I create a new commit. It seems the commit number is global (see the screen shot of my private repo history below). Recent revisions 3796. By Shuang Qiu 20 hours ago prototype with GSSAPI enabled 3795. By Shuang Qiu on 2013-06-15 Kerberos plugin prototype with raw krb5 API 3794. By Shuang Qiu on 2013-06-13 doodle 3793. By Michael Widenius on 2013-06-11 Fixed tests that failed on 32 bit because of my earlier fixes of 32 bit limits. 3792. By Sergei on 2013-06-07 MDEV-4468 Assertion `error != 0' fails or timeout occurs on select from a FEDERATED table which points at a non-existent table I'm afraid that my commits may pollute the code base. So the question is that will others (e.g. Sergei since commit 3792) see my changes next time he gets local branch updated. Is it the best practice to use bazaar like this (push after each commit)? If not, could you please give me some suggestion to use bazaar and launchpad correctly in our community? The other two are about the exception case. When get the initial credential from KDC, clients/services are asked for a password or asked to provide a keytab file to derive a long-term key that can be used encrypting messages between clients and KDC. On the server/service side, I think the keytab approach is better. Otherwise, each time we restart the service, a password will be asked. It may be annoying, I think :p. Then how to figure out the name of keytab file? I can come up with three approaches: 1. hard coded the filename and document the keytab file name clearly in the manual for users to follow (not the best choice, I think); 2. use the default Kerberos setting, which is /etc/krb5.keytab when properly configured; 3. make the filename an option in configuration file, e.g. ~/.my.cnf, with a default value /etc/krb5.keytab, since it is the default Kerberos setting. I prefer the third solution. As to the client side, I think both are OK. The problem of keytab choice is the same. Then I try to get some hints from similar password requesting authentication plugin. I read code in plugin/auth_dialog and plugin/auth_examples where server plugin and client plugin are implemented separately. In the client plugin, a external function of type mysql_authenticaton_dialog_ask_t static mysql_authentication_dialog_ask_t ask; can be provided as an alternative to command line input. Should I also implement them in two parts? And also provide the mysql_authentication_dialog_ask_t hook? The last problem. Kerberos requires user to provide a service name to resolve before requesting service ticket. How to name our service? Hard coded it as MYSQL or set in configuration file (I prefer the latter one)? Let me know your thoughts :). Thank you very much and best regards! Sincerely, Shuang

2 2

Re: [Maria-developers] [GSoC] Kerberize MariaDB -- some unclear point about the project
by QIU Shuang 19 Jun '13

19 Jun '13

Hi serg, Thank you for your comment. I will clean my repo at some later time follow your advice. :) Sincerely, Shuang On Jun 18, 2013, at 4:51 PM, Sergei Golubchik <serg(a)mariadb.org> wrote: > Hi, QIU Shuang! > > Just to answer the bzr part: > > On Jun 16, QIU Shuang wrote: >> The first one is about the private repo and bazaar (I'm a git guy and >> have little experience working with bazaar). In order to make some >> demonstrated progress, I push my local update to the private repo each >> time I create a new commit. It seems the commit number is global (see >> the screen shot of my private repo history below). >> >> Recent revisions >> 3796. By Shuang Qiu 20 hours ago >> prototype with GSSAPI enabled >> 3795. By Shuang Qiu on 2013-06-15 >> Kerberos plugin prototype with raw krb5 API >> 3794. By Shuang Qiu on 2013-06-13 >> doodle >> 3793. By Michael Widenius on 2013-06-11 >> Fixed tests that failed on 32 bit because of my earlier fixes of 32 bit limits. >> 3792. By Sergei on 2013-06-07 >> MDEV-4468 Assertion `error != 0' fails or timeout occurs on select from a FEDERATED table which points at a non-existent table >> >> I'm afraid that my commits may pollute the code base. So the question >> is that will others (e.g. Sergei since commit 3792) see my changes >> next time he gets local branch updated. Is it the best practice to >> use bazaar like this (push after each commit)? > > Don't worry about it. Revision numbers in bzr aren't stable, during a > merge they might be renumbered (*). Revision ids are stable and unique, > but they look like > > sergii(a)pisem.net-20130607133513-p27s1bg8gt0amkiw > > (this is for my commit with revno 3792 above). > Think of revision numbers as of convenient shortcuts that help to avoid > typing and that are easy to remember. But don't use them for long-term > references or something. > > And no, I won't see your changes when I pull from 5.5. > Not unless I pull from your repository. > > Regards, > Sergei > > [1]. Details. Example. Say you have a repo with only one changeset. It > has revno 1. You've branched. And you've made changes and committed > something in both branches, independently. Every branch has now two > changesets. Changeset 1 (it's the same in both, save revid), and > Changeset 2 (revno is the same, revid is different). Now if you merge > the second branch into the first, the changeset 2 of the second branch > will be renumbered, it'll get revno 1.1 (but its revid won't change). > The history will look like (Changeset 3 is the merge changeset): > > 1 > | \ > 2 1.1 > | / > 3 > >

1 0

[Maria-developers] mysql 5.6, 5.7
by Roberto Spadim 18 Jun '13

18 Jun '13

hi guys, i was reading mysql 5.6 and 5.7 manual i know that mariadb is very different (a bit) than mysql internally there's a mysql merge issue or something like that i can see what will be merged and what not? i like the partition lock prune, memcache and optimizer tracer of mysql 5.6, but today i'm not using it in mariadb (not a problem, but it's nice features) -- Roberto Spadim

3 6

Re: [Maria-developers] review / close MDEV-4581
by Roberto Spadim 18 Jun '13

18 Jun '13

nice! thanks sergei! i think it's now rich in features, maybe should be used by community to check if it's usefull (there's many columns) thanks again! 2013/6/18 Sergei Golubchik <serg(a)mariadb.org> > Hi, Roberto! > > On Jun 18, Roberto Spadim wrote: > > hi guys, could review / close > > MDEV-4581<https://mariadb.atlassian.net/browse/MDEV-4581> ? > > Roberto, as you can see in Jira, > MDEV-4581 is assigned to me, it's targeted for the 10.0.4 release. > So yes, I could and I will. > > Regards, > Sergei > -- Roberto Spadim SPAEmpresarial

1 0

Re: [Maria-developers] close some mdev
by Roberto Spadim 18 Jun '13

18 Jun '13

hi sergei! thanks about others mdev, and yes mdev 4583 shouldn't be closed, i will study a bit more to see what can be done to write a patch about it and test the speed of a procedure removing query cache queries 2013/6/18 Sergei Golubchik <serg(a)mariadb.org> > Hi, Roberto! > > On Jun 17, Roberto Spadim wrote: > > hi guys could anyone close some opened and fixed (won't fix) mdev's ? the > > comment explain why to close it > > > > MDEV-4571 <https://mariadb.atlassian.net/browse/MDEV-4571> > > MDEV-4589 <https://mariadb.atlassian.net/browse/MDEV-4589> > > MDEV-575 <https://mariadb.atlassian.net/browse/MDEV-575> > > Closed. > > > MDEV-4583 <https://mariadb.atlassian.net/browse/MDEV-4583> > > Not closed, it doesn't have a comment like the three above? > Should I still close it? > > Regards, > Sergei > -- Roberto Spadim SPAEmpresarial

1 0

[Maria-developers] a problem of Parallel replication
by nanyi607rao 18 Jun '13

18 Jun '13

hi, when i read patchs from http://bazaar.launchpad.net/~knielsen/maria/dingqi-parallel-replication/cha…, i found a problem (maybe would cause bug) :main SQL thread read a event from relay log, it make the log_event->thd point to the thd of main SQL thread. But when a log_event is disaptched to a sub SQL thread, the log_event->thd still point to the old value instead of the thd of sub SQL thread. this problem will lead check_stack_overrun() fails, then lead open_table() fails... maybe this will be helpful to you . 2013-06-18 nanyi607rao

1 0

[Maria-developers] review / close MDEV-4581
by Roberto Spadim 18 Jun '13

18 Jun '13

hi guys, could review / close MDEV-4581<https://mariadb.atlassian.net/browse/MDEV-4581> ? -- Roberto Spadim

1 0

[Maria-developers] Implementing implicit primary key in mysql server
by Lixun Peng 18 Jun '13

18 Jun '13

Hi, As we know, InnoDB has implicit primary key if a table hasn't defined a primary key. However MySQL server doesn't know this primary key, so this primary key will not apear in binlog. When we are using ROW format binlog for replication, if a table has no any indexes, that's a disaster. If a table without indexes do a DML (UPDATE/DELETE), of course it will run a long time in master, but in slave, it still need a long time. It will cause serious slave replication delay. For example: binlog_format=ROW; CREATE TABLE table1 ( col1 varchar(32), col2 int ); table1 hasn't any indexes, if table1 has 100 million rows, I do a query "UPDATE table1 SET col1='xxx' WHERE col2=100", and this query will effect 1 million rows. Then it has 1 million UPDATE events in binlog. After slave received these events, each events should scan all rows!!! It will cause serious replication delay! I saw:https://kb.askmonty.org/en/row-based-replication-with-no-primary-key/ But if no any indexes, it still a problem. I think mysql server can do the same thing as InnoDB do, if user doesn't define the primary key, mysql can add the primary key automatically. For example: If user create a table like "table1", we can do "ALTER TABLE table1 ADD __id bigint auto_increment, ADD PRIMARY KEY(__id)", so in fact, the "table1" structure is CREATE TABLE table1 ( col1 varchar(32), col2 int, __id bigint auto_increment primary key ); But when user do "SELECT * FROM table xxx" or "INSERT INTO table VALUES(xx)", we should ignore implicit primary key, "__id" will not apear in the result set. How do you think? Thanks, Lixun -- Senior Database Engineer @ Taobao.com Mobile Phone: +86 18658156856 (Hangzhou) Gtalk: penglixun(at)gmail.com Twitter: http://www.twitter.com/plinux Blog: http://www.penglixun.com

1 0

[Maria-developers] close some mdev
by Roberto Spadim 18 Jun '13

18 Jun '13

hi guys could anyone close some opened and fixed (won't fix) mdev's ? the comment explain why to close it 1. MDEV-4571 <https://mariadb.atlassian.net/browse/MDEV-4571> 1. MDEV-4583 <https://mariadb.atlassian.net/browse/MDEV-4583> 1. MDEV-4589 <https://mariadb.atlassian.net/browse/MDEV-4589> 1. MDEV-575 <https://mariadb.atlassian.net/browse/MDEV-575> -- Roberto Spadim SPAEmpresarial

1 0

Re: [Maria-developers] Jira new field
by Roberto Spadim 18 Jun '13

18 Jun '13

ok =) thanks 2013/6/17 Sergei Golubchik <serg(a)mariadb.org> > Hi, Roberto! > > On Jun 17, Roberto Spadim wrote: > > > > today i'm 'working around ' this, and putting the 'group information' in > > MDEV issue, example.. query cache things i put > > QUERY CACHE - some issue title > > processlist things i put > > PROCESSLIST - some issue title > > > > this works too, but well. mariadb team can change and when i lost my > memory > > (some times =) ) i need to read all issues again and find where i put the > > information > > You can certainly add labels like QUERY CACHE and PROCESSLIST. This > shouldn't be a problem. > > Regards, > Sergei > -- Roberto Spadim SPAEmpresarial

1 0

Re: [Maria-developers] Jira new field
by Roberto Spadim 18 Jun '13

18 Jun '13

heeh sometimes i just add a label that don't have interest to mariadb team, it's just to 'group' my issues if there's a interesting and nice way to do this could be nice it's just to organize my issues and when i have time see what changed and where, not changing the work of mariadb placing the right label in the right issue 2013/6/17 Sergei Golubchik <serg(a)mariadb.org> > Hi, Roberto! > > On Jun 17, Roberto Spadim wrote: > > hi guys, could we add a Reporter Label inside jira issue? instead of only > > Label (used by mariadb team) > > it's easier to filter issues, thinking about the reporter filter not for > > mariadb team filters > > But you can add your labels to the Label field, why would one need > another list of Labels? > > Regards, > Sergei > -- Roberto Spadim SPAEmpresarial

1 3

[Maria-developers] Jira new field
by Roberto Spadim 17 Jun '13

17 Jun '13

hi guys, could we add a Reporter Label inside jira issue? instead of only Label (used by mariadb team) it's easier to filter issues, thinking about the reporter filter not for mariadb team filters -- Roberto Spadim

1 0

[Maria-developers] Is it possible to export last seen GTID from IO thread?
by Pavel Ivanov 15 Jun '13

15 Jun '13

Kristian, We had an idea that can make master failovers much easier, please could you tell us how feasible to do that in MariaDB. Currently when master dies and failover is necessary we have to wait while SQL thread catches up with IO thread on all replicas before we'll be able to decide which replica should be new master. Otherwise we can loose some transactions. But if IO thread could export some variable indicating which GTID it's seen last then we'll be able to understand which replica is furthest in replication stream faster, without waiting. Of course for multi-domain replication it will be more complicated and IO thread will have to synthesize gtid_slave_pos which server would have after applying all relay logs. But exporting this synthesized position will still help a lot. So what do you think? Is it possible to do this in MariaDB? Thank you, Pavel

1 0

Re: [Maria-developers] Process suggestion for minor issues
by Elena Stepanova 15 Jun '13

15 Jun '13

(warning: it's a long one) Hi Sergei, On 6/14/2013 6:26 PM, Sergei Golubchik wrote: > Hi, Elena! > > On Jun 14, Elena Stepanova wrote: >> On 6/14/2013 2:27 PM, Rasmus Johansson wrote: >>> Here is a suggested process to handle the ever growing amount of >>> minors. If a minor issue isn't fixed after two consecutive releases >>> it's re-reviewed and either priority changed to major (or higher), >>> or the minor issue is closed with resolution Won't Fix and a >>> comment. Example of comment can be that "this is an edge case where >>> effort and benefits doesn't correlate". >> >> I don't think the algorithm is quite fair. It means that if a really >> minor bug happens to be found during a relatively slow period, it gets >> fixed, but another bug, even considerably more important (but still >> falling into "not major => minor" category), can be by this time closed >> as "not worth fixing" just because people had been busy with some other >> stuff and couldn't fix it over two previous releases; while logically, >> if somebody suddenly has time to fix a minor bug, the latter should be >> fixed first. > > This algorithm has a "re-reviewed and either priority changed to > major..." step, so it's not just the timing. Not like automatic closing > of old bugs. Yes, I got that part; but it doesn't change the potential faulty result. Lets say we have imaginary priorities 1-5 which are all "minor", 1 being the lowest. For a while you had had a bug of "priority 4". After 2 releases the reviewing force reviewed it, decided it was minor (rightly so), and closed it. A few days later you got a bug of "priority 1". A few more days later you suddenly have time to fix a bug -- you are stuck at your main task and need a distraction, or you've got a headache and can't work on anything serious, or you only have half an hour before you need to go pick up kids... whatever. Naturally, you go to the list of your open bugs, take a quick look and pick whatever fits. If you still had the bug of priority 4, you'd fix it now, but instead you fix the bug of priority 1. I don't know about others, but I do keep a bunch of open low-priority tasks, and while rarely, sometimes it happens that I can do something about a minor one. And then I look at them and pick by importance and "suitability for the moment", not by the creation time. > >> I'm fine with closing bugs as "Won't Fix", but imho it should be based >> on the merits of the bug itself, not on the bad timing of its >> creation. Since all our developers are very senior and can totally >> make such a decision on their own, I suppose anyone who a bug is >> assigned to can take a look at it and either close it as "Won't fix" >> with a proper explanation, regardless the timing (even right after >> receiving it), or keep it open. > > The main issue I didn't like - what that algorithm is making an attempt > of solving - is a growing pile of minor bugs that we aren't going to > fix. I agree, there is no point at keeping those that we are really not going to fix. But I don't see anything critically bad at preserving those that we *want* to fix, just currently don't have time to -- as long as we don't schedule them for the closest release, thus lying to users and to ourselves (more on that later). I also think it's two essentially different categories. Later here you are saying that "a closed bug can always be reopened later and fixed". If we put the "don't have resources to fix" bugs into the same trash bin as we currently use for "really won't fix" bugs, re-opening will never happen. Who will go through the whole bunch of "Won't fix" bugs, review them *again* and see if they are really not worth fixing, or if we had to close them because we didn't have resources? Certainly you can't do it when you just have little time to fix one-two small bugs, you'll spend more time searching. > > A bug is real, so it's difficult to say "won't fix", it's much easier, > psychologically, to say, "okay, I'll fix it, *when I have time*". But > really, for minor bugs this might never happen. But sooner or later somebody who's reviewing them will dare close it anyway?.. Okay, I agree that psychologically it is easier to reject a bug *together with somebody* than make such a decision on your own, group mentality and all that, I get it and have it myself; so if it's all about having some kind of committee for closing such bugs, let it be. (I don't really know who could this committee consist of. Generally, I'm a big supporter of a triage team, it makes the process less personal. But in a company where everyone is pretty much unique in his area, it's a bit strange, because at the end only the expert can bring in an educated verdict -- and this would normally be the person the bug is already assigned to anyway; others will just be there to hold his hand.) > And de-facto this leads to wrong expectations (users believe the bug > will be fixed in a specific version, but it isn't) and us lying to our > users. True, but I think the new approach will mess up more than it will fix. Now we have Fix version evolution like that: 5.5.30 => 5.5.31 => 5.5.32 => 5.5.33 => .... It's bad. I still think that we shouldn't schedule minor bugs for any particular version. We've discussed it before, and yes, I know that we do it to keep them on somebody's radar, but I would think that having it assigned to somebody should be enough, since it's already on a radar of the person who it's assigned to. But maybe I'm wrong and nobody looks at lists of bugs assigned to them -- shouldn't we try to change it instead, then? But with the new system, we will have 5.5.30 => 5.5.31 => "Won't fix, it isn't worth it". If I were a user, now I'd be mad. If it isn't worth fixing, why didn't you say so right away? Why schedule it and even move it to the next release then? That's even more of lying to users than we currently do, and in a more irritating way. > My idea is to be honest about it, and if we don't have resources to fix > a specific bug, we won't drag it forever from one release to another, > but close it eventually. I agree. But I think we shouldn't even start dragging it then. If you are going to have a committee which would regularly review bugs (those that have waited for 2 releases) and decide if they're worth fixing or not, why not do it right away -- let the same committee decide it for *new* bugs instead, so that unimportant ones don't even get scheduled and go to the trash bin immediately. It will spare users wrong expectations. > And yes, I hate closing non-fixed bugs too. > > If you can suggest another solution, please do. But "keep it open" > doesn't work, as we've seen already. Well, I'd suggest to consider adjusting yours a little, in the following way: - define who will be reviewing the bugs; - those person(s) will screen new bugs instead of those that have waited for 2 releases; - , set "Won't fix" when we really mean it (example: PBXT bugs, etc.), and create a new resolution "Don't have time to fix" or "Wishlist" or whatever, and put corresponding bugs there, while keeping them assigned. The last point is to preserve at least hypothetical possibility for the bugs to be re-opened later. So, if somebody has time to fix a minor bug, first he looks at the list of his open bugs (we already know that the committee kept only most worthy ones there). If nothing fits the moment, or (a miracle!) the list is empty, the somebody goes to the "wishlist" bugs assigned to him, and picks from there. At the same time, those bugs aren't shown as open, aren't scheduled for any particular release, the users aren't misled, etc. A couple more notes: - I think external bugs should generally have higher priority than otherwise equal ones found internally; - a different approach might be needed for bugs in new feature trees, it should be decided separately (the basic question is -- when we consider the feature good enough to be pushed to the main tree, whether it requires fixing minor bugs or not). /E > > Besides, a closed bug can always be reopened later and fixed. > > Regards, > Sergei >

