- developers - lists.mariadb.org

[Maria-developers] ASAN segfaults on vcol case
by Aleksey Midenkov 18 Mar '19

18 Mar '19

This case segfaults in ASAN code: delimiter ~~ create or replace procedure t1() begin create or replace table t1 ( b tinytext, v text as (b) virtual ); end~~ delimiter ; call t1; Both in 10.3 and 10.4 in different places. Does someone know about such ASAN faults? -- All the best, Aleksey Midenkov @midenok

2 2

Re: [Maria-developers] 21a15ea673c: MDEV-16240: Assertion `0' failed in row_sel_convert_mysql_key_to_innobase
by Sergei Golubchik 18 Mar '19

18 Mar '19

Hi, Oleksandr! On Mar 18, Oleksandr Byelkin wrote: > revision-id: 21a15ea673c (mariadb-10.2.22-72-g21a15ea673c) > parent(s): bb8c82c66ab > author: Oleksandr Byelkin <sanja(a)mariadb.com> > committer: Oleksandr Byelkin <sanja(a)mariadb.com> > timestamp: 2019-03-14 15:12:47 +0100 > message: > > MDEV-16240: Assertion `0' failed in row_sel_convert_mysql_key_to_innobase > > Set table in row ID position mode before using this function. > > diff --git a/mysql-test/r/multi_update_innodb.result b/mysql-test/r/multi_update_innodb.result > index 5890fd24f5f..294ebfcebdf 100644 > --- a/mysql-test/r/multi_update_innodb.result > +++ b/mysql-test/r/multi_update_innodb.result > @@ -151,3 +151,43 @@ create table t2 like t1; > insert into t2 select * from t1; > delete t1,t2 from t2,t1 where t1.a<'B' and t2.b=t1.b; > drop table t1,t2; > +# > +# MDEV-16240: Assertion `0' failed in > +# row_sel_convert_mysql_key_to_innobase > +# > +SET @save_sql_mode=@@sql_mode; > +set sql_mode=''; > +CREATE TABLE `t3` ( > +`f1` TIMESTAMP NOT NULL DEFAULT '0000-00-00 00:00:00' ON UPDATE current_timestamp(), > +`f2` datetime DEFAULT '2000-01-01 00:00:00' ON UPDATE current_timestamp(), > +`f3` TIMESTAMP NULL DEFAULT '2000-01-01 00:00:00', > +`pk` datetime NOT NULL DEFAULT '0000-00-00 00:00:00', > +`f4` datetime DEFAULT current_timestamp(), > +PRIMARY KEY (`pk`), > +UNIQUE KEY `f2k` (`f2`), > +KEY `f4k` (`f4`) > +) ENGINE=InnoDB; > +INSERT INTO `t3` VALUES ('2018-05-18 15:08:07','2018-05-18 17:08:07','0000-00-00 00:00:00','0000-00-00 00:00:00','0000-00-00 00:00:00'),('0000-00-00 00:00:00','0000-00-00 00:00:00','1999-12-31 23:00:00','2002-07-03 23:04:40','0000-00-00 00:00:00'); > +CREATE VIEW `v1` AS > +SELECT `t3`.`pk` AS `pk`, > +`t3`.`f3` AS `f3`, > +`t3`.`f4` AS `f4`, > +`t3`.`f2` AS `f2`, > +`t3`.`f1` AS `f1` > +FROM `t3`; > +CREATE TABLE `t4` ( > +`f1` datetime DEFAULT current_timestamp() ON UPDATE current_timestamp(), > +`f3` timestamp NULL DEFAULT NULL, > +`f2` timestamp NULL DEFAULT '1999-12-31 23:00:00' ON UPDATE current_timestamp(), > +`pk` int(11) NOT NULL, > +`f4` timestamp NOT NULL DEFAULT current_timestamp() ON UPDATE current_timestamp(), > +PRIMARY KEY (`pk`) > +) ENGINE=InnoDB; > +INSERT INTO `t4` VALUES ('2018-05-18 17:08:06','0000-00-00 00:00:00',NULL,1,'2018-05-18 15:08:06'),('2018-05-18 17:08:06','0000-00-00 00:00:00',NULL,2,'2018-05-18 15:08:06'),('2018-05-18 17:08:06','0000-00-00 00:00:00',NULL,3,'2018-05-18 15:08:06'),('0000-00-00 00:00:00','0000-00-00 00:00:00',NULL,1976,'0000-00-00 00:00:00'),('2018-05-18 17:08:06','0000-00-00 00:00:00',NULL,2000,'2018-05-18 15:08:06'),('2018-05-18 17:08:06','0000-00-00 00:00:00',NULL,2001,'2018-05-18 15:08:06'),('2018-05-18 17:08:06','0000-00-00 00:00:00',NULL,2002,'2018-05-18 15:08:06'),('2018-05-18 17:08:06','0000-00-00 00:00:00',NULL,2003,'2018-05-18 15:08:06'),('2018-05-18 17:08:06','0000-00-00 00:00:00',NULL,2004,'2018-05-18 15:08:06'),('2018-05-18 