2 1

[Maria-developers] Process suggestion for minor issues
by Rasmus Johansson 14 Jun '13

14 Jun '13

Hi MariaDB Developers, This is a suggestion on how to deal with minor or lower priority bugs and tasks in JIRA, http://mariadb.org/jira . Currently, the process we use makes a clear differentiation between major or higher prioirty issues (called majors altogether after this) and minor or lower priority issues (called minors altogether after this). The majors are the ones that should be fixed for a release and the minors are rolling, meaning that they are done when the corresponding developer has finished his/her majors. This of course sounds logical, but it also leads to a growing number of minors for every consecutive release, many of them which never will be fixed but in the current process will hang around. Here is a suggested process to handle the ever growing amount of minors. If a minor issue isn't fixed after two consecutive releases it's re-reviewed and either priority changed to major (or higher), or the minor issue is closed with resolution Won't Fix and a comment. Example of comment can be that "this is an edge case where effort and benefits doesn't correlate". It should also be noted that after closing an issue in JIRA, it doesn't mean that it disappears. It will still show up in free text searches and among the issues for the version where it was closed. Please provide your thoughts. Do you think it's an OK way to deal with the minors or do you prefer some other way? Best regards, Rasmus

6 8

Re: [Maria-developers] Review request: mdev-4635
by Alexander Barkov 14 Jun '13

14 Jun '13

Hi Sergei, On 06/14/2013 04:30 PM, Sergei Golubchik wrote: > Hi, Alexander! > > On Jun 13, Alexander Barkov wrote: >> On 06/13/2013 08:16 PM, Sergei Golubchik wrote: >>> >>>> === modified file 'mysql-test/r/strict.result' >>>> --- mysql-test/r/strict.result 2011-06-05 02:56:06 +0000 >>>> +++ mysql-test/r/strict.result 2013-06-13 12:15:43 +0000 >>>> @@ -213,11 +213,11 @@ ERROR 22007: Incorrect date value: '2004 >>>> INSERT INTO t1 (col1) VALUES(STR_TO_DATE('0.10.2004 15.30','%d.%m.%Y %H.%i')); >>>> ERROR 22007: Incorrect date value: '2004-10-00 15:30:00' for column 'col1' at row 1 >>>> INSERT INTO t1 (col1) VALUES(STR_TO_DATE('31.9.2004 15.30','%d.%m.%Y %H.%i')); >>>> -ERROR 22007: Incorrect date value: '2004-09-31 15:30:00' for column 'col1' at row 1 >>>> +ERROR HY000: Incorrect datetime value: '31.9.2004 15.30' for function str_to_date >>> >>> Hmm, I don't know. I liked the old error message more. >>> Indeed, datetime value '31.9.2004 15.30' is just fine for a function str_to_date(). >>> But '2004-09-31 15:30:00' is not fine for column 'col1' in the strict mode. >>> >>> Why did that change? >> >> Before the change the warning was generated in >> Field_timestamp::store_TIME_with_warning. >> >> Now this warning is generated much earlier, >> in Item_func_str_to_date::get_date(), >> after check_date() is called. > > Which, I think, it wrong. The date is perfectly ok, as far as > STR_TO_DATE is concerted, it's wrong for a field. > >> The problem was that Item_func_str_to_date: >> a. checked TIME_NO_ZERO_DATE >> b. did not check TIME_NO_ZERO_IN_DATE (the bug) >> c. did not check invalid dates >> >> After the patch it: >> >> a. checks TIME_NO_ZERO_DATE >> b. checks TIME_NO_ZERO_IN_DATE >> c. checks invalid dates >> >> I thought it would be good to check everything, >> for consistency. But a side effect is a different warning. > > I agree, it should check everything. But, perhaps, the field should not > pass TIME_NO_ZERO_DATE|TIME_NO_ZERO_IN_DATE down to the arg->get_date() ? It does not. The above warnings are not about zeros. They are about wrong dates: for example, there is no September 31, like in '31.9.2004 15.30'. The last version of the patch does not change the warnings. It works as follows: - str_to_date checks TIME_NO_ZERO_DATE and TIME_NO_ZERO_IN_DATE with help of check_date(), but only if the caller wants it (CONVERT_TZ wants, field does not). - str_to_date suppresses checking of invalid dates inside check_date() and leaves this responsibility to the caller (like it was before the patch). If we want check invalid dates inside check_date() in str_to_date, then this warning "Incorrect date value: '2004-09-31 15:30:00' for column 'col1' at row 1" won't be possible, because str_to_date does not know the name of the column it's being inserted into. But we could try to produce something like this: "Incorrect datetime value: '2004-09-31 15:30:00'" I.e. a better formatted value, but without the column name. > > Regards, > Sergei >

1 0

[Maria-developers] Review request: new SHUTDOWN server command
by James Briggs 14 Jun '13

14 Jun '13

Hi Monty. Here's the shutdown command (WL#232) I wrote for review. I think there's 2 steps to improve server shutdown: 1) apply this patch now 2) later on, fix the old shutdown thread race problems, clean up the old kill_mysql() design and code, then add a few options to the shutdown command like WAIT|IMMEDIATE|ABORT. This code is released under the Open Source BSD-new License, according to the MariaDB Contributor Agreement. Thanks, James Briggs. james.briggs(a)yahoo.com http://www.jebriggs.com/blog/2013/05/patch-to-add-shutdown-statement-to-mys… --- sql_parse.cc 2013-03-11 03:29:13.000000000 -0700 +++ /home/james/mariadb-5.5.30-new/sql/sql_parse.cc 2013-05-15 13:17:05.000000000 -0700 @@ -1305,7 +1305,6 @@ my_ok(thd); break; } -#ifndef EMBEDDED_LIBRARY case COM_SHUTDOWN: { status_var_increment(thd->status_var.com_other); @@ -1333,7 +1332,6 @@ error=TRUE; break; } -#endif case COM_STATISTICS: { STATUS_VAR *current_global_status_var; // Big; Don't allocate on stack @@ -3736,6 +3734,31 @@ lex->kill_signal); break; } + case SQLCOM_SHUTDOWN: + { + // jeb - This code block is copied from COM_SHUTDOWN above. Since kill_mysql(void) {} doesn't take a level argument, the level code is pointless. + // jeb - In fact, the level code should be removed and Oracle Database statements implemented: SHUTDOWN, SHUTDOWN IMMEDIATE and SHUTDOWN ABORT. See WL#232. + + status_var_increment(thd->status_var.com_other); + if (check_global_access(thd,SHUTDOWN_ACL)) + break; /* purecov: inspected */ + + enum mysql_enum_shutdown_level level; + level= SHUTDOWN_DEFAULT; + if (level == SHUTDOWN_DEFAULT) + level= SHUTDOWN_WAIT_ALL_BUFFERS; // soon default will be configurable + else if (level != SHUTDOWN_WAIT_ALL_BUFFERS) + { + my_error(ER_NOT_SUPPORTED_YET, MYF(0), "this shutdown level"); + break; + } + DBUG_PRINT("SQLCOM_SHUTDOWN",("Got shutdown command for level %u", level)); + my_eof(thd); + kill_mysql(); + res=TRUE; + break; + } + #ifndef NO_EMBEDDED_ACCESS_CHECKS case SQLCOM_SHOW_GRANTS: { --- sql_yacc.yy 2013-03-11 03:29:19.000000000 -0700 +++ /home/james/mariadb-5.5.30-new/sql/sql_yacc.yy 2013-05-15 11:12:03.000000000 -0700 @@ -791,7 +791,7 @@ Currently there are 174 shift/reduce conflicts. We should not introduce new conflicts any more. */ -%expect 174 +%expect 196 /* Comments for TOKENS. @@ -1645,6 +1645,7 @@ definer_opt no_definer definer parse_vcol_expr vcol_opt_specifier vcol_opt_attribute vcol_opt_attribute_list vcol_attribute + shutdown END_OF_INPUT %type call sp_proc_stmts sp_proc_stmts1 sp_proc_stmt @@ -1796,6 +1797,7 @@ | savepoint | select | set + | shutdown | signal_stmt | show | slave @@ -13715,6 +13717,17 @@ ; +shutdown: + SHUTDOWN + { + LEX *lex=Lex; + lex->value_list.empty(); + lex->users_list.empty(); + lex->sql_command= SQLCOM_SHUTDOWN; + } + ; + + set_expr_or_default: expr { $$=$1; } | DEFAULT { $$=0; } --- sql_prepare.cc 2013-03-11 03:29:11.000000000 -0700 +++ /home/james/mariadb-5.5.30-new/sql/sql_prepare.cc 2013-05-15 03:07:00.000000000 -0700 @@ -2173,6 +2173,7 @@ case SQLCOM_GRANT: case SQLCOM_REVOKE: case SQLCOM_KILL: + case SQLCOM_SHUTDOWN: break; case SQLCOM_PREPARE: --- mysqld.cc 2013-03-11 03:29:14.000000000 -0700 +++ /home/james/mariadb-5.5.30-new/sql/mysqld.cc 2013-05-15 01:20:11.000000000 -0700 @@ -3333,6 +3333,7 @@ {"savepoint", (char*) offsetof(STATUS_VAR, com_stat[(uint) SQLCOM_SAVEPOINT]), SHOW_LONG_STATUS}, {"select", (char*) offsetof(STATUS_VAR, com_stat[(uint) SQLCOM_SELECT]), SHOW_LONG_STATUS}, {"set_option", (char*) offsetof(STATUS_VAR, com_stat[(uint) SQLCOM_SET_OPTION]), SHOW_LONG_STATUS}, + {"shutdown", (char*) offsetof(STATUS_VAR, com_stat[(uint) SQLCOM_SHUTDOWN]), SHOW_LONG_STATUS}, {"signal", (char*) offsetof(STATUS_VAR, com_stat[(uint) SQLCOM_SIGNAL]), SHOW_LONG_STATUS}, {"show_authors", (char*) offsetof(STATUS_VAR, com_stat[(uint) SQLCOM_SHOW_AUTHORS]), SHOW_LONG_STATUS}, {"show_binlog_events", (char*) offsetof(STATUS_VAR, com_stat[(uint) SQLCOM_SHOW_BINLOG_EVENTS]), SHOW_LONG_STATUS}, --- sql_lex.h 2013-03-11 03:29:13.000000000 -0700 +++ /home/james/mariadb-5.5.30-new/sql/sql_lex.h 2013-05-15 01:19:17.000000000 -0700 @@ -193,6 +193,7 @@ SQLCOM_SHOW_RELAYLOG_EVENTS, SQLCOM_SHOW_USER_STATS, SQLCOM_SHOW_TABLE_STATS, SQLCOM_SHOW_INDEX_STATS, SQLCOM_SHOW_CLIENT_STATS, + SQLCOM_SHUTDOWN, /*

1 0

Re: [Maria-developers] Review request: mdev-4635
by Alexander Barkov 13 Jun '13

13 Jun '13

Hi, Serg, On 06/13/2013 08:16 PM, Sergei Golubchik wrote: > Hi, Alexander! > > On Jun 13, Alexander Barkov wrote: >> >> please review a fix for MDEV-4635 > > Is it for 5.1 or 5.5? For 5.3. > >> === modified file 'mysql-test/r/func_time.result' >> --- mysql-test/r/func_time.result 2013-03-20 15:13:00 +0000 >> +++ mysql-test/r/func_time.result 2013-06-13 08:39:15 +0000 >> @@ -1910,3 +1910,13 @@ SELECT 1 FROM DUAL WHERE MINUTE(TIMEDIFF >> 1 >> SELECT 1 FROM DUAL WHERE SECOND(TIMEDIFF(NULL, '12:12:12')); >> 1 >> +# >> +# MDEV-4635 Crash in UNIX_TIMESTAMP(STR_TO_DATE('2020','%Y')) >> +# >> +SET TIME_ZONE='+02:00'; >> +SELECT UNIX_TIMESTAMP(STR_TO_DATE('2020','%Y')); >> +UNIX_TIMESTAMP(STR_TO_DATE('2020','%Y')) >> +NULL >> +Warnings: >> +Error 1411 Incorrect datetime value: '2020' for function str_to_date >> +SET TIME_ZONE=DEFAULT; >> >> === modified file 'mysql-test/r/strict.result' >> --- mysql-test/r/strict.result 2011-06-05 02:56:06 +0000 >> +++ mysql-test/r/strict.result 2013-06-13 12:15:43 +0000 >> @@ -213,11 +213,11 @@ ERROR 22007: Incorrect date value: '2004 >> INSERT INTO t1 (col1) VALUES(STR_TO_DATE('0.10.2004 15.30','%d.%m.%Y %H.%i')); >> ERROR 22007: Incorrect date value: '2004-10-00 15:30:00' for column 'col1' at row 1 >> INSERT INTO t1 (col1) VALUES(STR_TO_DATE('31.9.2004 15.30','%d.%m.%Y %H.%i')); >> -ERROR 22007: Incorrect date value: '2004-09-31 15:30:00' for column 'col1' at row 1 >> +ERROR HY000: Incorrect datetime value: '31.9.2004 15.30' for function str_to_date > > Hmm, I don't know. I liked the old error message more. > Indeed, datetime value '31.9.2004 15.30' is just fine for a function str_to_date(). > But '2004-09-31 15:30:00' is not fine for column 'col1' in the strict mode. > > Why did that change? Before the change the warning was generated in Field_timestamp::store_TIME_with_warning. Now this warning is generated much earlier, in Item_func_str_to_date::get_date(), after check_date() is called. The problem was that Item_func_str_to_date: a. checked TIME_NO_ZERO_DATE b. did not check TIME_NO_ZERO_IN_DATE (the bug) c. did not check invalid dates After the patch it: a. checks TIME_NO_ZERO_DATE b. checks TIME_NO_ZERO_IN_DATE c. checks invalid dates I thought it would be good to check everything, for consistency. But a side effect is a different warning. But we can suppress invalid dates detection, so it: a. checks TIME_NO_ZERO_DATE b. checks TIME_NO_ZERO_IN_DATE c. does not check invalid dates The test for invalid dates can be suppressed by passing "fuzzy_date | TIME_INVALID_DATES" to check_date(). See this version of the patch attached. I agree that the old warning is nicer. But I'm not sure what's better. Suppressing the test rather slightly looks like hiding more potential bugs... > >> === modified file 'sql/item_timefunc.cc' >> --- sql/item_timefunc.cc 2013-03-20 15:13:00 +0000 >> +++ sql/item_timefunc.cc 2013-06-13 08:20:13 +0000 >> @@ -1126,7 +1125,7 @@ bool Item_func_unix_timestamp::get_times >> } >> >> MYSQL_TIME ltime; >> - if (get_arg0_date(&ltime, 0)) >> + if (get_arg0_date(&ltime, TIME_NO_ZERO_IN_DATE)) >> return 1; > > This is correct, of course. > > Regards, > Sergei >

1 0

[Maria-developers] Review request: mdev-4651
by Alexander Barkov 13 Jun '13

13 Jun '13

Hi Serg, Please review a patch for "mdev-4651 Crash in my_decimal2decimal in a ORDER BY query". Thanks.

1 0

[Maria-developers] Review request: mdev-4635
by Alexander Barkov 13 Jun '13

13 Jun '13

Hi Serg, please review a fix for MDEV-4635 Thanks.

1 0

Re: [Maria-developers] Review request: mdev-4634
by Alexander Barkov 13 Jun '13

13 Jun '13

Hi Sergei, Thanks for review! On 06/11/2013 09:33 PM, Sergei Golubchik wrote: > Hi, Alexander! > > This looks like a lot of changes for what you explained the reason of > the bug was. There will be lots of "whys" below... The fix itself was quite small. However, an attempt to write a reasonable test case revealed a few other bugs. I just tried to fixing everything in a single patch. Okey. Let's do everything separately. These are separate reports for the problems found during writing a test case for 4634: https://mariadb.atlassian.net/browse/MDEV-4652 https://mariadb.atlassian.net/browse/MDEV-4653 https://mariadb.atlassian.net/browse/MDEV-4654 Here's a new one, found today: https://mariadb.atlassian.net/browse/MDEV-4651 > > On Jun 11, Alexander Barkov wrote: >> === modified file 'sql/item.cc' >> --- sql/item.cc 2013-03-17 06:41:22 +0000 >> +++ sql/item.cc 2013-06-11 11:52:43 +0000 >> @@ -252,7 +252,8 @@ String *Item::val_string_from_decimal(St >> String *Item::val_string_from_date(String *str) >> { >> MYSQL_TIME ltime; >> - if (get_date(&ltime, TIME_FUZZY_DATE) || >> + uint time_flag= field_type() == MYSQL_TYPE_TIME ? TIME_TIME_ONLY : 0; >> + if (get_date(&ltime, TIME_FUZZY_DATE | time_flag) || > > Why? > >> str->alloc(MAX_DATE_STRING_REP_LENGTH)) >> { >> null_value= 1; >> >> === modified file 'sql/item_func.cc' >> --- sql/item_func.cc 2013-04-06 13:14:46 +0000 >> +++ sql/item_func.cc 2013-06-11 10:54:50 +0000 >> @@ -2456,7 +2456,8 @@ bool Item_func_min_max::get_date(MYSQL_T >> { >> Item **arg= args + i; >> bool is_null; >> - longlong res= get_datetime_value(thd, &arg, 0, compare_as_dates, &is_null); >> + longlong res= get_datetime_value(thd, &arg, 0, compare_as_dates, >> + &is_null, fuzzy_date); > > why is that? to handle CONVERT_TZ(GREATEST(...)) ? > That is to pass NO_ZERO_DATE down? > > I'm not sure it's a good idea. Strictly speaking, only the greatest > value should be NO_ZERO_DATE, not every single one. > > I'd suggest to add check_date in Item_func_min_max::get_date() instead. > >> /* Check if we need to stop (because of error or KILL) and stop the loop */ >> if (thd->is_error() || args[i]->null_value) >> @@ -2469,6 +2470,11 @@ bool Item_func_min_max::get_date(MYSQL_T >> min_max= res; >> } >> unpack_time(min_max, ltime); >> + >> + if (!(fuzzy_date & TIME_TIME_ONLY) && >> + ((null_value= check_date_with_warn(ltime, fuzzy_date)))) >> + return true; > > Yes, that's what I mean. If you check the return value anyway, there's > no need to check it inside get_datetime_value(). > >> if (compare_as_dates->field_type() == MYSQL_TYPE_DATE) >> { >> ltime->time_type= MYSQL_TIMESTAMP_DATE; >> >> === modified file 'sql/item_timefunc.cc' >> --- sql/item_timefunc.cc 2013-03-20 15:13:00 +0000 >> +++ sql/item_timefunc.cc 2013-06-11 10:57:09 +0000 >> @@ -1923,8 +1921,10 @@ void Item_date_add_interval::fix_length_ >> bool Item_date_add_interval::get_date(MYSQL_TIME *ltime, uint fuzzy_date) >> { >> INTERVAL interval; >> - >> - if (args[0]->get_date(ltime, TIME_NO_ZERO_DATE | TIME_FUZZY_DATE | TIME_NO_ZERO_IN_DATE) || >> + if (args[0]->get_date(ltime, >> + TIME_NO_ZERO_DATE | TIME_FUZZY_DATE | TIME_NO_ZERO_IN_DATE | >> + ((args[0]->field_type() == MYSQL_TYPE_TIME) ? >> + TIME_TIME_ONLY : 0)) || > > why is that? > >> get_interval_value(args[1], int_type, &value, &interval)) >> return (null_value=1); >> >> @@ -2290,6 +2290,12 @@ bool Item_time_typecast::get_date(MYSQL_ >> return 1; >> if (decimals < TIME_SECOND_PART_DIGITS) >> ltime->second_part= sec_part_truncate(ltime->second_part, decimals); >> + if (!(fuzzy_date & TIME_TIME_ONLY)) >> + { >> + if (ltime->time_type == MYSQL_TIMESTAMP_TIME) >> + ltime->time_type= MYSQL_TIMESTAMP_DATETIME; > > why? > >> + return ((null_value= check_date_with_warn(ltime, fuzzy_date))); >> + } >> /* >> MYSQL_TIMESTAMP_TIME value can have non-zero day part, >> which we should not lose. >> === modified file 'sql/item_timefunc.h' >> --- sql/item_timefunc.h 2012-12-28 12:41:46 +0000 >> +++ sql/item_timefunc.h 2013-06-11 11:19:03 +0000 >> @@ -492,6 +492,8 @@ class Item_timefunc :public Item_tempora >> Item_timefunc(Item *a,Item *b) :Item_temporal_func(a,b) {} >> Item_timefunc(Item *a, Item *b, Item *c) :Item_temporal_func(a, b ,c) {} >> enum_field_types field_type() const { return MYSQL_TYPE_TIME; } >> + int save_in_field(Field *field, bool no_conversions) >> + { return save_time_in_field(field); } > > why? > >> }; >> >> === modified file 'sql/time.cc' >> --- sql/time.cc 2013-01-08 20:23:03 +0000 >> +++ sql/time.cc 2013-06-11 10:51:31 +0000 >> @@ -213,6 +213,20 @@ ulong convert_month_to_period(ulong mont >> } >> >> >> +bool >> +check_date_with_warn(const MYSQL_TIME *ltime, uint fuzzy_date) >> +{ >> + int dummy_warnings; >> + if (check_date(ltime, fuzzy_date, &dummy_warnings)) >> + { >> + Lazy_string_time str(ltime); >> + make_truncated_value_warning(current_thd, MYSQL_ERROR::WARN_LEVEL_WARN, >> + &str, MYSQL_TIMESTAMP_ERROR, 0); > > please, use this function also in Item_date_typecast::get_date() and > may be in other places. you might need to pass MYSQL_TIMESTAMP_xxx > value as an argument, though. > >> + return true; >> + } >> + return false; >> +} >> + >> /* >> Convert a timestamp string to a MYSQL_TIME value and produce a warning >> if string was truncated during conversion. >> > > Regards, > Sergei >

1 0

[Maria-developers] I_S plugin doubts
by Roberto Spadim 13 Jun '13

13 Jun '13

Hi i was reading locale_info and qc_info source that i sent in jira changing the last parameter from 0 to SKIP_OPEN_TABLE, change the table from aria engine to memory engine, why? this change something internally in I_S table ? example: from static ST_FIELD_INFO qc_info_fields_queries[]= { {"ENTRY_POSITION_IN_CACHE", MY_INT32_NUM_DECIMAL_DIGITS, MYSQL_TYPE_LONG, 0, 0, 0, 0}, ... to static ST_FIELD_INFO qc_info_fields_queries[]= { {"ENTRY_POSITION_IN_CACHE", MY_INT32_NUM_DECIMAL_DIGITS, MYSQL_TYPE_LONG, 0, 0, 0, SKIP_OPEN_TABLE}, ... -- Roberto Spadim

2 5

Re: [Maria-developers] MDEV-4581 , more query cache information fields
by Roberto Spadim 12 Jun '13

12 Jun '13

hi sergei, i sent the patch, could you check it? i think that's easy to read and understand maybe some code standard should be done since i don't know how code in mariadb cs https://mariadb.atlassian.net/browse/MDEV-4581

1 0

[Maria-developers] MDEV-4581 , more query cache information fields
by Roberto Spadim 12 Jun '13

12 Jun '13

Hi guys, i sent a patch for qc_info plugin in MDEV-4581 Could mariadb team review and tell me what i should do, to put it in the next release? i will rewrite it, at 10.0.3 release, and change some things: the add to query to cache, i will change to only a function call at sql_cache.cc, and put the function at sql_cache.h something like: qc->qc_info_set_informations(insert time, rows_sent, rows_read, etc etc etc) rewrite the part of query hit. to execute only a command too qc->qc_info_hit(current_time, expend_qc_time, etc etc etc) this will make sql_cache.cc code cleaner and sql_cache.h "dirty" but will be more easly to read and patch future versions. only add 2 lines to sql_cache.cc , and many lines to sql_cache.h since sql_cache.h have the structure it will be easly to write function call at struct than sql_cache.cc file bye guys -- Roberto Spadim

1 0

Re: [Maria-developers] audit plugin help with MDEV-575 (fail2ban like)
by Roberto Spadim 12 Jun '13

12 Jun '13

well i'm thinking about again, maybe just a syslog could do the job with fail2ban but in windows... well it will not work maybe the second option of a internal table could do the job too Host | bad_login_datetime if host have more than 'bad_login_tries' bad logins in 'bad_login_seconds' seconds before now it will be disconnect after xxx seconds or after each minute, or after each disconnect clean the internal table that have bad_login_datetime < now() - 'bad_login_seconds' the clean should be optimized in a not paintfull way (maybe like a garbage colector that runs after X disconnects, or a event that happens each x seconds, or both of them, i prefer the disconnect method) any other idea?

1 2

Re: [Maria-developers] [Commits] Rev 3551: MDEV-4612 SQ pushdown: Server crashes in make_join_statistics with materialization+semijoin, IN subqueries, constant table, impossible condition in file:///home/tsk/mprog/src/10.0-md83/
by Oleksandr Byelkin 11 Jun '13

11 Jun '13

Hi! 11.06.2013 11:11, timour(a)askmonty.org пишет: > At file:///home/tsk/mprog/src/10.0-md83/ > > ------------------------------------------------------------ > revno: 3551 > revision-id: timour(a)askmonty.org-20130611081135-rge7p00h8e9nhypj > parent: timour(a)askmonty.org-20130528122232-kisslvdw5vhrna0q > fixes bug: https://mariadb.atlassian.net/browse/MDEV-4612 > committer: timour(a)askmonty.org > branch nick: 10.0-md83 > timestamp: Tue 2013-06-11 11:11:35 +0300 > message: > MDEV-4612 SQ pushdown: Server crashes in make_join_statistics with materialization+semijoin, IN subqueries, constant table, impossible condition > > Analysis: > The subquery materialization execution method subselect_hash_sj_engine::exec() > didn't restore thd->lex->current_select through all return paths. This was > not a problem when the method is called during execution. When a subquery > was executed during the optimization phase of the outer query, current_select > remained to point to the wrong subquery. This resulted in update_ref_and_keys() > getting the wrong number of conditions via thd->lex->current_select->cond_count, > thus allocating a smaller buffer for sargable conditions. This later resulted > in memory overwrite. > > Solution: > Make sure that subselect_hash_sj_engine::exec() always restores current_select. The patch looks OK to push. [skip]

1 0

[Maria-developers] Group commit comparison between mysql and mariadb
by Sergei Petrunia 11 Jun '13

11 Jun '13

Hi Kristian, I'm wondering, do we have any comparison of MariaDB's binlog group commit implementation vs MySQL's binlog's group commit? I recall Mats writing a blog about the differences last year, but he didn't have the code back then. I'm wondering if Oracle had implemented what he described, or did something else... I understand that in order to do benchmark we need to finish merging InnoDB from MySQL 5.6 into MariaDB 10.0. One can't benchmark group commit per se, it is tied to innodb. Do we still have any clue, before the merge happens? Btw, what should the merge do - take group commit from MariaDB or from MySQL? BR Sergei -- Sergei Petrunia, Software Developer Monty Program AB, http://askmonty.org Blog: http://s.petrunia.net/blog

4 3

Re: [Maria-developers] Buildbot question
by Kristian Nielsen 11 Jun '13

11 Jun '13

Sergey Vojtovich <svoj(a)mariadb.org> writes: > I'm observing pretty odd buildbot behavior with some of my pushes (probably > merges). Bar and Daniel were unable to answer my questions and suggested to > ask you instead. The problem seems to be that Buildbot has become confused about revisions. Look at the top of the page you linked, at the revisions for each column. You will notice that they are not in order. The 3rd column is the highest revision, which is probably why spider suite is enabled there (I assume you pushed something to add the suite). I see your current top revision in the tree is 3760, while the Buildbot page has a newer revision 3767. It also looks like the Buildbot history contains _another_ earlier revision that was also number 3760, maybe that is the cause of the confusion (Buildbot perhaps thinks it is rebuilding the old 3760). You can see your revision sorted as the first columns in this link: http://buildbot.askmonty.org/buildbot/grid?branch=10.0-spider&width=20 This confusion usually happens because the tree was overwritten and revisions became renumbered. This can happen due to bzr push --overwrite, but in your case it is probably because you merged in the opposite direction. When merging a tree A with a tree B and pushing to B, it is generally best to merge A into B (cd to B and bzr merge A). This keeps the existing revision numbers of B intact. In the main trees we have protection that enforces this, but not in feature trees. (I would be nice if we could fix Buildbot to prevent this confusion, but so far noone has taken it upon themselves. At one point I asked the Launchpad/bzr developers to make the revision id (as opposed to revision numbers) available so that Buildbot could avoid using revision numbers completely, but the answer was that "hiding revision ids was a feature" ...). Usually the confusion is sorted out by itself on a following push, or you can use the trick above adding width=N to the link to see the column of the revision you want. Hope this helps, - Kristian.

2 1

[Maria-developers] MariaDB 10.0.3 release prep
by Daniel Bartholomew 11 Jun '13

11 Jun '13

All, I've begun the prep work for the MariaDB 10.0.3 alpha release. The packages are uploaded and the mirrors should pick them up soon (if they haven't already). The draft changelog and release notes are here: - https://kb.askmonty.org/en/mariadb-1003-release-notes/ - https://kb.askmonty.org/en/mariadb-1003-changelog/ Please let me know ASAP if there's anything that needs to be mentioned/highlighted/removed/etc... Assuming all goes well, the plan is to activate the release tomorrow (11 Jun 2013) at around 10am US Eastern Time. Thanks. -- Daniel Bartholomew Google+ - http://gplus.to/dbart Twitter - http://twitter.com/daniel_bart MariaDB: An Enhanced Drop-in Replacement for MySQL Website - http://mariadb.org Twitter - http://twitter.com/mariadb Google+ - http://google.com/+mariadb Facebook - http://fb.com/MariaDB.dbms Knowledgebase - http://kb.askmonty.org

1 0

[Maria-developers] audit plugin help with MDEV-575 (fail2ban like)
by Roberto Spadim 09 Jun '13

09 Jun '13

hi guys i want get users / ip that try to connect to server with bad password/user and retry very frequently (normally a brute force attack) some things i need: 1)get information about connection when it fail or not (maybe an audit plugin could do this?), information that i need: 1.1)protocol (tcp/unix/others) 1.2)user 1.3)remote ip (when tcp only) it must be executed in every login 2) log this information maybe we could use CSV,aria, myisam, to log bad logins and put it in a log table: PROTOCOL, USER, REMOTE IP, DATATIME 3) add columns to user BAD_LOGIN_MAX_TRIES, BAD_LOGIN_LOG_TIME, BAD_LOGIN_MIN_RETRY_TIME 4) add server global bad_login_tcp_log_time, bad_login_tcp_max_errors 5) implement checks before connection: when openned a connection to server via TCP execute something like: SELECT COUNT(*) FROM bad_login_log WHERE remote_ip="remote_ip" AND datetime>NOW() - @@bad_login_tcp_log_time if COUNT(*) > @@bad_login_tcp_max_erros disconnect 6) implement user checks: SELECT COUNT(*),(SELECT BAD_LOGIN_MAX_TRIES FROM users WHERE user="user") FROM bad_login_log WHERE user="user" AND datetime>NOW() - (SELECT BAD_LOGIN_LOG_TIME FROM users WHERE user="user") if count(*) > bad_login_max_tries form that user disconnect that's all -- Roberto Spadim SPAEmpresarial

1 0

[Maria-developers] Help with MDEV-4419, ENUM optimization
by Roberto Spadim 09 Jun '13

09 Jun '13

Hi guys, i want to optimize fields with ENUM() in item comparation the point is: SELECT COUNT(*) FROM table WHERE enum_field!='x' column 'enum_field' is ENUM('x,'y','z') NOT NULL DEFAULT 'x' i want to rewrite the query in a way that we can execute it with index: SELECT COUNT(*)FROM table WHERE enum_field IN ('y','z','') the "" is not allowed values ('a' for example) the feature is 'easy' to understand but i don't know where to implement it will read possible values of enum fields, and rewrite "!=","<","<=",">=",">" operators to IN operator with fields that match the condition of the primary operator some think like: column "operator" const value if(operator in ('!=','<','<=','>=','>' AND column is ENUM){ get all possible values from column copy all values to a return variable execute the "operator" in all values with const values if operator return false remove the value from the return variable change the "operator" and "const value" to : "IN" and return variable comma separated }else{ no optimization leave operation as it is } example: column = enum('x','y','z'), operator = ">" const value = "3" if(column is ENUM and operator = ">"){ possible values => "x","y","z","" (check that "" is used when insert into, value = "a" for example) return variable = "x","y","z", "" for each value.... "x" > "3" ? yes, ASCII 120 > ASCII 51 -> do nothing "y" > "3" ? yes, ASCII 121 > ASCII 51 -> do nothing "z" > "3" ? yes, ASCII 122 > ASCII 51 -> do nothing "" > "3" ? no, "" < ASCII 51 -> remove the value "" from return variable end for now the return variable is : "x","y","z" the "" was removed rewrite the operator to: "column IN ('x','y','z')" } ----------------------- this will optimize a lot: using != in a big table (446577 rows) result = Using where, 446577 rows, type= ALL using IN() in a big table: result = Using where, 28 rows, key = index, type = range in other words, a lot of optimization with ENUM fields, since they are a pseudo "index" of what we can found in the column =) thank guys any help is wellcome to make a patch -- Roberto Spadim SPAEmpresarial

1 0

[Maria-developers] GSoC MariaDB Kerberos Authentication Plugin Proposal
by QIU Shuang 06 Jun '13

06 Jun '13

Hi Sergei, Thank you and the community for accepting my proposal of this year's GSoC: Kerberize MariaDB -- add kerberos authentication support to MariaDB. I'm Shuang Qiu, a second year graduate student from Shanghai Jiao Tong University, Shanghai, China. My field of research is distributed computing, especially distributed storage. I just get in touch with my project mentor Mr. Vaintroub and join the #maria IRC channel at freenode.net, where I can be found as shuang. Thank you again and hope we can have a happy and productive summer! Cheers! Sincerely, Shuang