17:08:06','0000-00-00 00:00:00','2018-05-18 15:08:06',2005,'2018-05-18 15:08:06'),('2018-05-18 17:08:06','0000-00-00 00:00:00','2018-05-18 15:08:06',2018,'2018-05-18 15:08:06'),('2018-05-18 17:08:06','0000-00-00 00:00:00','2018-05-18 15:08:06',2019,'2018-05-18 15:08:06'),('2018-05-18 17:08:06','0000-00-00 00:00:00','2018-05-18 15:08:06',2024,'2018-05-18 15:08:06'),('2018-05-18 17:08:06','0000-00-00 00:00:00','1999-12-31 23:00:00',2025,'2018-05-18 15:08:06'),('0000-00-00 00:00:00',NULL,'2018-05-18 15:08:06',2026,'2018-05-18 15:08:06'),('2018-05-18 17:08:07','0000-00-00 00:00:00','0000-00-00 00:00:00',2027,'0000-00-00 00:00:00'); > +UPDATE `v1` t1, `t4` t2 > +SET t1.`f2` = 6452736 WHERE t1.`f4` = 6272000; > +ERROR 23000: Duplicate entry '0000-00-00 00:00:00' for key 'f2k' actually, when I wrote "please, rename all columns in the test to have unique names" I meant _all_ columns, that is, in the t3 columnd could've been f1-f5, in v1 they could be called f6-f10, in t4 - f11-f15. but don't bother renaming them now, apparently the most important name was 'f2k' and it's unique now, so it's sufficiently clear, thanks. Still, for the future, it's generally helpful to have all names different to be able to understand what the error message refers to. > +DROP VIEW v1; > +DROP TABLE t3,t4; > +SET @@sql_mode=@save_sql_mode; > +# End of 10.2 tests > diff --git a/sql/handler.cc b/sql/handler.cc > index ae63640ae5c..d307433cf5d 100644 > --- a/sql/handler.cc > +++ b/sql/handler.cc > @@ -2648,8 +2648,7 @@ int handler::ha_rnd_pos(uchar *buf, uchar *pos) > DBUG_ENTER("handler::ha_rnd_pos"); > DBUG_ASSERT(table_share->tmp_table != NO_TMP_TABLE || > m_lock_type != F_UNLCK); > - /* TODO: Find out how to solve ha_rnd_pos when finding duplicate update. */ > - /* DBUG_ASSERT(inited == RND); */ > + DBUG_ASSERT(inited == RND); > > TABLE_IO_WAIT(tracker, m_psi, PSI_TABLE_FETCH_ROW, MAX_KEY, 0, > { result= rnd_pos(buf, pos); }) > @@ -3307,6 +3306,10 @@ void handler::get_auto_increment(ulonglong offset, ulonglong increment, > ulonglong nr; > int error; > MY_BITMAP *old_read_set; > + bool rnd_inited= (inited == RND); > + when is that true? > diff --git a/sql/item_subselect.cc b/sql/item_subselect.cc > index 8a9dd083911..cc1649b3c92 100644 > --- a/sql/item_subselect.cc > +++ b/sql/item_subselect.cc > @@ -5855,12 +5855,17 @@ Ordered_key::cmp_keys_by_row_data_and_rownum(Ordered_key *key, > } > > > -void Ordered_key::sort_keys() > +bool Ordered_key::sort_keys() > { > + /* Prepare table for random positioning (importent for innodb) */ 1) typo: "important" but 2) I don't think you need a comment here at all now. There's an assert in ha_rnd_pos(), it's rather more clear and more reliable anyway. > + if (tbl->file->ha_rnd_init(0)) > + return TRUE; > my_qsort2(key_buff, (size_t) key_buff_elements, sizeof(rownum_t), > (qsort2_cmp) &cmp_keys_by_row_data_and_rownum, (void*) this); > /* Invalidate the current row position. */ > cur_key_idx= HA_POS_ERROR; > + tbl->file->ha_rnd_end(); > + return FALSE; > } > > > diff --git a/sql/sql_insert.cc b/sql/sql_insert.cc > index 307473cecbd..5593580dbca 100644 > --- a/sql/sql_insert.cc > +++ b/sql/sql_insert.cc > @@ -3177,6 +3193,10 @@ bool Delayed_insert::handle_inserts(void) > max_rows= ULONG_MAX; // Do as much as possible > } > > + /* Prepare table for random positioning (importent for innodb) */ same > + if (table->file->ha_rnd_init(0)) > + goto err; what's that for? duplicate key error message? > + > /* > We can't use row caching when using the binary log because if > we get a crash, then binary log will contain rows that are not yet > @@ -3649,7 +3671,12 @@ select_insert::prepare(List<Item> &values, SELECT_LEX_UNIT *u) > > thd->cuted_fields=0; > if (info.ignore || info.handle_duplicates != DUP_ERROR) > + { > table->file->extra(HA_EXTRA_IGNORE_DUP_KEY); > + if (table->file->ha_table_flags() & HA_DUPLICATE_POS && > + table->file->ha_rnd_init(0)) > + DBUG_RETURN(1); here and in other places, you don't do my_error(). Is there some catch-all my_error somewhere up the stack or it'll be the usual assert in Diagnostic_area? > + } > if (info.handle_duplicates == DUP_REPLACE && > (!table->triggers || !table->triggers->has_delete_triggers())) > table->file->extra(HA_EXTRA_WRITE_CAN_REPLACE); > @@ -3929,6 +3959,11 @@ void select_insert::abort_result_set() { > if (thd->locked_tables_mode <= LTM_LOCK_TABLES) > table->file->ha_end_bulk_insert(); > > + if (table->file->inited) why not "if (table->file->ha_table_flags() & HA_DUPLICATE_POS)" ? > + table->file->ha_rnd_end(); > + table->file->extra(HA_EXTRA_NO_IGNORE_DUP_KEY); > + table->file->extra(HA_EXTRA_WRITE_CANNOT_REPLACE); > + > /* > If at least one row has been inserted/modified and will stay in > the table (the table doesn't have transactions) we must write to > diff --git a/sql/sql_load.cc b/sql/sql_load.cc > index 8c2f17dac3f..3283b962f65 100644 > --- a/sql/sql_load.cc > +++ b/sql/sql_load.cc > @@ -638,6 +638,13 @@ int mysql_load(THD *thd,sql_exchange *ex,TABLE_LIST *table_list, > > thd->abort_on_warning= !ignore && thd->is_strict_mode(); > > + /* Prepare table for random positioning (importent for innodb) */ same > + if ((table_list->table->file->ha_table_flags() & HA_DUPLICATE_POS) && > + table_list->table->file->ha_rnd_init(0)) > + { > + error= 1; > + goto err; > + } > thd_progress_init(thd, 2); > if (table_list->table->validate_default_values_of_unset_fields(thd)) > { > diff --git a/sql/sql_select.cc b/sql/sql_select.cc > index 91a6445c870..4c401ea9011 100644 > --- a/sql/sql_select.cc > +++ b/sql/sql_select.cc > @@ -20317,6 +20317,12 @@ end_unique_update(JOIN *join, JOIN_TAB *join_tab __attribute__((unused)), > table->file->print_error(error,MYF(0)); /* purecov: inspected */ > DBUG_RETURN(NESTED_LOOP_ERROR); /* purecov: inspected */ > } > + /* Prepare table for random positioning */ > + bool rnd_inited= (table->file->inited == handler::RND); when is that true? and when false? > + if (!rnd_inited && > + ((error= table->file->ha_index_end()) || > + (error= table->file->ha_rnd_init(0)))) > + DBUG_RETURN(NESTED_LOOP_ERROR); > if (table->file->ha_rnd_pos(table->record[1],table->file->dup_ref)) > { > table->file->print_error(error,MYF(0)); /* purecov: inspected */ > diff --git a/sql/sql_update.cc b/sql/sql_update.cc > index 11ffa684216..6d4c11d494a 100644 > --- a/sql/sql_update.cc > +++ b/sql/sql_update.cc > @@ -211,6 +211,10 @@ static void prepare_record_for_error_message(int error, TABLE *table) > bitmap_union(table->read_set, &unique_map); > /* Tell the engine about the new set. */ > table->file->column_bitmaps_signal(); > + /* Prepare table for random positioning (importent for innodb) */ same > + if (table->file->ha_index_or_rnd_end() || > + table->file->ha_rnd_init(0)) > + DBUG_VOID_RETURN; > /* Read record that is identified by table->file->ref. */ > (void) table->file->ha_rnd_pos(table->record[1], table->file->ref); > /* Copy the newly read columns into the new record. */ > Regards, Sergei Chief Architect MariaDB and security(a)mariadb.org