1 0

Re: [Maria-developers] [Maria-discuss] Clear-text plugin
by Peter Laursen 06 Jun '13

06 Jun '13

"no support for dlopen()" . How does this apply to Windows clients? We always build first on Windows and once we have no errors of any kind we build on Linux if the application exists for that platform (what MONyog does and SQLyog does not) and fix what issues may occur there. This issue has occurred with SQLyog (and as such for a Windows client). I am not coder/developer myself, but on Windows I think we have no other option than statically linking? We could of could compile the API as a .dll, but we don't (because we had a lot of problems with it earlier). -- Peter On Thu, Jun 6, 2013 at 3:25 PM, Sergei Golubchik <serg(a)mariadb.org> wrote: > Hi, Peter! > > On Jun 04, Peter Laursen wrote: > > We have the same problem. We cannot test either. But we have a few user > > reports. We will discuss what we can do. > > > > Is there any parameter or compile-time option we should be aware of? Any > > specific documentation somewhere? Anything? > > Not really. If the cleartext plugin is built and installed, it'll be > used when the server requests it. > > I've just checked that this plugin is present both in rpm and deb > packages. I didn't check binary tarballs, but it should be there too. > > So, there's nothing to configure at run-time or compile-time, everything > should work automatically and transparently. > > Hmm, unless you build a completely static client library with no support > for dlopen(). > > Regards, > Sergei >

2 1

Re: [Maria-developers] [Maria-discuss] Clear-text plugin
by Peter Laursen 04 Jun '13

04 Jun '13

We have the same problem. We cannot test either. But we have a few user reports. We will discuss what we can do. Is there any parameter or compile-time option we should be aware of? Any specific documentation somewhere? Anything? -- Peter On Tue, Jun 4, 2013 at 2:19 PM, Sergei Golubchik <serg(a)mariadb.org> wrote: > Hi, Peter! > > On Jun 04, Peter Laursen wrote: > > For clarity: a 'dummy' parameter doing nothing does not solve our > > problem. We actually want to build clients that can (also) connect to > > Oracle MySQL Enterprise servers configured to make use of Oracle's PAM > > implementation. > > We build and distribute a cleartext plugin. And it can be used if the > server requests. In a sense, in MariaDB a cleartext plugin is always > enabled. So a dummy option is needed only to restore the binary protocol > compatibility. That's my understanding, at least, but I don't have > Oracle's PAM plugin, so I cannot test it. > > But if you have verified that your client with our C connector can not > connect to Oracle MySQL Enterprise server with the Oracle's PAM > implementation - please submit a bug report, and we will fix it. > > Regards, > Sergei > >

1 0

Re: [Maria-developers] [Maria-discuss] Clear-text plugin
by Peter Laursen 04 Jun '13

04 Jun '13

For clarity: a 'dummy' parameter doing nothing does not solve our problem. We actually want to build clients that can (also) connect to Oracle MySQL Enterprise servers configured to make use of Oracle's PAM implementation. -- Peter On Tue, Jun 4, 2013 at 11:05 AM, Peter Laursen <peter_laursen(a)webyog.com>wrote: > yes .. in the Connector/C. We use C/C++ for both SQLyog and MONyog. > > -- Peter > > > On Tue, Jun 4, 2013 at 10:17 AM, Sergei Golubchik <serg(a)mariadb.org>wrote: > >> Hi, Peter! >> >> On Jun 04, Peter Laursen wrote: >> > This is a continuation of >> > https://mariadb.atlassian.net/browse/CONC-25. We requested support for >> > the clear-text plugin (or a workaround of any kind) that would allow a >> > client compiled with the MariaDB connector to use PAM authentication >> > with MySQL Enterprise servers from Oracle. >> >> Another related issue is https://mariadb.atlassian.net/browse/MDEV-4574 >> which should be fixed in 5.5.32, next 5.5 release. >> >> But you, probably, want it in Connector/C, don't you? >> >> Regards, >> Sergei >> >> >

1 0

[Maria-developers] Clear-text plugin
by Peter Laursen 04 Jun '13

04 Jun '13

This is a continuation of https://mariadb.atlassian.net/browse/CONC-25. We requested support for the clear-text plugin (or a workaround of any kind) that would allow a client compiled with the MariaDB connector to use PAM authentication with MySQL Enterprise servers from Oracle. I will not dispute Sergei's opinions about Oracle's implementation. But I believe it misses the point. The question is if the MariaDB connector should be a replacement for the C-IPA from Oracle. And the question is nothing else than that. Not the quality of solutions bundled with Enterprise servers from Oracle. I would like to hear opinions not only from 'core developers' like Sergei, but also from wider circles in MariaDB organization (management, marketing) as well as the user community. For us the problem is like this: We introduced support for the clear-text plugin in SQLyog 11.1. In 11.12 beta2 we replaced Oracles connector with MariaDB (and had to removed it again). We had a few crashes and memory leaks with mariaDB Connector but that is getting in control now (ther were bugs in both the API and in oru code). We probably could patch the MariaDB code without much problems to support the full-text plugin - but we don't have a MySQL Enterprise server to test with and we cannot release without proper testing. We have already started getting complaints from users/customers about that what we introduced in 11.1 has been removed in current development tree. It seems that we have quite a lot of customers using Oracle's Enterprise servers wiht Oracle's PAM authentication. A solution/workaround would very much be appreciated. Regards Peter Laursen Webyog

1 0

Re: [Maria-developers] MDEV-4568 - Port Percona response time distribution as audit plugin
by Sergey Vojtovich 03 Jun '13

03 Jun '13

Hi Sergei, 03.06.2013, в 16:36, Sergei Golubchik <serg(a)mariadb.org> написал(а): > Hi, Sergey! > > On May 31, Sergey Vojtovich wrote: >> revno: 3753 >> revision-id: svoj(a)mariadb.org-20130531131833-zs83x5wntvl1i15f >> parent: wlad(a)montyprogram.com-20130523125248-22m56m01r24t0dmb >> committer: Sergey Vojtovich <svoj(a)mariadb.org> >> branch nick: maria-qrtp >> timestamp: Fri 2013-05-31 17:18:33 +0400 >> message: >> MDEV-4568 - Port Percona response time distribution as audit plugin >> >> sql/sp_head.cc: >> Trigger MYSQL_AUDIT_GENERAL_STATUS event for each statement in stored >> procedures. Not strictly needed for query_response_time plugin, but >> makes it behave more like Percona patch. > > I'm sorry, I couldn't quite understand what you have changed, as > compared to the first patch (and the first was not in emails, so I > couldn't compare them). The difference is rather minor: - renamed query_response_time_exec_time to query_response_time_exec_time_debug - updated plugin author and added GPL header to plugin.cc - fixed I_S plugin description What I didn't touch were query_response_time.cc and .h, but below you agree that we shouldn't touch them. > Looks pretty much the same as the previous one, although you need to run > the tests and to see that they pass (you've changed plugin description > in the code, but the test result wasn't changed accordingly). Hmm… you're right, somehow I forgot to update basic test. Other tests are up to date. > Otherwise, looks ok to push, I suppose we can let it be as close to > Percona's version as possible. I don't think you need to rewrite it, > it's in a plugin now anyway. > > Regards, > Sergei Thanks, Sergey

1 0

[Maria-developers] Partition prune and lock
by Roberto Spadim 02 Jun '13

02 Jun '13

Hi guys i'm studing mariadb code and reading user manuals to know more about features. My last study was about query cache with many ideas in some MDEV and some patchs sent Now i'm reading about partition My questions are 1)when a update or a delete is done on a expecific partition like: Update table set field=2 where partition=1 How lock is done? Lock is done at table level or partition level? 2)Today we have vertical partition, could we implement 'virtual' horizontal partition update and delete in some group of fields (horizontal partition) only lock the horizontal partition being changed and used in where clausule? Any idea and problems on how to develop it? 3) can query cache be partition pruned? instead of removing entire table, just remove the partition that was changed It will consume more memory but will do a better work in cache Maybe a new key like SQL_CACHE_PARTITON or a new variable, query_cache_min_unit=PARTITION or TABLE could change how the query will be cached and pruned Thanks guys

1 1

Re: [Maria-developers] [codership-team] Hanging Galera Cluster, looking for the cause
by Jan Kirchhoff 01 Jun '13

01 Jun '13

I have seen exactly the same issue again, second day in a row. This time at 7:30 in the morning and again, no logs or anything that would give me an idea of what could have caused this. "killall -9 mysqld" and "service mysql start" on the hanging node is the only solution I have so far... I am crossposting this to maria-developers, maybe somebody on that list has an idea where to start? Jan Am 31.05.2013 10:29, schrieb Jan Kirchhoff: > Hi, > > we have been testing MariaDB 5.5.29 with Galera > (wsrep_provider_version 23.2.4(r147)) for a few months now and have had > a freeze of the cluster for the second time now. I didn't report the > first hang about 6 weeks ago as I wanted to be sure it wasn't our fault, > but I can't find any type of network issue or such in our monitoring > system and nothing in the logs that could be related to this problem: > > We have 2 database servers running Ubuntu 12.10 with Kernel 3.5.0 > (names: db8, db9), with a bunch of 15k-disks, some SSDs, 128GB RAM and a > test database of about 600-700GB. The amount of binlogs written every > day is around 500GB, so you get an idea of the workload. > db8 is a slave (traditional replication) of our main database. We have a > garbd running on another server as a third galera instance. > > Everything runs smooth for some weeks, but since this morning at 3:50am > db9 seems to be frozen. The server itself looks good, mysqld is taking > about 50% of one CPU core, everything else looks normal. Nothing in the > logs. Connecting to mysql works, but any query (like "show status" or > normal selects) hang forever. Only "show engine innodb status" returns. > There are almost 3000 processes that piled up like this: > > ---TRANSACTION 2F89D7CB, not started > MySQL thread id 113575, OS thread handle 0x7fab31821700, query id > 45249946 cl4.logical-line.de 192.168.0.16 stats_pflege sleeping > > but none of them ever gets any further. I can issue a "kill 113575" and > the prompt returns, but nothing happens. > > This is what db8's global status says: > wsrep_local_state_uuid 41d7259d-7200-11e2-0800-943dd0e2f659 > wsrep_protocol_version 4 > wsrep_last_committed 179518511 > wsrep_replicated 53590833 > wsrep_replicated_bytes 14768686288759 > wsrep_received 894995 > wsrep_received_bytes 7161712 > wsrep_local_commits 53583808 > wsrep_local_cert_failures 60 > wsrep_local_bf_aborts 0 > wsrep_local_replays 0 > wsrep_local_send_queue 8 > wsrep_local_send_queue_avg 0.000000 > wsrep_local_recv_queue 0 > wsrep_local_recv_queue_avg 0.000000 > wsrep_flow_control_paused 1.000000 > wsrep_flow_control_sent 0 > wsrep_flow_control_recv 0 > wsrep_cert_deps_distance 73.229730 > wsrep_apply_oooe 0.000000 > wsrep_apply_oool 0.000000 > wsrep_apply_window 0.000000 > wsrep_commit_oooe 0.000000 > wsrep_commit_oool 0.000000 > wsrep_commit_window 0.000000 > wsrep_local_state 4 > wsrep_local_state_comment Synced > wsrep_cert_index_size 25606 > wsrep_causal_reads 0 > wsrep_incoming_addresses 192.168.10.30:3306,192.168.10.31:3306, > wsrep_cluster_conf_id 15 > wsrep_cluster_size 3 > wsrep_cluster_state_uuid 41d7259d-7200-11e2-0800-943dd0e2f659 > wsrep_cluster_status Primary > wsrep_connected ON > wsrep_local_index 0 > wsrep_provider_name Galera > wsrep_provider_vendor Codership Oy <info(a)codership.com> > wsrep_provider_version 23.2.4(r147) > wsrep_ready ON > > > > > Here is the innodb status from db9 (the one that hangs). This is the > only output I could get out of db9 at all: > > Type Name Status > InnoDB > ===================================== > 130531 10:02:08 INNODB MONITOR OUTPUT > ===================================== > Per second averages calculated from the last 18 seconds > ----------------- > BACKGROUND THREAD > ----------------- > srv_master_thread loops: 907033 1_second, 906789 sleeps, 89831 > 10_second, 22491 background, 22489 flush > srv_master_thread log flush and writes: 478222 > ---------- > SEMAPHORES > ---------- > OS WAIT ARRAY INFO: reservation count 17017782, signal count 287266881 > Mutex spin waits 98936299, rounds 91970366, OS waits 868715 > RW-shared spins 139527705, rounds 455067309, OS waits 6488041 > RW-excl spins 58122849, rounds 842712424, OS waits 8565312 > Spin rounds per wait: 0.93 mutex, 3.26 RW-shared, 14.50 RW-excl > -------- > FILE I/O > -------- > I/O thread 0 state: waiting for completed aio requests (insert buffer > thread) > I/O thread 1 state: waiting for completed aio requests (log thread) > I/O thread 2 state: waiting for completed aio requests (read thread) > I/O thread 3 state: waiting for completed aio requests (read thread) > I/O thread 4 state: waiting for completed aio requests (read thread) > I/O thread 5 state: waiting for completed aio requests (read thread) > I/O thread 6 state: waiting for completed aio requests (write thread) > I/O thread 7 state: waiting for completed aio requests (write thread) > I/O thread 8 state: waiting for completed aio requests (write thread) > I/O thread 9 state: waiting for completed aio requests (write thread) > Pending normal aio reads: 0 [0, 0, 0, 0] , aio writes: 0 [0, 0, 0, 0] , > ibuf aio reads: 0, log i/o's: 0, sync i/o's: 0 > Pending flushes (fsync) log: 0; buffer pool: 0 > 49405970 OS file reads, 277420410 OS file writes, 6653801 OS fsyncs > 1 pending preads, 0 pending pwrites > 49.44 reads/s, 16384 avg bytes/read, 0.61 writes/s, 0.61 fsyncs/s > ------------------------------------- > INSERT BUFFER AND ADAPTIVE HASH INDEX > ------------------------------------- > Ibuf: size 1, free list len 698, seg size 700, 300226 merges > merged operations: > insert 410879, delete mark 8032461, delete 82075 > discarded operations: > insert 0, delete mark 0, delete 0 > Hash table size 212497717, node heap has 450852 buffer(s) > 16846.79 hash searches/s, 77045.28 non-hash searches/s > --- > LOG > --- > Log sequence number 18522888544071 > Log flushed up to 18522888544071 > Last checkpoint at 18522888544051 > Max checkpoint age 2035833201 > Checkpoint age target 1972213414 > Modified age 20 > Checkpoint age 20 > 0 pending log writes, 0 pending chkp writes > 6895012 log i/o's done, 0.39 log i/o's/second > ---------------------- > BUFFER POOL AND MEMORY > ---------------------- > Total memory allocated 110310195200; in additional pool allocated 0 > Internal hash tables (constant factor + variable factor) > Adaptive hash index 9086745120 \t(1699981736 + 7386763384) > Page hash 106249768 (buffer pool 0 only) > Dictionary cache 426512815 \t(424997488 + 1515327) > File system 429392 \t(82672 + 346720) > Lock system 296805680 \t(295674296 + 1131384) > Recovery system 0 \t(0 + 0) > Dictionary memory allocated 1515327 > Buffer pool size 6553599 > Buffer pool size, bytes 107374166016 > Free buffers 0 > Database pages 6102747 > Old database pages 2252752 > Modified db pages 1 > Pending reads 1 > Pending writes: LRU 0, flush list 0, single page 0 > Pages made young 48135625, not young 0 > 64.83 youngs/s, 0.00 non-youngs/s > Pages read 49405737, created 7213276, written 266046626 > 49.44 reads/s, 0.00 creates/s, 0.11 writes/s > Buffer pool hit rate 1000 / 1000, young-making rate 0 / 1000 not 0 / 1000 > Pages read ahead 0.00/s, evicted without access 40.11/s, Random read > ahead 0.00/s > LRU len: 6102747, unzip_LRU len: 0 > I/O sum[2662]:cur[22], unzip sum[0]:cur[0] > -------------- > ROW OPERATIONS > -------------- > 1 queries inside InnoDB, 0 queries in queue > 2 read views open inside InnoDB > ---OLDEST VIEW--- > Normal read view > Read view low limit trx n:o 2F754555 > Read view up limit trx id 2F754555 > Read view low limit trx id 2F754555 > Read view individually stored trx ids: > ----------------- > Main thread process no. 15888, id 140381271201536, state: sleeping > Number of rows inserted 46059494, updated 2501543763, deleted 28801859, > read 17046037029 > 0.00 inserts/s, 0.00 updates/s, 0.00 deletes/s, 48965.50 reads/s > ------------ > TRANSACTIONS > ------------ > Trx id counter 2F89DD34 > Purge done for trx's n:o < 2F754554 undo n:o < 0 > History list length 140 > LIST OF TRANSACTIONS FOR EACH SESSION: > ---TRANSACTION 0, not started > MySQL thread id 113586, OS thread handle 0x7fab314fe700, query id > 45250432 localhost root sleeping > SHOW engine INNODB STATUS > ---TRANSACTION 2F89DC7B, not started > MySQL thread id 113583, OS thread handle 0x7fab315d9700, query id > 45250367 cl4.logical-line.de 192.168.0.16 stats_pflege sleeping > ---TRANSACTION 2F89DB4F, not started > MySQL thread id 113582, OS thread handle 0x7fab31622700, query id > 45250261 cl4.logical-line.de 192.168.0.16 stats_pflege sleeping > ---TRANSACTION 2F89DA23, not started > MySQL thread id 113578, OS thread handle 0x7fab31746700, query id > 45250159 cl4.logical-line.de 192.168.0.16 stats_pflege sleeping > ---TRANSACTION 0, not started > <many many rows like these, around 3000 although max_connections is 300?! > > > As soon as I trigger the restart of mysql on db9 (service mysql > restart), I see normal log entries like this: > May 31 10:16:59 db9 mysqld: 130531 10:16:59 [Note] WSREP: Stop replication > May 31 10:16:59 db9 mysqld: 130531 10:16:59 [Note] WSREP: Closing send > monitor... > May 31 10:16:59 db9 mysqld: 130531 10:16:59 [Note] WSREP: Closed send > monitor. > May 31 10:16:59 db9 mysqld: 130531 10:16:59 [Note] WSREP: gcomm: > terminating thread > > and db8 is accepting inserts again and catches up with replication. > > meanwhile, db9 says: > May 31 10:17:02 db9 mysqld: 130531 10:17:02 [Note] WSREP: killing local > connection: 113529 > May 31 10:17:02 db9 mysqld: 130531 10:17:02 [Note] WSREP: killing local > connection: 113199 > May 31 10:17:02 db9 mysqld: 130531 10:17:02 [Note] WSREP: killing local > connection: 110221 > > and nothing happens any more. No CPU and no IO load coming from the > mysqld processes. > After 5 minutes I killed the mysqld with "killall -9 mysqld" and then > started it again the normal way. > db9 syncs via IST and everything is fine again. But I have no idea what > caused all this. According to a colleague killing mysql on db9 was the > thing that helped 6 weeks ago, too. > > Thanks for any pointers where we should look further. > > Jan Kirchhoff >

1 0

[Maria-developers] Open Questions in the knowledgebase
by Ian Gilfillan 31 May '13

31 May '13

I've been cleaning up the questions on the knowledgebase, but there are still a few open ones that need a response. If anyone can help, please pay a visit. Here are the ones still needing feedback: Multiple error logs: https://kb.askmonty.org/en/multiple-err-logs/ Query optimizer: https://kb.askmonty.org/en/does-the-query-optimize-a-low-cardinality-index-… Strange SSL messages: https://kb.askmonty.org/en/strange-ssl-messages-in-error-log/ Multi-source replication and time zones: https://kb.askmonty.org/en/time-zones-multi-source-replication/ And of course, the old faithful, when will it be released (10.x as stable) - https://kb.askmonty.org/en/release-date-for-10x-stable-version/ Also, a comment asking for specifics on JSON implementation going forward: https://kb.askmonty.org/en/column_json/ thanks, ian

1 0

[Maria-developers] a bug also in mariadb
by xiaobing jiang 30 May '13

30 May '13

hi all: I submit a bug: http://bugs.mysql.com/bug.php?id=69352 . I review the mariadb code, I think it is also in mariadb. I have read the page. https://kb.askmonty.org/en/reporting-bugs/ should I report this bug to mariadb? Thanks!

1 0

[Maria-developers] GSoC 2013 - Interested Student - Roles Project
by Vicentiu Ciorbaru 28 May '13

28 May '13

Hello, My name is Vicențiu Ciorbaru. I am 3'rd year undergraduate student at the Polytechnic University of Bucharest, Romania and I am very interested in contributing to MariaDB as part of Google Summer of Code. I have not used MariaDB previously but I have taken it for a spin just now. I have participated in GSoC once before and would like to do so again. I have contributed to other open source organisations (Mozilla, Samba, ROSEdu) and I submit patches when time permits. The project that interests me the most is MDEV-4397: Roles. I want to ask you what would be a good place to start in the codebase to help me get an idea what I need to do for this project? Looking forward to hearing from you, Vicențiu

2 7

Re: [Maria-developers] GTID work priorities
by Kristian Nielsen 28 May '13

28 May '13

Pavel Ivanov <pivanof(a)google.com> writes: > we would really love to see "strict gtid mode" and related stuff next. I've now pushed strict mode to 10.0-base, and documented it in the knowledgebase. However, one thing was not clear from your description in MDEV-4478: > 4) Probably with a separate flag we want to prohibit any writes to binlog on > the slave except from the slave thread(s), i.e. whenever MASTER_HOST is > configured server works as if sql_log_bin = 0 globally except for the slave > thread(s) which work as if sql_log_bin = 1 and SELECT @@global.sql_log_bin > still returns 1. Can you elaborate exactly what semantics you have in mind for this new option? And why a new option is needed/how it differs from the already existing read-only and sql-log-bin options? It sounds to me what you want is simply to be able to start the server with --sql-log-bin=0 and still have the slave sql thread log its events to the binlog (if --log-slave-updates=1)? I would expect that to work already, but maybe there is a bug (that we could fix), or maybe you had something else in mind? - Kristian.

2 1

[Maria-developers] Query cache information plugin - MDEV-4516
by Roberto Spadim 28 May '13

28 May '13

Hi guys i'm reading query cache code, and i have some doubts... mariadb-10.0.1 source: 1)PROTOCOL_LOCAL: at line: 1412 (Query_cache::store_query) we have this comment: /* PROTOCOL_LOCAL results are not cached. */ what's PROTOCOL_LOCAL ? it's the unix socket? windows pipe? embedded? memcache? mysql? another thing like this? or it's the internal mariadb source running queries? 2)the query_size=0 (MDEV-4516) i was reading at line 1680 if (is_disabled() || thd->locked_tables_mode || thd->variables.query_cache_type == 0) goto err; and line : 1694 DBUG_ASSERT(query_cache_size != 0); // otherwise cache would be disabled why it's isn't used in query_cache_information plugin? *check that in line 1680 it test the "query_cache_type" too, * *in MDEV-4516 report about size=0 we didn't tested this case!!!* 3) i was at line 1886: DBUG_PRINT("qcache", ("\ long %d, 4.1: %d, bin_proto: %d, more results %d, pkt_nr: %d, \ CS client: %u, CS result: %u, CS conn: %u, limit: %lu, TZ: 0x%lx, \ sql mode: 0x%llx, sort len: %lu, conncat len: %lu, div_precision: %lu, \ def_week_frmt: %lu, in_trans: %d, autocommit: %d", is this the "primary key" of the query cache!?!?!?!?!?!?! could we use a sha1 or another "primary key" method (index [] of the query in cache + '';'+ 5 bytes of a hash key to compare if the index still with the same query) to easly show query entry "number", and allow a "delete from query_cache_information where key=xxx" function? at line 1938 we have this maybe the query position on cache ("primary key of query cache")?: DBUG_PRINT("qcache", ("Query have result 0x%lx", (ulong) query)); 4) there's some more ideas about query cache information plugin at MDEV-4571, if anyone could read it =) well that's all =] -- Roberto Spadim

2 19

[Maria-developers] GTID replication and relay_log.info
by Pavel Ivanov 28 May '13

28 May '13

Kristian, There's a well-known problem in MySQL that relay_log.info is not always in sync with the database state, and killing and restarting server at an unfortunate time will cause it to re-execute last statement and potentially break replication. Did you fix this situation with GTIDs and rpl_slave_state table? It looks like with GTIDs it's really easy to solve the problem -- rpl_slave_state has GTID of the last executed transaction, so on restart one needs just to ignore all relay logs and re-connect to master with the last GTID. Did you do that? If not are you planning to do that? Thank you, Pavel

2 11

Re: [Maria-developers] MariaDB 5.5 EOL plans
by Honza Horak 27 May '13

27 May '13

On 05/24/2013 11:04 PM, Rasmus Johansson wrote: > Hi, > > On Tue, May 14, 2013 at 5:57 PM, Honza Horak <hhorak(a)redhat.com > <mailto:hhorak@redhat.com>> wrote: > > Hi guys, > > I'm not able to find any info about life cycle of MariaDB 5.5 or > even other versions. Particularly if the MariaDB 5.5 life cycle will > somehow follow the MySQL's 5.5 life cycle or if there is some > another vision/rough plans when supporting of 5.5 will end? > > > Thanks for bringing this up! It has been on the table quite some times, > but now finally I think we got it documented. So here it is: > Thank you very much, that looks really great. Honza

1 0

[Maria-developers] Syncs server status variable
by Ian Gilfillan 26 May '13

26 May '13

What does the Syncs server status variable record, and when was it introduced? I can't seem to find this documented anywhere. thanks, ian gilfillan

2 1

[Maria-developers] MariaDB 5.5 EOL plans
by Honza Horak 26 May '13

26 May '13

Hi guys, I'm not able to find any info about life cycle of MariaDB 5.5 or even other versions. Particularly if the MariaDB 5.5 life cycle will somehow follow the MySQL's 5.5 life cycle or if there is some another vision/rough plans when supporting of 5.5 will end? Cheers, Honza

3 3

[Maria-developers] query cache restrictions
by Roberto Spadim 25 May '13

25 May '13

hi guys, why we can't cache "union" queries with query cache? http://www.mysqlperformanceblog.com/2006/07/27/mysql-query-cache/ *Caching full queries only *– Meaning it does not work for subselects, inline views, parts of the UNION. This is also common missunderstanding. -- Roberto Spadim

1 1

Re: [Maria-developers] To Sergei:
by Sergei Golubchik 24 May '13

24 May '13

Hi, Samarth! On May 24, Samarth Mathur wrote: > I put the <plugin>.c file and the CmakeLists.txt in a folder on my desktop > and ran > *cmake .* > > CMake Warning (dev) in CMakeLists.txt: > No cmake_minimum_required command is present. A line of code such as > > cmake_minimum_required(VERSION 2.8) > > should be added at the top of the file. The version specified may be > lower > if you wish to support older CMake versions for this project. For more > information run "cmake --help-policy CMP0000". > This warning is for project developers. Use -Wno-dev to suppress it. > > then i ran > *cmake . -Wno-dev* > > CMake Error at CMakeLists.txt:1 (MYSQL_ADD_PLUGIN): > Unknown CMake command "MYSQL_ADD_PLUGIN". You need to do that from the MariaDB source tree. Put your plugin in the plugin/ directory, where other plugins are. Regards, Sergei

1 0

[Maria-developers] Fwd: To Sergei:
by Samarth Mathur 24 May '13

24 May '13

I put the <plugin>.c file and the CmakeLists.txt in a folder on my desktop and ran *cmake .* CMake Warning (dev) in CMakeLists.txt: No cmake_minimum_required command is present. A line of code such as cmake_minimum_required(VERSION 2.8) should be added at the top of the file. The version specified may be lower if you wish to support older CMake versions for this project. For more information run "cmake --help-policy CMP0000". This warning is for project developers. Use -Wno-dev to suppress it. then i ran *cmake . -Wno-dev* CMake Error at CMakeLists.txt:1 (MYSQL_ADD_PLUGIN): Unknown CMake command "MYSQL_ADD_PLUGIN". -- Configuring incomplete, errors occurred! Thanks On Fri, May 24, 2013 at 10:40 PM, Sergei Golubchik <serg(a)askmonty.org>wrote: > Hi, Samarth! > > On May 24, Samarth Mathur wrote: > > Hi, > > > > How do I go about compiling the plugin in mariadb? I searched online, but > > only found the mysql method : > > https://dev.mysql.com/doc/refman/5.5/en/compiling-plugin-libraries.html > > I'm getting some errors there, but, I wanna know, if it's done > differently > > or like this? > > Like in the manual, at the link above. > What errors are you getting? > > Regards, > Sergei >

1 0

Re: [Maria-developers] To Sergei:
by Samarth Mathur 24 May '13

24 May '13

Hi, How do I go about compiling the plugin in mariadb? I searched online, but only found the mysql method : https://dev.mysql.com/doc/refman/5.5/en/compiling-plugin-libraries.html I'm getting some errors there, but, I wanna know, if it's done differently or like this? Thanks On Thu, May 23, 2013 at 6:28 PM, Sergei Golubchik <serg(a)askmonty.org> wrote: > Hi, Samarth! > > On May 23, Samarth Mathur wrote: > > 2 questions: > > > > to make .so file from .c, is it necessary for me to compile mariadb > > after each change? Or is there a shorter way as far as plugins are > > concerned? > > You only need to compile the plugin, you don't need to compile MariaDB > at all, you can use a binary from our site or as provided by your OS. > > > Is there any specific tip you would have as far as coding is > > concerned? Or does the normal convention of good code, good comments > > and indentation work alright? > > Just normal convention. The usual - see how it's done elsewhere in > MariaDB and do the same. > > Regards, > Sergei > >

2 1

[Maria-developers] Fusion-io atomic writes patch comments
by Laurynas Biveinis 24 May '13

24 May '13

Hi - We ported your patch to Percona Server, meanwhile we collected the following comments which perhaps are useful for you. These are addressed in our XtraDB port, so you will pick them up, but since you patched InnoDB as well, you might consider addressing them there as well. I tried to order them roughly in the order of importance. The comments below refer to rev 3720 of 5.5 tree (and one comment to MariaDB 10 trunk). - Where are the testcases? sys_vars test doesn't count :) - fil_extend_space_to_desired_ size() fails to consider that it might be extending a single node on a multi-node tablespace. It will calculate the desired size as if the last node was the only one, overextending the tablespace by the size of all the previous nodes. Fix is replacing offset_high = size_after_extend * page_size / (4ULL*1024*1024*1024); offset_low = size_after_extend * page_size % (4ULL*1024*1024*1024); with offset_high = (size_after_extend - file_start_page_no) * page_size / (4ULL * 1024 * 1024 * 1024); offset_low = (size_after_extend - file_start_page_no) * page_size % (4ULL * 1024 * 1024 * 1024); - In the same function in InnoDB 5.6 / MariaDB 10.0 it is necessary to set node->being_extended = FALSE; before jumping to complete_io. - In the same function, complete_io: label may skip over the fil_node_complete_io() call, which is required for os_aio() tablespace-extending call only. - os_file_set_atomic_writes() should return not an int, but bool or ibool instead, also its return value currently is raw ioctl() result, which is used as a bool at callers anyway and reversed from the InnoDB convention of true/non-zero meaning success. While at that, it's also nice to add __attribute__((warn_unused_result)) to it. - If ioctl() failed, then it's customary for the failed I/O call handling to call one of the os_file_handle_error... functions, here os_file_handle_error_no_exit() looks like a good fit. - If atomic writes are unsupported, then I'm not sure why set errno / LastError. No syscall has failed, the atomic writes unsupported-ness is communicated by the os_file_set_atomic_writes() return value. - os_file_set_atomic_writes() declaration should not depend on __linux__ etc, thus we pushed the #ifdef to function body. - InnoDB uses UNIV_LINUX instead of __linux__, the build sys checks for the presence of sys/ioctl.h and defines corresponding define, thus we replaced #ifdef __linux__ #include <sys/ioctl.h> #ifndef DFS_IOCTL_ATOMIC_WRITE_SET #define DFS_IOCTL_ATOMIC_WRITE_SET _IOW(0x95, 2, uint) #endif with #if defined(UNIV_LINUX) && defined(HAVE_SYS_IOCTL_H) # include <sys/ioctl.h> # ifndef DFS_IOCTL_ATOMIC_WRITE_SET # define DFS_IOCTL_ATOMIC_WRITE_SET _IOW(0x95, 2, uint) # endif #endif at the top of the file, and the #ifdef inside the function is #ifdef DFS_IOCTL... - Is this bit inside __WIN__ in os_file_create_func() actually needed? if (srv_use_atomic_writes && type == OS_DATA_FILE && os_file_set_atomic_writes(file, name)) { CloseHandle(file); *success = FALSE; file = INVALID_HANDLE_VALUE; } - os_file_set_size() could call os_file_handle_error_no_exit() for failed posix_fallocate() instead of printing the errno itself. - Comments don't follow InnoDB style, missing header function comments, too long lines, spurious whitespace, etc. Our code is at https://bazaar.launchpad.net/~percona-core/percona-server/5.5/revision/518 https://bazaar.launchpad.net/~percona-core/percona-server/5.6/revision/353 We haven't addressed the testcase issue yet neither. Regards, -- Laurynas

2 2

Re: [Maria-developers] [Commits] Rev 3770: MDEV-4042 & MDEV-4536 fix. in file:///home/bell/maria/bzr/work-maria-5.5-MDEV-4536/
by Sergei Petrunia 24 May '13