1 0

Re: [Maria-developers] 34098adf32b: Get rid of rea_create_table()
by Sergei Golubchik 14 Mar '19

14 Mar '19

Hi, Sergey! On Mar 13, Sergey Vojtovich wrote: > revision-id: 34098adf32b (mariadb-10.4.1-138-g34098adf32b) > parent(s): d953dc02d9a > author: Sergey Vojtovich <svoj(a)mariadb.org> > committer: Sergey Vojtovich <svoj(a)mariadb.org> > timestamp: 2019-02-07 20:04:51 +0400 > message: > > Get rid of rea_create_table() > > Moved rea_create_table() to the sole caller. > > Visible difference: regular ALTER TABLE doesn't create .frm for temporary > and discoverable tables anymore. > > Also ha_create_partitioning_metadata(CHF_CREATE_FLAG) does cleanup on > error now. > > Part of MDEV-17805 - Remove InnoDB cache for temporary tables. ok to push Regards, Sergei Chief Architect MariaDB and security(a)mariadb.org

1 0

Re: [Maria-developers] d953dc02d9a: Fix inplace ALTER TABLE to not register tmp table
by Sergei Golubchik 14 Mar '19

14 Mar '19

Hi, Sergey! On Mar 13, Sergey Vojtovich wrote: > revision-id: d953dc02d9a (mariadb-10.4.1-137-gd953dc02d9a) > parent(s): 9496394002f > author: Sergey Vojtovich <svoj(a)mariadb.org> > committer: Sergey Vojtovich <svoj(a)mariadb.org> > timestamp: 2019-02-07 20:04:38 +0400 > message: > > Fix inplace ALTER TABLE to not register tmp table > > Do not register intermediate tables created by inplace ALTER TABLE in > THD::temporary_tables. > > For inplace ALTER TABLE moved .frm creation to > create_table_for_inplace_alter(). For regular ALTER TABLE it is (and was > redundantly) created by ha_create_table(). > > Removed open_in_engine argument of create_and_open_tmp_table() and > open_temporary_table(): it became unused after this patch. > > Part of MDEV-17805 - Remove InnoDB cache for temporary tables. Looks ok code wise. But I don't see what are you gaining by this transformation. Regards, Sergei Chief Architect MariaDB and security(a)mariadb.org

1 0

[Maria-developers] Self Introduction (gsoc)
by shreya pohekar 14 Mar '19

14 Mar '19

Hey! I am Shreya Pohekar. I am currently in my 3rd year of BE in Computer Science and Engineering from University Institute of Technology, RGPV Bhopal. I have a conceptual proficiency on Java and C++. On scrolling through the projects of mariadb, i found most of the projects interesting. But i would like to contribute to the project, "Implement multi-table UPDATE/DELETE returning a result set".@Igor Babaev, If you can guide me on where to get started, that would be great. Looking forward to hearing from you. Thank you.

1 0

[Maria-developers] Self Introduction (gsoc)
by shreya pohekar 14 Mar '19

14 Mar '19

Hey! I am Shreya Pohekar. I am currently in my 3rd year of BE in Computer Science and Engineering from University Institute of Technology, RGPV Bhopal. I have a conceptual proficiency on Java and C++. On scrolling through the projects of mariadb, i found most of the projects interesting. But i would like to contribute to the project, "Implement multi-table UPDATE/DELETE returning a result set". If you can guide me on where to get started, that would be great. Looking forward to hearing from you. Thank you.

1 0

Re: [Maria-developers] 9496394002f: Removed redundant partitioning check
by Sergei Golubchik 14 Mar '19

14 Mar '19

Hi, Sergey! On Mar 13, Sergey Vojtovich wrote: > revision-id: 9496394002f (mariadb-10.4.1-136-g9496394002f) > parent(s): 311beaffa63 > author: Sergey Vojtovich <svoj(a)mariadb.org> > committer: Sergey Vojtovich <svoj(a)mariadb.org> > timestamp: 2019-02-06 13:56:42 +0400 > message: > > Removed redundant partitioning check > > This check was introduced in 602a222 and then became redundant in ad1553e, > where we attempt to open a table even for non-copy algorithms. > > Added missing test case from 602a222. > > Part of MDEV-17805 - Remove InnoDB cache for temporary tables. ok to push Regards, Sergei Chief Architect MariaDB and security(a)mariadb.org

1 0

Re: [Maria-developers] 311beaffa63: Simplified dd_recreate_table()
by Sergei Golubchik 14 Mar '19