24 May '13

Hi Sanja, The patch doesn't take one thing into account. Consider a query: explain select * from t12 where b>3 or exists (select max(t10.b) Z from t10 where t10.a in ( select t11.a from t11) group by t10.c having Z<100); +--+------------+-----------+------+-------------+------------+-------+----+----+-------------------------------+ |id|select_type |table |type |possible_keys|key |key_len|ref |rows|Extra | +--+------------+-----------+------+-------------+------------+-------+----+----+-------------------------------+ |1 |PRIMARY |t12 |ALL |NULL |NULL |NULL |NULL|10 | | |2 |SUBQUERY |t10 |ALL |NULL |NULL |NULL |NULL|6 |Using temporary; Using filesort| |2 |SUBQUERY |<subquery3>|eq_ref|distinct_key |distinct_key|4 |func|1 | | |3 |MATERIALIZED|t11 |ALL |NULL |NULL |NULL |NULL|6 | | +--+------------+-----------+------+-------------+------------+-------+----+----+-------------------------------+ Here, the EXISTS subquery is the one that will have the problem. I have added "in ( select t11.a from t11)" so that EXISTS subquery's join_tab array is not linear. The problem is not visible to the user, but it can be seen in debugger. Put a breakpoint in JOIN::cleanup(), with these commands: p table->alias->Ptr p table->file Run the query and you'll see: Breakpoint 5, st_join_table::cleanup (this=0x7fff280dbea0) at /home/psergey/dev2/5.5/sql/sql_select.cc:10248 $30 = 0x7fff28000af0 "t12" $31 = (ha_innobase *) 0x7fff28047198 ... Breakpoint 5, st_join_table::cleanup (this=0x7fff28053078) at /home/psergey/dev2/5.5/sql/sql_select.cc:10248 $32 = 0xd05418 "" $33 = (ha_heap *) 0x7fff2805ab98 ... Breakpoint 5, st_join_table::cleanup (this=0x7fff28051ef8) at /home/psergey/dev2/5.5/sql/sql_select.cc:10248 $36 = 0x7fff28025720 "t10" $37 = (ha_innobase *) 0x7fff28009558 ... Breakpoint 5, st_join_table::cleanup (this=0x7fff28052218) at /home/psergey/dev2/5.5/sql/sql_select.cc:10248 $38 = 0xd30a69 "sj-materialize" $39 = (ha_heap *) 0x7fff28054398 ... Note that in subquery join_tab->cleanup() was called for t10, and for sj-materialize. For t11 (which is inside the sj-materialize) it was not called. This caused by this loop being flat loop: > @@ -10646,6 +10646,12 @@ void JOIN::cleanup(bool full) > { > tab->cleanup(); > } > + if (join_tab != table_access_tabs) > + { > + for (uint i= 0; i < top_table_access_tabs_count; i++)a ^^^^^^^^^^^ here. > + table_access_tabs[i].cleanup(); > + table_access_tabs= join_tab; // avoid second cleanup() > + } > cleaned= true; > } > else It turns out, I was incorrect with the statement that "table_access_tabs array is always flat". On Tue, May 21, 2013 at 08:30:35AM +0300, sanja(a)montyprogram.com wrote: > At file:///home/bell/maria/bzr/work-maria-5.5-MDEV-4536/ > > ------------------------------------------------------------ > revno: 3770 > revision-id: sanja(a)montyprogram.com-20130521053031-xe34093yhvey0l5o > parent: monty(a)askmonty.org-20130515132812-pnjhza9iugz89356 > committer: sanja(a)montyprogram.com > branch nick: work-maria-5.5-MDEV-4536 > timestamp: Tue 2013-05-21 08:30:31 +0300 > message: > MDEV-4042 & MDEV-4536 fix. > > Optimization JOIN tabs also should be cleaned up (in case of EXPLAIN when we cleaning up everythibng at the end of execution). > === added file 'mysql-test/r/innodb_explain.result' > --- a/mysql-test/r/innodb_explain.result 1970-01-01 00:00:00 +0000 > +++ b/mysql-test/r/innodb_explain.result 2013-05-21 05:30:31 +0000 > @@ -0,0 +1,34 @@ > +drop table if exists t1,t2,t3; > +# > +# MDEV-4536: sql_base.cc:1598: bool close_thread_table(THD*, TABLE**): > +# Assertion `table->key_read == 0' failed. > +# > +set optimizer_switch="index_merge=on,index_merge_union=on,index_merge_sort_union=on,index_merge_intersection=on,index_merge_sort_intersection=off,engine_condition_pushdown=off,index_condition_pushdown=on,derived_merge=on,derived_with_keys=on,firstmatch=on,loosescan=on,materialization=on,in_to_exists=on,semijoin=on,partial_match_rowid_merge=on,partial_match_table_scan=on,subquery_cache=on,mrr=off,mrr_cost_based=off,mrr_sort_keys=off,outer_join_with_cache=on,semijoin_with_cache=on,join_cache_incremental=on,join_cache_hashed=on,join_cache_bka=on,optimize_join_buffer_size=off,table_elimination=on,extended_keys=off"; > +CREATE TABLE t1 (pk int auto_increment primary key, `col_int_key` int(11), > +key `col_int_key` (`col_int_key`),`col_varchar_key` varchar(128), key > +(col_varchar_key)) engine=innodb; > +EXPLAIN SELECT 1 FROM t1 AS alias1 WHERE EXISTS ( SELECT SQ2_alias1 . col_int_key AS SQ2_field1 FROM ( t1 AS SQ2_alias1 RIGHT OUTER JOIN t1 AS SQ2_alias2 ON (SQ2_alias2 . col_int_key = SQ2_alias1 . col_int_key ) ) GROUP BY SQ2_field1 HAVING SQ2_alias1 . col_int_key >= 7 ); > +id select_type table type possible_keys key key_len ref rows Extra > +1 PRIMARY NULL NULL NULL NULL NULL NULL NULL Impossible WHERE > +2 SUBQUERY SQ2_alias2 index NULL col_int_key 5 NULL 1 Using index; Using temporary; Using filesort > +2 SUBQUERY SQ2_alias1 ref col_int_key col_int_key 5 test.SQ2_alias2.col_int_key 1 Using where; Using index > +set optimizer_switch=default; > +drop table t1; > +# > +# MDEV-4042: Assertion `table->key_read == 0' fails in > +# close_thread_table on EXPLAIN with GROUP BY and HAVING in > +# EXISTS SQ, InnoDB > +# > +CREATE TABLE t1 (a INT) ENGINE=InnoDB; > +CREATE TABLE t2 (b INT PRIMARY KEY, c INT) ENGINE=InnoDB; > +CREATE TABLE t3 (d INT) ENGINE=InnoDB; > +EXPLAIN > +SELECT * FROM t1 > +WHERE EXISTS ( SELECT b FROM t2, t3 > +GROUP BY b > +HAVING b != 3 ); > +id select_type table type possible_keys key key_len ref rows Extra > +1 PRIMARY NULL NULL NULL NULL NULL NULL NULL Impossible WHERE > +2 SUBQUERY t2 index NULL PRIMARY 4 NULL 1 Using index; Using temporary; Using filesort > +2 SUBQUERY t3 ALL NULL NULL NULL NULL 1 Using join buffer (flat, BNL join) > +drop table t1,t2,t3; > > === added file 'mysql-test/t/innodb_explain.test' > --- a/mysql-test/t/innodb_explain.test 1970-01-01 00:00:00 +0000 > +++ b/mysql-test/t/innodb_explain.test 2013-05-21 05:30:31 +0000 > @@ -0,0 +1,41 @@ > +# > +# Innodb specific explains > + > +--disable_warnings > +drop table if exists t1,t2,t3; > +--enable_warnings > + > +--source include/have_innodb.inc > + > +--echo # > +--echo # MDEV-4536: sql_base.cc:1598: bool close_thread_table(THD*, TABLE**): > +--echo # Assertion `table->key_read == 0' failed. > +--echo # > +set optimizer_switch="index_merge=on,index_merge_union=on,index_merge_sort_union=on,index_merge_intersection=on,index_merge_sort_intersection=off,engine_condition_pushdown=off,index_condition_pushdown=on,derived_merge=on,derived_with_keys=on,firstmatch=on,loosescan=on,materialization=on,in_to_exists=on,semijoin=on,partial_match_rowid_merge=on,partial_match_table_scan=on,subquery_cache=on,mrr=off,mrr_cost_based=off,mrr_sort_keys=off,outer_join_with_cache=on,semijoin_with_cache=on,join_cache_incremental=on,join_cache_hashed=on,join_cache_bka=on,optimize_join_buffer_size=off,table_elimination=on,extended_keys=off"; > + > +CREATE TABLE t1 (pk int auto_increment primary key, `col_int_key` int(11), > +key `col_int_key` (`col_int_key`),`col_varchar_key` varchar(128), key > +(col_varchar_key)) engine=innodb; > + > +EXPLAIN SELECT 1 FROM t1 AS alias1 WHERE EXISTS ( SELECT SQ2_alias1 . col_int_key AS SQ2_field1 FROM ( t1 AS SQ2_alias1 RIGHT OUTER JOIN t1 AS SQ2_alias2 ON (SQ2_alias2 . col_int_key = SQ2_alias1 . col_int_key ) ) GROUP BY SQ2_field1 HAVING SQ2_alias1 . col_int_key >= 7 ); > + > + > +set optimizer_switch=default; > +drop table t1; > + > +--echo # > +--echo # MDEV-4042: Assertion `table->key_read == 0' fails in > +--echo # close_thread_table on EXPLAIN with GROUP BY and HAVING in > +--echo # EXISTS SQ, InnoDB > +--echo # > +CREATE TABLE t1 (a INT) ENGINE=InnoDB; > +CREATE TABLE t2 (b INT PRIMARY KEY, c INT) ENGINE=InnoDB; > +CREATE TABLE t3 (d INT) ENGINE=InnoDB; > + > +EXPLAIN > +SELECT * FROM t1 > +WHERE EXISTS ( SELECT b FROM t2, t3 > +GROUP BY b > +HAVING b != 3 ); > + > +drop table t1,t2,t3; > > === modified file 'sql/sql_select.cc' > --- a/sql/sql_select.cc 2013-05-07 11:05:09 +0000 > +++ b/sql/sql_select.cc 2013-05-21 05:30:31 +0000 > @@ -3680,8 +3680,8 @@ make_join_statistics(JOIN *join, List<TA > if (pull_out_semijoin_tables(join)) > DBUG_RETURN(TRUE); > > - join->join_tab=stat; > - join->top_join_tab_count= table_count; > + join->table_access_tabs= join->join_tab=stat; > + join->top_table_access_tabs_count= join->top_join_tab_count= table_count; > join->map2table=stat_ref; > join->table= table_vector; > join->const_tables=const_count; > @@ -10646,6 +10646,12 @@ void JOIN::cleanup(bool full) > { > tab->cleanup(); > } > + if (join_tab != table_access_tabs) > + { > + for (uint i= 0; i < top_table_access_tabs_count; i++) > + table_access_tabs[i].cleanup(); > + table_access_tabs= join_tab; // avoid second cleanup() > + } > cleaned= true; > } > else > > _______________________________________________ > commits mailing list > commits(a)mariadb.org > https://lists.askmonty.org/cgi-bin/mailman/listinfo/commits -- BR Sergei -- Sergei Petrunia, Software Developer Monty Program AB, http://askmonty.org Blog: http://s.petrunia.net/blog

1 0

Re: [Maria-developers] Fwd: [Commits] Rev 3549: Fix bug MDEV-4407 SQ pushdown: Wrong result (extra rows) with IN and EXISTS subqueries, optimizer_use_condition_selectivity>1 in file:///home/tsk/mprog/src/10.0-md83/
by Sergei Petrunia 24 May '13

24 May '13

Hi Timour, Ok to push. On Thu, May 23, 2013 at 11:19:36PM +0300, Timour Katchaounov wrote: > Sergey, > > Please approve the following patch we already discussed. > > Timour > > > ------------------------------------------------------------ > revno: 3549 > revision-id: timour(a)askmonty.org-20130523201638-4vj8cx0yjuww29nt > parent: timour(a)askmonty.org-20130520190622-rdrru0c2gbdgtjy5 > fixes bug: https://mariadb.atlassian.net/browse/MDEV-4407 > committer: timour(a)askmonty.org > branch nick: 10.0-md83 > timestamp: Thu 2013-05-23 23:16:38 +0300 > message: > Fix bug MDEV-4407 SQ pushdown: Wrong result (extra rows) with IN and EXISTS subqueries, optimizer_use_condition_selectivity>1 > (Part of MDEV-83) > > Analysis: > MDEV-83 takes into account the cost of subqueries, which results > in a materialized semi-join where the materialized table is on > the right side of the join. At the same time, the WHERE clause > has an EXISTS predicate which is moved to the second table in > the join because of smaller join cardinality. > > The semi-join startup function setup_sj_materialization_part2 > didn't take into account that in addition to the injected > IN-EXISTS conditions, there may be other conditions, such as > ones moved by pushdown of subquery predicates in mdev-83. So > setup_sj_materialization_part2 just set the condition of the > semi-join to be the IN-EXISTS equality, this removing the > subquery moved by the subquery pushdown logic. > > Solution: > And the conditions instead of just setting them. > > > > === modified file 'mysql-test/r/subselect4.result' > --- a/mysql-test/r/subselect4.result 2013-05-20 19:06:22 +0000 > +++ b/mysql-test/r/subselect4.result 2013-05-23 20:16:38 +0000 > @@ -2393,5 +2393,31 @@ COUNT(b) > 1 > 2 > drop table t1, t2, t3; > +# > +# MDEV-4407 SQ pushdown: Wrong result (extra rows) with IN and EXISTS subqueries, optimizer_use_condition_selectivity>1 > +# > +SET optimizer_switch='expensive_pred_static_pushdown=on'; > +SET optimizer_use_condition_selectivity=3; > +SET use_stat_tables=PREFERABLY; > +CREATE TABLE t1 (a INT, b INT, KEY(b)) ENGINE=MyISAM; > +INSERT INTO t1 VALUES (1,1),(2,2),(1,1),(0,0); > +ANALYZE TABLE t1; > +Table Op Msg_type Msg_text > +test.t1 analyze status OK > +FLUSH TABLES; > +EXPLAIN > +SELECT * FROM t1 AS outer_t1 WHERE b IN ( SELECT a FROM t1 WHERE b != 1 ) AND EXISTS ( SELECT 1 FROM t1 WHERE b < outer_t1.a ); > +id select_type table type possible_keys key key_len ref rows Extra > +1 PRIMARY outer_t1 ALL b NULL NULL NULL 4 > +1 PRIMARY <subquery2> eq_ref distinct_key distinct_key 4 func 1 Using where > +2 MATERIALIZED t1 ALL b NULL NULL NULL 4 Using where > +3 DEPENDENT SUBQUERY t1 index b b 5 NULL 4 Using where; Using index > +SELECT * FROM t1 AS outer_t1 WHERE b IN ( SELECT a FROM t1 WHERE b != 1 ) AND EXISTS ( SELECT 1 FROM t1 WHERE b < outer_t1.a ); > +a b > +2 2 > +drop table t1; > +SET optimizer_switch=@@global.optimizer_switch; > +SET optimizer_use_condition_selectivity=default; > +SET use_stat_tables=default; > SET optimizer_switch= @@global.optimizer_switch; > set @@tmp_table_size= @@global.tmp_table_size; > > === modified file 'mysql-test/t/subselect4.test' > --- a/mysql-test/t/subselect4.test 2013-05-20 19:06:22 +0000 > +++ b/mysql-test/t/subselect4.test 2013-05-23 20:16:38 +0000 > @@ -1917,5 +1917,29 @@ GROUP BY a; > > drop table t1, t2, t3; > > +--echo # > +--echo # MDEV-4407 SQ pushdown: Wrong result (extra rows) with IN and EXISTS subqueries, optimizer_use_condition_selectivity>1 > +--echo # > + > +SET optimizer_switch='expensive_pred_static_pushdown=on'; > +SET optimizer_use_condition_selectivity=3; > +SET use_stat_tables=PREFERABLY; > + > +CREATE TABLE t1 (a INT, b INT, KEY(b)) ENGINE=MyISAM; > +INSERT INTO t1 VALUES (1,1),(2,2),(1,1),(0,0); > +ANALYZE TABLE t1; > +FLUSH TABLES; > + > +EXPLAIN > +SELECT * FROM t1 AS outer_t1 WHERE b IN ( SELECT a FROM t1 WHERE b != 1 ) AND EXISTS ( SELECT 1 FROM t1 WHERE b < outer_t1.a ); > +SELECT * FROM t1 AS outer_t1 WHERE b IN ( SELECT a FROM t1 WHERE b != 1 ) AND EXISTS ( SELECT 1 FROM t1 WHERE b < outer_t1.a ); > + > +drop table t1; > + > +SET optimizer_switch=@@global.optimizer_switch; > +SET optimizer_use_condition_selectivity=default; > +SET use_stat_tables=default; > + > + > SET optimizer_switch= @@global.optimizer_switch; > set @@tmp_table_size= @@global.tmp_table_size; > > === modified file 'sql/opt_subselect.cc' > --- a/sql/opt_subselect.cc 2013-04-01 10:36:05 +0000 > +++ b/sql/opt_subselect.cc 2013-05-23 20:16:38 +0000 > @@ -3620,7 +3620,10 @@ bool setup_sj_materialization_part2(JOIN > emb_sj_nest->sj_subq_pred))) > DBUG_RETURN(TRUE); /* purecov: inspected */ > sjm_tab->type= JT_EQ_REF; > - sjm_tab->select_cond= sjm->in_equality; > + remove_sj_conds(&sjm_tab->select_cond); > + sjm_tab->select_cond= and_items(sjm_tab->select_cond, sjm->in_equality); > + if (!sjm_tab->select_cond->fixed) > + sjm_tab->select_cond->fix_fields(thd, &sjm_tab->select_cond); > } > else > { > > > _______________________________________________ > commits mailing list > commits(a)mariadb.org > https://lists.askmonty.org/cgi-bin/mailman/listinfo/commits -- BR Sergei -- Sergei Petrunia, Software Developer Monty Program AB, http://askmonty.org Blog: http://s.petrunia.net/blog

1 0

Re: [Maria-developers] [Commits] Rev 3769: MDEV-4520 fix
by Sergei Golubchik 23 May '13

23 May '13

Hi, Oleksandr! On May 23, Oleksandr Byelkin wrote: > 22.05.2013 20:27, Sergei Golubchik пишет: > > On May 22, sanja(a)montyprogram.com wrote: > >> ------------------------------------------------------------ > >> revno: 3769 > >> revision-id: sanja(a)montyprogram.com-20130522125526-aw778iepj1yceskd > >> parent: holyfoot(a)askmonty.org-20130514213637-3uvcda98gemgiquf > >> committer: sanja(a)montyprogram.com > >> branch nick: work-maria-5.5-MDEV-4520 > >> timestamp: Wed 2013-05-22 15:55:26 +0300 > >> message: > >> MDEV-4520 fix. > > First, please, fix the changeset comment. The first line should be > > MDEV-<number><space><complete issue summary> > > > > For example: > > > > MDEV-4520 Assertion `0' fails in Query_cache::end_of_result on concurrent drop event and event execution > > > > Then, after an empty line, you add whatever explanations you think are > > necessary. Like the one below: > The only question is how happened that I did not know standard of commit > comment (where that standard is/was described)? This is a de facto standard. It wasn't formally stated as a rule to follow, but everyone is doing it anyway, and it's really convenient to have changeset comments structured this way. But now I don't mind making it a written rule - because it practically won't change anything :) Where do you suggest it should be written? > >> If there is no net.vio then query cache cant't get data via > >> net_real_write() so it is better just do not try to cache such > >> query. > > Also, it'd be good to record the fixed bug in the changeset metadata. > > This can be done either with "bzr gcommit" and my patches to it (see the > > article in KB), or from the command line with "bzr commit --fixes". > > Again previous question... It's not at all a standard. It's nice to have, but very few are doing it. Feel free not to. > >> === modified file 'sql/sql_cache.cc' > >> --- a/sql/sql_cache.cc 2013-05-07 11:05:09 +0000 > >> +++ b/sql/sql_cache.cc 2013-05-22 12:55:26 +0000 > >> @@ -3977,7 +3977,8 @@ Query_cache::is_cacheable(THD *thd, LEX > >> if (thd->lex->safe_to_cache_query && > >> (thd->variables.query_cache_type == 1 || > >> (thd->variables.query_cache_type == 2 && (lex->select_lex.options & > >> - OPTION_TO_QUERY_CACHE)))) > >> + OPTION_TO_QUERY_CACHE))) && > >> + thd->net.vio) > >> { > >> DBUG_PRINT("qcache", ("options: %lx %lx type: %u", > >> (long) OPTION_TO_QUERY_CACHE, > > The fix itself looks ok. But please add a test case. > > > > And think if there can be more cases (besides events) where SELECT is > > executed with thd->net.vio == 0. E.g. init-file? or init-connect? > > Or INSERT (in the replication thread) that invokes a stored function or > > a trigger? Anything you can think of. > I spent more then half a day with this and here is result: > > - vio is definitely unset for replication and/or events (I have no idea > how that tests work, probably it will require several days to figure out) Right, that's why I suggested a test with events. And a test with a SELECT in the replication thread. > - one bug found in mysql-test-run.pl and options files (MDEV-4567) (the > bug really slows down any investigation about any test with options). Sanja, if you want your submitted bug to have any chance of ever being fixed, please at least specify the affected version and the fix version. > - putting SELECT statement in --init_connect does not work. What does it mean? How exactly it "does not work" ? > I was trying to catch it with putting assertion in my_net_write. there > is no great success so far. > > Should I continue or left it as is taking into account that bug was > reported against our standard test suite? But that's the point - it's not against our standard test suite. It's a test from the test suite run with the non-default command line options. Nobody runs the test suite this way, so if this will ever be broken again, it won't be noticed, no test will fail. That's why we need a test for this bug. Regards, Sergei

1 0

[Maria-developers] Proposed changes to GTID syntax
by Kristian Nielsen 23 May '13

23 May '13

I am planning some changes to the user interface of Global Transaction ID. Testing by Elena has shown usability issues that I want to address with these changes, but I wanted to give everyone a change to voice their opinion on the changes. The problem: The issue concerns when a slave server initially connects to the master using GTID. The slave needs to send to the master the GTID position it wants to start replicating from (the GTID of the last transaction executed within each replication domain). If the connecting server was already a slave to another master before, then the position is the last replicated event(s), stored in table mysql.rpl_slave_state. But if we server is the prior master that is now to become a slave, we need to connect at the position of the last event(s) logged by the prior master in the binlog. The problem occurs when users do manual updates on the slave and those updates get into the binlog of the slave. Of course, in principle one should not do that, as it creates differences between the master and the slave. But it is a very common thing to do among users no doubt. Maybe they need to fix some replication problem, maybe add an account only on the slave, whatever, and they forget to set @@sql_log_bin=0 while doing the change, so it gets into the slave binlog. With current code, if the slave connects to a new master right after doing the manual changes to the binlog on the slave, it will connect at the GTID position of those changes, not at the last replicated transaction. Because doing changes directly on the slave server in effect makes that server temporarily a "master". And besides, this is user error anyway, creating divergence betheen the binlogs on slave and master server. But I agree with Elena that this will be a nasty surprise to many users trying out GTID for the first time. It is inconsistent with old-style replication, where direct updates on the slave never change the slave replication position. And many users will not read the complete documentation on GTID before trying to use it, and will not understand the issues with manual changes on a slave in effect turning it into a master temporarily. Note that this affects usability when users abuse the replication, using it against the recommended way, not taking care to keep master and slaves in sync. For disciplined usage, eg. running with the "strict GTID" mode that we plan to implement, these kind of problems will never occur in the first place. But I want to do as much as possible to also make things work well for "undisciplined" users. The proposed solution: There will be two ways to configure a slave to use GTID. One will be for "disciplined" usage, same as the current code, where slave connects with whatever is most recent in mysql.rpl_slave_state or binlog. The other will be for "sloppy" usage, where the user will need to explicitly state whether to use the position from the last replicated transaction, or to use what is in the binlog. I hope this will solve the problem. Since new users need to learn new syntax to enable GTID anyway (CHANGE MASTER ... master_use_gtid=...), they will at least have made a conscious choice about whether to use one syntax or the other. And for "disciplined" users there is not really any change, they will just use the syntax for the current behaviour. I also think this improves things on the conceptual level. Currently, there is too much magic about the value of the variable @@GLOBAL.gtid_pos. You can set this manually (and that updates mysql.rpl_slave_state), but it also changes value whenever something is added to the binlog, and you cannot manually set it to something that conflicts with the binlog. This magic all becomes unnecessary with the proposed changes. So there will be some changes to the syntax for CHANGE MASTER, and some stuff will be renamed to make things clearer. Details: CHANGE MASTER syntax will be modified to this: 1. CHANGE MASTER TO master_use_gtid = current_gtid_pos This is the "disciplined" mode, the same as master_use_gtid=1 in current code. It will use as replication start position whatever is most recent, last replicated GTID or last binlogged GTID. 2. CHANGE MASTER TO master_use_gtid = slave_gtid_pos This is "sloppy" mode, for pointing an existing slave to a new master. It will only look at the last replicated GTID for the start position, not at any transactions logged directly to the slave binlog. 3. CHANGE MASTER TO master_use_gtid = off For completeness, to turn of using GTID to connect, go back to using old-style replication. There will be three new system variables introduced, replacing the current @@GLOBAL.gtid_pos: 1. @@GLOBAL.slave_gtid_pos. Read-write. This is the set of last replicated GTID, per replication domain. 2. @@GLOBAL.binlog_gtid_pos. Read only. This is the set of last binlogged GTID, per domain. 3. @@GLOBAL.current_gtid_pos. Read only. This is a combination of @@GLOBAL.slave_gtid_pos and @@GLOBAL.binlog_gtid_pos, per-domain it holds the most recent GTID either replicated or binlogged. The current mysql.rpl_slave_state table will be renamed to mysql.slave_gtid_pos. Note how the naming is now more consistent. slave_gtid_pos means the most recently replicated GTIDs, both for table, variable, and CHANGE MASTER syntax. Similarly, current_gtid_pos is used for both the variable and CHANGE MASTER to mean the most recent GTID either replicated or binlogged. The user can see the GTID position the slave will start replicating at by inspecting the value of the corresponding variable. When using master_use_gtid=slave_gtid_pos, the following statement can be used on a server that used to be the master but should now become a slave: SET GLOBAL slave_gtid_pos = @@binlog_gtid_pos; This will load whatever was logged to the binlog into the slave position. When using master_use_gtid=current_gtid_position, there is no need to do anything special, the binlog will have the most recent GTID on a prior master so it will be used automatically. I think these changes will be a good improvement, justifying doing the change this late during 10.0 release state. But I very much welcome input and suggestions for doing things differently. - Kristian.

1 2

[Maria-developers] Ubuntu Hardy and Oneiric EOL
by Daniel Bartholomew 23 May '13

23 May '13

FYI: As of the end of April, Ubuntu 8.04 "Hardy" and 11.10 "Oneiric" have reached EOL and are no longer supported by Canonical. Sometime in the next week or so I plan to turn off these builders in buildbot. The VMs will be archived, in case we need them, but no regular builds will be made. This includes Debian packages for Hardy and Oneiric and the binary tarballs which are built on Hardy (binary tarballs are also built on Ubuntu Precise, and they are not going away). The MariaDB 5.5.31 release will be the last MariaDB release to feature packages for these two distros. The repos for them will remain online until the 5.5.32 release. At that time they will go away. This also affects the Mint 9 and Mint 12 as they use the hardy and oneiric repos, respectively. Let me know if you have any questions or concerns. Thanks! -- Daniel Bartholomew Google+ - http://gplus.to/dbart Twitter - http://twitter.com/daniel_bart MariaDB: An Enhanced Drop-in Replacement for MySQL Website - http://mariadb.org Twitter - http://twitter.com/mariadb Google+ - http://google.com/+mariadb Facebook - http://fb.com/MariaDB.dbms Knowledgebase - http://kb.askmonty.org

1 0

[Maria-developers] MariaDB 5.5.31 Release Prep
by Daniel Bartholomew 23 May '13

23 May '13

All, I've uploaded almost all of the packages for the release (working on a fix for an error with the sol-10 package) and the mirrors will be picking them up soon. Draft Changelog and Release Notes are here: - https://kb.askmonty.org/en/mariadb-5531-release-notes/ - https://kb.askmonty.org/en/mariadb-5531-changelog/ One of the big items in this release is TokuDB. We didn't have a proper category for TokuDB yet, so I created a stub TokuDB category with some links to the Tokutek website: - https://kb.askmonty.org/en/tokudb/ If there is anything I should be sure to mention in the release notes, please let me know ASAP. My current plan is to activate the release tomorrow, 23 May 2013. Thanks! -- Daniel Bartholomew Google+ - http://gplus.to/dbart Twitter - http://twitter.com/daniel_bart MariaDB: An Enhanced Drop-in Replacement for MySQL Website - http://mariadb.org Twitter - http://twitter.com/mariadb Google+ - http://google.com/+mariadb Facebook - http://fb.com/MariaDB.dbms Knowledgebase - http://kb.askmonty.org

3 4

Re: [Maria-developers] [Commits] Rev 3769: MDEV-4520 fix
by Sergei Golubchik 22 May '13

22 May '13

Hi, Sanja! On May 22, sanja(a)montyprogram.com wrote: > ------------------------------------------------------------ > revno: 3769 > revision-id: sanja(a)montyprogram.com-20130522125526-aw778iepj1yceskd > parent: holyfoot(a)askmonty.org-20130514213637-3uvcda98gemgiquf > committer: sanja(a)montyprogram.com > branch nick: work-maria-5.5-MDEV-4520 > timestamp: Wed 2013-05-22 15:55:26 +0300 > message: > MDEV-4520 fix. First, please, fix the changeset comment. The first line should be MDEV-<number><space><complete issue summary> For example: MDEV-4520 Assertion `0' fails in Query_cache::end_of_result on concurrent drop event and event execution Then, after an empty line, you add whatever explanations you think are necessary. Like the one below: > If there is no net.vio then query cache cant't get data via > net_real_write() so it is better just do not try to cache such > query. Also, it'd be good to record the fixed bug in the changeset metadata. This can be done either with "bzr gcommit" and my patches to it (see the article in KB), or from the command line with "bzr commit --fixes". > === modified file 'sql/sql_cache.cc' > --- a/sql/sql_cache.cc 2013-05-07 11:05:09 +0000 > +++ b/sql/sql_cache.cc 2013-05-22 12:55:26 +0000 > @@ -3977,7 +3977,8 @@ Query_cache::is_cacheable(THD *thd, LEX > if (thd->lex->safe_to_cache_query && > (thd->variables.query_cache_type == 1 || > (thd->variables.query_cache_type == 2 && (lex->select_lex.options & > - OPTION_TO_QUERY_CACHE)))) > + OPTION_TO_QUERY_CACHE))) && > + thd->net.vio) > { > DBUG_PRINT("qcache", ("options: %lx %lx type: %u", > (long) OPTION_TO_QUERY_CACHE, The fix itself looks ok. But please add a test case. And think if there can be more cases (besides events) where SELECT is executed with thd->net.vio == 0. E.g. init-file? or init-connect? Or INSERT (in the replication thread) that invokes a stored function or a trigger? Anything you can think of. Regards, Sergei

1 0

[Maria-developers] History of table changes
by Roberto Spadim 22 May '13

22 May '13

Hi guys, i was looking git and i don't know if there's something similar to mariadb alter table command could we develop something like git for create database,alter tables, create table, drop table, drop database? a log that could be useful to update databases table definition and read changes done and 'commits' done by dba or developer or a tool to read my table and read remote table and show the create command to make they in sync (ok here history is lost) thanks guys -- Roberto Spadim

1 0

[Maria-developers] unexpected "Duplicate entry" error
by Kazuhiko Shiozaki 21 May '13

21 May '13

Hello, We sometimes (not always, not often but sometimes) have unexpected "Duplicate entry" error on the following table and with the following query : * table CREATE TABLE `category` ( `uid` bigint(20) unsigned NOT NULL, `category_uid` bigint(20) unsigned NOT NULL DEFAULT '0', `base_category_uid` bigint(20) unsigned NOT NULL DEFAULT '0', `category_strict_membership` tinyint(1) NOT NULL DEFAULT '0', PRIMARY KEY (`uid`,`category_uid`,`base_category_uid`,`category_strict_membership`), KEY `Membership` (`category_uid`,`base_category_uid`) ) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_unicode_ci * query BEGIN; (some queries for other tables) DELETE FROM category WHERE uid=14026268; INSERT INTO category VALUES (14026268, 3131502, 2074, 1),(14026268, 2992, 2992, 0),(14026268, 3133077, 191, 1),(14026268, 1212264, 2992, 1); (some queries for other tables) COMMIT; * error Duplicate entry '14026268-3131502-2074-1' for key 'PRIMARY' But this duplicate entry error should not happen because DELETE is called in the same transaction, even if similar queries are sent from another connection at the same time. We confirmed that this kind of issue happens on the following environments using MySQL_python-1.2.3 : * MariaDB 5.5.25, XtraDB * MariaDB 5.5.29, XtraDB * MariaDB 5.5.29, InnoDB plugin Since it happens SOMETIMES, I cannot reproduce this issue by small query. Do you have any idea so that we can investigate this issue further ? Thanks in advance ! -- Kazuhiko Shiozaki Nexedi: Consulting and Development of Free / Open Source Software http://www.nexedi.com/ ERP5: Full Featured High End Open Source ERP http://www.erp5.com/

4 12

Re: [Maria-developers] MariaDB and MySQL replication
by Kristian Nielsen 21 May '13

21 May '13

Jan Lindström <jplindst(a)mariadb.org> writes: > My question is how compatible replication between > > 1) MySQL (master) -> MariaDB (slave) > 2) MariaDB (master) -> MySQL (slave) In general, it should work, though there are some limitations. It also depends on versions. I always try to make MariaDB (master) -> MySQL (slave) work. For example, when new binlog events are introduced to MariaDB, the MariaDB master will not send them to slaves that do not understand the new events (whether MySQL slave, or older version of MariaDB), it will replace them with something compatible. This is not the best tested part of the code though, so there may be bugs. MySQL (master) -> MariaDB (slave) generally has limitations. This is the same as MySQL (high version master) -> MySQL (low version master). This is due to the MySQL replication teams approach, where you need to disable new features on the master for things to work with old slaves. For example, MySQL 5.6 (master) -> MySQL 5.5 (slave) does not work out-of-the box, so MariaDB 5.5 slave is the same. Generally, when incompatibilities are introduced in new MySQL versions used as a master, we will backport to MariaDB at least to the extent that new MariaDB versions will be able to replicate as slaves from a MySQL master without needing any special configuration. Though it depends on time available and priority against other work, for example there are known incompatibilities that have not yet been fixed in 10.0 alpha but will hopefully be fixed by 10.0 GA. Officially, replicating from new master to old slave is not supported (you should upgrade slaves first). This applies mostly the same whether MySQL or MariaDB or both are involved. Unofficially, as explained above, we do try to make it work as well as reasonably possible on the MariaDB side. > Before you say they are fully compatible, I point out at least > following types at mysql_com.h that do not exist on MariaDB mysql_com.h: > > MYSQL_TYPE_TIMESTAMP2, > MYSQL_TYPE_DATETIME2, > MYSQL_TYPE_TIME2, I do not know the details of this particular point (also you did not mention any version info). Best is probably to test it and see, and then report your findings and if needed we can file a bug for fixing things. - Kristian.

2 1

Re: [Maria-developers] Getting community to write missing test cases
by Kristian Nielsen 21 May '13

21 May '13

Elena Stepanova <elenst(a)montyprogram.com> writes: > If we start doing it, Oracle will soon realize that this activity is > based at least as much on commit comments as on the code. As a result, > they will probably start making much vaguer comments than they do now, > which in turn might make merge and related activities even more > complicated than they are now. I wouldn't worry about that. First, I do not believe day-to-day developers spend any effort to actively conceal things for outside developers. The ban on bug test cases seems to have come from a generic security team mandate, which has roots some years back where Oracle were hit quite hard with countless vulnerabilities after their "can't break it - can't break in" campain. Second, even if they did, anything that makes Oracle close down more benefits the adoption of the forks. But most importantly, we must not get into the mentality that we let what Oracle might or might not do influence how we work. We need to go in whatever direction we believe is best for the project. Oracle has made it plain that they do not want to participate in the community, and we should just ignore them until such time as they might decide otherwise. - Kristian.