14 Mar '19

Hi, Sergey! On Mar 13, Sergey Vojtovich wrote: > revision-id: 311beaffa63 (mariadb-10.4.1-135-g311beaffa63) > parent(s): 3323b013c13 > author: Sergey Vojtovich <svoj(a)mariadb.org> > committer: Sergey Vojtovich <svoj(a)mariadb.org> > timestamp: 2019-02-06 13:56:42 +0400 > message: > > Simplified dd_recreate_table() > > It is used only with persistent tables, so remove path argument, which > was introduced in 9594107f and became useless after ce6a63e. > > Part of MDEV-17805 - Remove InnoDB cache for temporary tables. ok to push Regards, Sergei Chief Architect MariaDB and security(a)mariadb.org

1 0

Re: [Maria-developers] 3323b013c13: Removed redundant SE lock for tmp tables
by Sergei Golubchik 14 Mar '19

14 Mar '19

Hi, Sergey! On Mar 13, Sergey Vojtovich wrote: > revision-id: 3323b013c13 (mariadb-10.4.1-134-g3323b013c13) > parent(s): 4b3656a44d6 > author: Sergey Vojtovich <svoj(a)mariadb.org> > committer: Sergey Vojtovich <svoj(a)mariadb.org> > timestamp: 2019-02-06 13:56:28 +0400 > message: > > Removed redundant SE lock for tmp tables > > CREATE TEMPORARY TABLE locks SE plugin 6 times. 5 of these locks are > released by the end of the statement. And only 1 acquired by > init_from_binary_frm_image() / plugin_lock() remains. > > The lock removed in this patch was clearly redundant. > > Part of MDEV-17805 - Remove InnoDB cache for temporary tables. > > diff --git a/sql/temporary_tables.cc b/sql/temporary_tables.cc > index 917a85e6c3b..438b670b5f5 100644 > --- a/sql/temporary_tables.cc > +++ b/sql/temporary_tables.cc > @@ -951,8 +947,6 @@ TMP_TABLE_SHARE *THD::create_temporary_table(handlerton *hton, > init_tmp_table_share(this, share, saved_key_cache, key_length, > strend(saved_key_cache) + 1, tmp_path); > > - share->db_plugin= ha_lock_engine(this, hton); > - I suppose, share->db_plugin will be set in the following share->init_from_binary_frm_image ? Ok to push. Regards, Sergei Chief Architect MariaDB and security(a)mariadb.org

1 0

Re: [Maria-developers] 14c2f90ad08: Idempotent INSERT events for system versioning
by Aleksey Midenkov 13 Mar '19

13 Mar '19

Hi, Sergei, Thank you for observations! This task is in progress. While doing it I found out that RBR doesn't replicate timestamp fractions. That is: it gets seconds from event timestamp marker and adds fractions from system time. This causes bugs in System Versioning. More on it in Bug 6 here: https://github.com/tempesta-tech/mariadb/issues/578#issuecomment-470533050 I propose to fix it by adding timestamp fractions field to Log_event header after FLAGS_OFFSET. To support backwards compatibility it is needed to increment fdle->binlog_version and probably add new flag LOG_EVENT_HAS_SEC_PART_F (for support sending Log_event without fractions). Do you agree? On Thu, Jan 31, 2019 at 1:10 PM Sergei Golubchik <serg(a)mariadb.com> wrote: > Hi, Aleksey! > > On Jan 31, Aleksey Midenkov wrote: > > revision-id: 14c2f90ad08 (versioning-1.0.7-1-g14c2f90ad08) > > parent(s): a8efe7ab1f2 > > author: Aleksey Midenkov <midenok(a)gmail.com> > > committer: Aleksey Midenkov <midenok(a)gmail.com> > > timestamp: 2018-11-17 16:30:10 +0300 > > message: > > > > Idempotent INSERT events for system versioning > > > > Case 1: Rows_log_event::write_row() always overwrite historical row. > > > > Related to MDEV-16370. > > I realized that the whole RBR logging of system versioned tables is > wrong. > > For an update you log update_row event and write_row event separately > and execute them separately. It allows trivially to manipulate the history. > See a test case attached. > > To fix this, the slave should ignore all events that modify the history. > And generate them locally. Say, an update on the master generates > update_row and write_row events. The slave executes update_row event > which updates a current row and this also should generate the historical > row. Then the slave ignores write_row event. > > Of course, as an optimization, the master should not generate historical > write_row events, but as my test case shows, they can be forged, so the > slave should ignore them anyway, if they happen to come. > > I've created a new https://jira.mariadb.org/browse/MDEV-18432 for that, > let's not have it tied to MDEV-16370. > > Regards, > Sergei > Chief Architect MariaDB > and security(a)mariadb.org > -- All the best, Aleksey Midenkov @midenok

3 7