1 0

[Maria-developers] Feature Request - Working auto_increment on BlackHole Engine
by Gordan Bobic 20 May '13

20 May '13

How difficult would be be to implement working auto-increment functionality on the BlackHole engine? This would be really useful for what I'm trying to do at the moment. Since it is not available and there are no sequences in MariaDB and MySQL, I have to resort to a hack such as a single-int-field table doing the auto-increment, and an insert trigger on the blackhole table to insert a null into the auto-increment tracking table, get LAST_INSERT_ID() to set the value in the auto-increment table, and then delete the record. Even with InnoDB log files and tablespace on tmpfs this makes the performance nosedive by a factor of 4 if the trigger doesn't bother cleaning up the auto-increment table, and a factor of 13 if it deletes the record immediately after inserting it. The actual times I tested with 24 inserter threads on a 24-thread server are 8m vs. 30m vs. 105m, i.e. a massive drop. Is there an better place than this list for a feature request? Gordan

3 6

[Maria-developers] Ogg vorbis based table
by Roberto Spadim 18 May '13

18 May '13

Hi guys, i was listening some musics, and i had a idea... let's me explain Some systems are based in time and value, an example of scada systems. In this kind of system we have an 'channel' (like right or left sound channel) a timestamp and a value, normally a table like CREATE TABLE history( Channel int not null default 0, Datetimestamp decimal(22,7) not null default 0, Value decimal(22,5) default NULL, Primary key (channel,datetimestamp) ) Some explanations and features that scada systems use... Value can be NULL, in this case we could think like MUTE function on audio files Datetime normally are gmt, or unixtimestamp values with microsecond precision, they could be fixed interval or not, example... 1 2 3 4 5... (q second interval), in audio it's fixed sample rate, or not fixed interval (variable sample rate), like 1 3 4 9... Well what i'm thinking... today mariadb 10.0.2 started the CONNECT table type (please correct me if i'm wrong), that can 'connect' to a xml,txt,csv and others kinds of files to read and write it's values... Could we implement mp3, wave, ogg vorbis, file type type? This is a really nice 'table compression' for scada systems, and allow a new era of table type, think about... Select datetimestamp from someaudiofile where someniceaudiofunction()>1 In this example we could check date time where some scada value get over a fixed level, or where a frequency get higher and start a alarm condition, it could be used to get more information using statistics functions... i'm talking now about a new table type and new functions... but thats the idea... Anyone intested in this? About table compression think about compressing tax values in nasdaq for example, and execute a statistic analyse with this information... Well just some uses... Thanks for the space to show my idea and your time to read it Maybe a video table in future, or image based... just to use jpeg compreesion (x,y,color values) and others kinds of files... Bye

1 0

[Maria-developers] Empty_queries vs Com_empty_queries
by Ian Gilfillan 18 May '13

18 May '13

I'm documenting the status variables at the moment, and have a question about Empty_queries vs Com_empty_queries. Com_empty_queries is not incremented when an empty result is returned, and appears to be incremented only when an empty query is sent to the server, eg: mysql --comments -e '-- sql comment' Empty_queries appears to match the empty queries returned with SHOW USER_STATISTICS. It increments each time a query is performed that returns an empty result. Currently the documentation on the user statistics page at https://kb.askmonty.org/en/user-statistics/ says that EMPTY_QUERIES (and therefore the Empty_queries status variable) is "The number of times this user's connections sent empty queries to the server", which would describe Com_empty_queries rather than Empty_queries, and would need to be corrected. The documentation on https://kb.askmonty.org/en/server-status-variables/ is what I've just written based on the above. Have I understood this correctly, or is there something else to it? thanks, ian

3 2

[Maria-developers] To Sergei:
by Samarth Mathur 16 May '13

16 May '13

Hi Sergei.. just a few questions. Hope not to bother too much! does the .so file of two_questions exist in the mariadb 10.0.2 source code or was that just an example? Also, if I'm looking to implement pluggable authentication for LDAP, does the version of mariadb, i'm doing this on matter? if it does, which version would you recommend? I'm currently on v10.0.2.. What is the advantage of going for a pluggable authentication over say a module made for the PAM plugin in mariadb? Regards Sam

3 6

Re: [Maria-developers] Feature Request - Working auto_inc rement on BlackHole Engine
by Gordan Bobic 16 May '13

16 May '13

On 05/15/2013 11:27 PM, Jocelyn Fournier wrote: > Hi Gordan, > > > Le 15/05/13 22:18, Gordan Bobic a écrit : >> On 05/15/2013 09:06 PM, Michael Widenius wrote: >>> >>> Hi! >>> >>>>>>>> "Gordan" == Gordan Bobic <gordan(a)bobich.net> writes: >>> >>> Gordan> How difficult would be be to implement working >>> Gordan> auto-increment functionality on the BlackHole engine? >>> >>> Gordan> This would be really useful for what I'm trying to do at >>> Gordan> the moment. Since it is not available and there are no >>> Gordan> sequences in MariaDB and MySQL, I have to resort to a >>> Gordan> hack such as a single-int-field table doing the >>> Gordan> auto-increment, and an insert trigger on the blackhole >>> Gordan> table to insert a null into the auto-increment >>> Gordan> tracking table, get LAST_INSERT_ID() to set the >>> Gordan> value in the auto-increment table, and then delete >>> Gordan> the record. >>> >>> What problem is that you wanted to solve? >>> Normally blackhole is used just for creating a slave of a slave that >>> doesn't have to hold any data (only the logs). >> >> The problem is a huge volume logging application that needs to have >> uniquely enumerated records for auditing purposes. > > Have you tried TokuDB engine ? It isn't suitable for what I am doing in this case - note that the blackhole engine in this case suits all of my requirements except for the working auto_increment functionality. I don't need to store the data on these nodes, merely binlog it up for further processing downstream. Gordan

1 0

[Maria-developers] MariaDB GTID first impression
by Giuseppe Maxia 15 May '13

15 May '13

Kristian, Thanks for the docs. It's a very good start. Compared to MySQL 5.6, GTID in MariaDB has some advantages: it is enabled by default and it is easy to read. There are a few items, though, that should need some more thought: 1) rpl_slave_state is probably ill-named. In this architecture, it is only written by the slaves, and thus the name matches, but I wonder if the master should write this table (or a similar one) to keep track of its progress in the binary log. Rationale: all the info related to replication should be in tables and easy to query. 2) rpl_slave_state does not have enough metadata. What is missing here, and would be extremely useful, is the corresponding binlog+position for each GTID, a timestamp of when the transaction was saved to the binary log, and the timestamp of when it was applied. (The host name of the origin would also be useful). Also very important (when you will implement parallel replication) is a shard identifier. 3) Multiple domains do not integrate well with GTID. In multi-source replication, we use "connection names", but for GTID we use "domain numbers". They do not play well together : there is no correspondence between the gtid_domain_id and the connection name, and a "SHOW ALL SLAVES STATUS" lists all the GTID for every connection. For example: mysql -e 'show all slaves status\G' |grep 'Connection_name\|Running\|Master_Host\|Master_Port\|Gtid_Pos\|Master_Log' Connection_name: Bear Master_Host: 127.0.0.1 Master_Port: 8203 Master_Log_File: mysql-bin.000001 Read_Master_Log_Pos: 1020 Relay_Master_Log_File: mysql-bin.000001 Slave_IO_Running: Yes Slave_SQL_Running: Yes Exec_Master_Log_Pos: 1020 Gtid_Pos: 100-101-21,400-104-23,200-102-8,300-103-12,500-105-20 Connection_name: Wolf Master_Host: 127.0.0.1 Master_Port: 8204 Master_Log_File: mysql-bin.000001 Read_Master_Log_Pos: 896 Relay_Master_Log_File: mysql-bin.000001 Slave_IO_Running: Yes Slave_SQL_Running: Yes Exec_Master_Log_Pos: 896 Gtid_Pos: 100-101-21,400-104-23,200-102-8,300-103-12,500-105-20 Connection_name: Fox Master_Host: 127.0.0.1 Master_Port: 8205 Master_Log_File: mysql-bin.000002 Read_Master_Log_Pos: 365 Relay_Master_Log_File: mysql-bin.000002 Slave_IO_Running: Yes Slave_SQL_Running: Yes Exec_Master_Log_Pos: 365 Gtid_Pos: 100-101-21,400-104-23,200-102-8,300-103-12,500-105-20 Connection_name: Deer Master_Host: 127.0.0.1 Master_Port: 8206 Master_Log_File: mysql-bin.000001 Read_Master_Log_Pos: 1144 Relay_Master_Log_File: mysql-bin.000001 Slave_IO_Running: Yes Slave_SQL_Running: Yes Exec_Master_Log_Pos: 1144 Gtid_Pos: 100-101-21,400-104-23,200-102-8,300-103-12,500-105-20 Which connection is running which gtid number? Not easy to tell, as the correspondence between domain and connections is only in the DBA's head. 4) gtid_pos is text that includes all sequences from all domains. The docs say that I can modify it if needed, but in a case like the one above, modifying this string can be very error prone. For example, if server_ids are assigned using inet_aton(@IP), you could have a gtid_pos like this '1001-170062692-230000,1002-170062593-12893445,1003-170062700-672009,1003-170062699-700002' Instead of editing a string, and risk breaking replication in all domains, I would expect the availability of commands like: CHANGE GTID for domain 'Wolf' set gtid_pos=1900 That's it for now. Hope there is still room for improvements. Cheers Giuseppe

3 16

Re: [Maria-developers] MariaDB build logs
by Kristian Nielsen 14 May '13

14 May '13

Marcin Kulisz <debian(a)kulisz.net> writes: > as promised I'm attaching build log for maria DB, it's much smaller then > anticipated. Thanks! > From what I can tell it was just one test which failed and rendered whole build as a > failure. > My build env is based on cowbuilder on fully patched SID chroot. Ah, I see what is going on: > plugins.unix_socket [ fail ] > Test ended at 2013-05-11 00:12:52 > --- /tmp/buildd/mariadb-5.5-5.5.29/builddir/mysql-test/suite/plugins/r/unix_socket.result 2013-05-10 21:49:20.510929790 +0000 > +++ /tmp/buildd/mariadb-5.5-5.5.29/builddir/mysql-test/suite/plugins/r/unix_socket.reject 2013-05-11 00:12:52.562350485 +0000 > @@ -8,7 +8,7 @@ > # > select user(), current_user(), database(); > user() current_user() database() > -USER@localhost USER@% test > +USER@localhost USER@localhost test > # > # name does not match = failure > # Right, so the problem is that this test is poorly written. Its output depends on the name of unix username of the account running the test! I am guessing that in your case, you build as user root. Since the MariaDB server running in the test happens to also have a (database) account named root, we get an authenticated user USER@localhost instead of an anonymous user. I have now pushed a fix for this problem to the MariaDB tree: http://lists.askmonty.org/pipermail/commits/2013-May/004715.html So this fix will appear in MariaDB 5.5.31 or possibly 5.5.32 (not sure if it made it in 5.5.31, which I think is due for release). Until that fix propagates to the MariaDB packaging, some options for a work-around are: - Run the build as a non-root user, with fakeroot. - Skip the test suite during build (DEB_BUILD_OPTIONS=nocheck) (this will also greatly speed up the build). - Apply the linked patch manually Hope this helps, - Kristian.

1 0

Re: [Maria-developers] Forked mytop - conflicts, issues | fullqueries overflow bug patch
by Jean Weisbuch 14 May '13

14 May '13

Hi, I attached to this message a patch to fix the overflow when fullqueries is activated and full queries are taking more than a line that might be resulting in a page overflow (the informations shown on mytop will exceed the terminal height). As for the latest commit upstream (21, mar 2013 https://github.com/jzawodn/mytop/commit/437f2ef8d3fce02eafe935ddbf860d1dfbc…) its only a small fix that seems to already have been fixed on the MariaDB version. Regards.

1 0

Re: [Maria-developers] [Commits] Rev 3654: MDEV-4290: in file:///home/bell/maria/bzr/work-maria-5.3-MDEV-4290/
by Sergei Golubchik 13 May '13

13 May '13

Hi, Sanja! On May 03, sanja(a)askmonty.org wrote: > revno: 3654 > revision-id: sanja(a)askmonty.org-20130503130713-0n5lgrniay7n3jua > parent: igor(a)askmonty.org-20130430033140-qhx9wsgc5w2dx14r > committer: sanja(a)askmonty.org > branch nick: work-maria-5.3-MDEV-4290 > timestamp: Fri 2013-05-03 16:07:13 +0300 > message: > MDEV-4290: > Fix agregate function resolution in derived tables (no name > resolution over a derived table border) ok to push Regards, Sergei

1 0

[Maria-developers] An invalid memory access bug in HandlerSocket
by ahiguti100＠gmail.com 10 May '13

10 May '13

Hi MariaDB developers, I found an invalid memory access bug of HandlerSocket is not yet fixed in MariaDB. https://github.com/DeNA/HandlerSocket-Plugin-for-MySQL/issues/78 This bug can be reproduced by running the following script: https://github.com/DeNA/HandlerSocket-Plugin-for-MySQL/blob/master/regtest/… You may need to run mysqld with valgrind to observe the invalid memory access, because hitting this bug may not cause mysqld to crash. I've filed a merge request. https://code.launchpad.net/~ahiguti100/maria/handlersocket-fix-78/+merge/16… Rev #3750 fixes the bug, and #3751 updates documents and tests for HandlerSocket. I want docs to be updated because my github url for HandlerSocket has been changed. All the changes are copied from my github repo, and are provided under the BSD license. Please let me know if something is wrong with the merge request. Thanks, Akira

3 3

[Maria-developers] Updated myisamchk to handle LARGE files
by Michael Widenius 10 May '13

10 May '13

Hi! Here is a patch that should fix your myisamchk issues. This is to be applied to MariaDB 5.5.30 source. Can you verify that this works so I can add it to MariaDB 10.0 ? Some notes: - If you run repair with two --force options, like this: myisamchk --repair --force --force ... Then myisamchk will use the full myisam_sort_buffer_size for keys. Other buffers will be allocated as needed. This means that with this option one will never get the "myisam_sort_buffer_size is too small" error. I also added the option --create-missing-keys, which will create keys just for any disabled keys. Other things: - Better error message for 'myisam_sort_buffer_size is too small' errors. - Somewhat larger buffers as default. - Changed a lot of variables from uint -> ulonglong to try to ensure that big files will not cause any problems. - Removed some checks that limited the sort_buffer_size to 4G. - Fixed code that would cause problems if sort_buffer_size would be > 8G - Fixed printing of longlong variables. - Fixed possible core dump if 'alloc_dynamic()' would fail. - Fixed some possible problems on system where uint would be <= ulong. Hope this helps! Regards, Monty

2 2

Re: [Maria-developers] GSoC projects related on data replication.
by Kristian Nielsen 09 May '13

09 May '13

QIU Shuang <qiush.summer(a)gmail.com> writes: > A quick question: are the codes related to replication in the path `sql/rpl_*' and `sql/repl_*'? Much of the replication code is in: sql/log_event.* sql/slave.* sql/log.* sql/sql_repl.* sql/rpl_* sql/repl* Hope this helps, - Kristian.

1 0

[Maria-developers] how to use TC_LOG_MMAP in MySQL
by nanyi607rao 09 May '13

09 May '13

hi , In http://bazaar.launchpad.net/~maria-captains/maria/10.0-mdev452/revision/3434 you make some change to the class TC_LOG_MMAP,and I want to debug those codes to see how it works.but i don't understand in what conditions will MySQL use the TC_LOG_MMAP.it only use TC_LOG_DUMMY or MYSQL_BIN_LOG depends on the value of opt_bin_log.so how will it use the TC_LOG_MMAP in MySQL. 2013-05-09 nanyi607rao

3 2

[Maria-developers] GTID and failovers with multi-domain replication
by Pavel Ivanov 08 May '13

08 May '13

Kristian, You've mentioned that in future implementation of parallel replication MariaDB will use multiple domains to replicate from one master to a slave. It means for me that I need to understand how multiple domains will work. And I realized that I don't know how the following situation can be handled. I don't know how parallel replication will work, so I'll explain situation in terms of several masters. Let's assume we have server S1 that is master working with domain_id=0, server S2 is master working with domain_id=1, servers S3 and S4 are slaves and replicate from both of these masters, i.e. they have both domains in their databases. Now let's say S1 has last GTID 0-1-100, S2 has last GTID 1-2-100. Before S3 and S4 were able to fully catch up with S1 and S2 power got cut out from S1 and S2. As replication from two masters goes independently it's possible that S3 will have last transactions 0-1-100, 1-2-99 while S4 will have last transactions 0-1-99, 1-2-100. As my masters are out I want either S3 or S4 to temporarily become master. But it looks like I won't be able to do so: S3 won't connect to S4 because S4 doesn't have 0-1-100 and S4 won't connect to S3 because S3 doesn't have 1-2-100. Ideally I'd want for S3 to replicate from S4 in domain 1 and S4 to replicate from S3 in domain 0, and when they are equal in their position I can declare one of them master for both domains. But it looks like there are no tools to do such operation. How would you suggest to resolve such situation? Thank you, Pavel

2 3

[Maria-developers] Documentation about GTID
by Giuseppe Maxia 08 May '13

08 May '13

Hi! I am testing MariaDB 10.0.2, and I am interested in GTID. The only documentation I found is the worklog. Without docs at hand, I tried something: slave1 [localhost] {msandbox} ((none)) > CHANGE MASTER TO master_gtid_pos=''; ERROR 1064 (42000): You have an error in your SQL syntax; check the manual that corresponds to your MariaDB server version for the right syntax to use near 'master_gtid_pos=''' at line 1 slave1 [localhost] {msandbox} ((none)) > CHANGE MASTER TO master_gtid_pos=auto; ERROR 1064 (42000): You have an error in your SQL syntax; check the manual that corresponds to your MariaDB server version for the right syntax to use near 'master_gtid_pos=auto' at line 1 This is supposed to work, according to a blog post and some worklogs, but it doesn't. As for GTID in practice, it seems to be active. slave1 [localhost] {msandbox} ((none)) > show variables like '%gtid%'; +----------------+--------+ | Variable_name | Value | +----------------+--------+ | gtid_domain_id | 0 | | gtid_pos | 0-1-11 | | gtid_seq_no | 0 | +----------------+--------+ 3 rows in set (0.00 sec) slave1 [localhost] {msandbox} ((none)) > use mysql Reading table information for completion of table and column names You can turn off this feature to get a quicker startup with -A Database changed slave1 [localhost] {msandbox} (mysql) > select * from rpl_slave_state; +-----------+--------+-----------+--------+ | domain_id | sub_id | server_id | seq_no | +-----------+--------+-----------+--------+ | 0 | 12 | 1 | 11 | +-----------+--------+-----------+--------+ However, when I run "show slave status", I get: Using_Gtid: 0 Could you provide a pointer to the docs of the expected behavior at this stage? Thanks Giuseppe

9 35