- developers - lists.mariadb.org

[Maria-developers] bzr commit into MariaDB 5.1, with Maria 1.5:maria branch (knielsen:2731) Bug#46640
by knielsen＠knielsen-hq.org 09 Sep '09

09 Sep '09

#At lp:maria 2731 knielsen(a)knielsen-hq.org 2009-09-09 Bug#46640: output from mysqlbinlog command in 5.1 breaks replication The BINLOG statement was sharing too much code with the slave SQL thread, introduced with the patch for Bug#32407. This caused statements to be logged with the wrong server_id, the id stored inside the events of the BINLOG statement rather than the id of the running server. Fix by rearranging code a bit so that only relevant parts of the code are executed by the BINLOG statement. modified: mysql-test/extra/binlog_tests/binlog.test mysql-test/suite/binlog/r/binlog_row_binlog.result mysql-test/suite/binlog/r/binlog_stm_binlog.result sql/log_event.cc sql/slave.cc sql/slave.h sql/sql_binlog.cc per-file messages: mysql-test/extra/binlog_tests/binlog.test Add test case. mysql-test/suite/binlog/r/binlog_row_binlog.result Add test case. mysql-test/suite/binlog/r/binlog_stm_binlog.result Add test case. sql/log_event.cc Move rows_event_stmt_clean() call from update_pos() to apply_event(). This in any case makes more sense, and is needed as update_pos() is no longer called when executing BINLOG statements. sql/slave.cc The skip flag is no longer needed, as the code path for BINLOG statement has been cleaned up. sql/slave.h The skip flag is no longer needed, as the code path for BINLOG statement has been cleaned up. sql/sql_binlog.cc Don't invoke the update_pos() code path for the BINLOG statement, as it contains code that is redundant and/or harmful (especially setting thd->server_id). === modified file 'mysql-test/extra/binlog_tests/binlog.test' --- a/mysql-test/extra/binlog_tests/binlog.test 2009-03-25 19:41:16 +0000 +++ b/mysql-test/extra/binlog_tests/binlog.test 2009-09-09 15:27:23 +0000 @@ -258,3 +258,39 @@ dec $it; } show master status /* must show new binlog index after rotating */; drop table t3; + + +# +# Bug#46640: output from mysqlbinlog command in 5.1 breaks replication +# + +connect (fresh,localhost,root,,test); +connection fresh; + +RESET MASTER; +CREATE TABLE t1 (a INT PRIMARY KEY); + +# Format description event, with server_id = 10; +BINLOG ' +3u9kSA8KAAAAZgAAAGoAAAABAAQANS4xLjM1LW1hcmlhLWJldGExLWRlYnVnLWxvZwAAAAAAAAAA +AAAAAAAAAAAAAAAAAADe72RIEzgNAAgAEgAEBAQEEgAAUwAEGggAAAAICAgC +'; + +# What server_id is logged for a statement? Should be our own, not the +# one from the format description event. +INSERT INTO t1 VALUES (1); + +# INSERT INTO t1 VALUES (2), with server_id=20. Check that this is logged +# with our own server id, not the 20 from the BINLOG statement. +BINLOG ' +3u9kSBMUAAAAKQAAAJEBAAAAABoAAAAAAAAABHRlc3QAAnQxAAEDAAA= +3u9kSBcUAAAAIgAAALMBAAAQABoAAAAAAAEAAf/+AgAAAA== +'; + +# Show binlog events to check that server ids are correct. +--replace_column 1 # 2 # 5 # +--replace_regex /Server ver: .*, Binlog ver: .*/Server ver: #, Binlog ver: #/ /table_id: [0-9]+/table_id: #/ +SHOW BINLOG EVENTS; + +DROP TABLE t1; +disconnect fresh; === modified file 'mysql-test/suite/binlog/r/binlog_row_binlog.result' --- a/mysql-test/suite/binlog/r/binlog_row_binlog.result 2009-03-25 19:41:16 +0000 +++ b/mysql-test/suite/binlog/r/binlog_row_binlog.result 2009-09-09 15:27:23 +0000 @@ -1298,3 +1298,27 @@ show master status /* must show new binl File Position Binlog_Do_DB Binlog_Ignore_DB master-bin.000002 106 drop table t3; +RESET MASTER; +CREATE TABLE t1 (a INT PRIMARY KEY); +BINLOG ' +3u9kSA8KAAAAZgAAAGoAAAABAAQANS4xLjM1LW1hcmlhLWJldGExLWRlYnVnLWxvZwAAAAAAAAAA +AAAAAAAAAAAAAAAAAADe72RIEzgNAAgAEgAEBAQEEgAAUwAEGggAAAAICAgC +'; +INSERT INTO t1 VALUES (1); +BINLOG ' +3u9kSBMUAAAAKQAAAJEBAAAAABoAAAAAAAAABHRlc3QAAnQxAAEDAAA= +3u9kSBcUAAAAIgAAALMBAAAQABoAAAAAAAEAAf/+AgAAAA== +'; +SHOW BINLOG EVENTS; +Log_name Pos Event_type Server_id End_log_pos Info +# # Format_desc 1 # Server ver: #, Binlog ver: # +# # Query 1 # use `test`; CREATE TABLE t1 (a INT PRIMARY KEY) +# # Query 1 # use `test`; BEGIN +# # Table_map 1 # table_id: # (test.t1) +# # Write_rows 1 # table_id: # flags: STMT_END_F +# # Query 1 # use `test`; COMMIT +# # Query 1 # use `test`; BEGIN +# # Table_map 1 # table_id: # (test.t1) +# # Write_rows 1 # table_id: # flags: STMT_END_F +# # Query 1 # use `test`; COMMIT +DROP TABLE t1; === modified file 'mysql-test/suite/binlog/r/binlog_stm_binlog.result' --- a/mysql-test/suite/binlog/r/binlog_stm_binlog.result 2009-03-25 19:41:16 +0000 +++ b/mysql-test/suite/binlog/r/binlog_stm_binlog.result 2009-09-09 15:27:23 +0000 @@ -773,3 +773,24 @@ show master status /* must show new binl File Position Binlog_Do_DB Binlog_Ignore_DB master-bin.000002 106 drop table t3; +RESET MASTER; +CREATE TABLE t1 (a INT PRIMARY KEY); +BINLOG ' +3u9kSA8KAAAAZgAAAGoAAAABAAQANS4xLjM1LW1hcmlhLWJldGExLWRlYnVnLWxvZwAAAAAAAAAA +AAAAAAAAAAAAAAAAAADe72RIEzgNAAgAEgAEBAQEEgAAUwAEGggAAAAICAgC +'; +INSERT INTO t1 VALUES (1); +BINLOG ' +3u9kSBMUAAAAKQAAAJEBAAAAABoAAAAAAAAABHRlc3QAAnQxAAEDAAA= +3u9kSBcUAAAAIgAAALMBAAAQABoAAAAAAAEAAf/+AgAAAA== +'; +SHOW BINLOG EVENTS; +Log_name Pos Event_type Server_id End_log_pos Info +# # Format_desc 1 # Server ver: #, Binlog ver: # +# # Query 1 # use `test`; CREATE TABLE t1 (a INT PRIMARY KEY) +# # Query 1 # use `test`; INSERT INTO t1 VALUES (1) +# # Query 1 # use `test`; BEGIN +# # Table_map 1 # table_id: # (test.t1) +# # Write_rows 1 # table_id: # flags: STMT_END_F +# # Query 1 # use `test`; COMMIT +DROP TABLE t1; === modified file 'sql/log_event.cc' --- a/sql/log_event.cc 2009-06-11 17:49:51 +0000 +++ b/sql/log_event.cc 2009-09-09 15:27:23 +0000 @@ -7427,6 +7427,13 @@ int Rows_log_event::do_apply_event(Relay const_cast<Relay_log_info*>(rli)->last_event_start_time= my_time(0); } + if (get_flags(STMT_END_F)) + { + error= rows_event_stmt_cleanup(rli, thd); + if (error) + DBUG_RETURN(error); + } + DBUG_RETURN(0); } @@ -7525,33 +7532,22 @@ Rows_log_event::do_update_pos(Relay_log_ if (get_flags(STMT_END_F)) { - if ((error= rows_event_stmt_cleanup(rli, thd)) == 0) - { - /* - Indicate that a statement is finished. - Step the group log position if we are not in a transaction, - otherwise increase the event log position. - */ - rli->stmt_done(log_pos, when); + /* + Indicate that a statement is finished. + Step the group log position if we are not in a transaction, + otherwise increase the event log position. + */ + rli->stmt_done(log_pos, when); - /* - Clear any errors pushed in thd->net.last_err* if for example "no key - found" (as this is allowed). This is a safety measure; apparently - those errors (e.g. when executing a Delete_rows_log_event of a - non-existing row, like in rpl_row_mystery22.test, - thd->net.last_error = "Can't find record in 't1'" and last_errno=1032) - do not become visible. We still prefer to wipe them out. - */ - thd->clear_error(); - } - else - { - rli->report(ERROR_LEVEL, error, - "Error in %s event: commit of row events failed, " - "table `%s`.`%s`", - get_type_str(), m_table->s->db.str, - m_table->s->table_name.str); - } + /* + Clear any errors pushed in thd->net.last_err* if for example "no key + found" (as this is allowed). This is a safety measure; apparently + those errors (e.g. when executing a Delete_rows_log_event of a + non-existing row, like in rpl_row_mystery22.test, + thd->net.last_error = "Can't find record in 't1'" and last_errno=1032) + do not become visible. We still prefer to wipe them out. + */ + thd->clear_error(); } else { === modified file 'sql/slave.cc' --- a/sql/slave.cc 2009-09-03 14:05:38 +0000 +++ b/sql/slave.cc 2009-09-09 15:27:23 +0000 @@ -1882,8 +1882,7 @@ static int has_temporary_error(THD *thd) @retval 2 No error calling ev->apply_event(), but error calling ev->update_pos(). */ -int apply_event_and_update_pos(Log_event* ev, THD* thd, Relay_log_info* rli, - bool skip) +int apply_event_and_update_pos(Log_event* ev, THD* thd, Relay_log_info* rli) { int exec_res= 0; @@ -1928,37 +1927,32 @@ int apply_event_and_update_pos(Log_event ev->when= my_time(0); ev->thd = thd; // because up to this point, ev->thd == 0 - if (skip) - { - int reason= ev->shall_skip(rli); - if (reason == Log_event::EVENT_SKIP_COUNT) - --rli->slave_skip_counter; - pthread_mutex_unlock(&rli->data_lock); - if (reason == Log_event::EVENT_SKIP_NOT) - exec_res= ev->apply_event(rli); + int reason= ev->shall_skip(rli); + if (reason == Log_event::EVENT_SKIP_COUNT) + --rli->slave_skip_counter; + pthread_mutex_unlock(&rli->data_lock); + if (reason == Log_event::EVENT_SKIP_NOT) + exec_res= ev->apply_event(rli); #ifndef DBUG_OFF - /* - This only prints information to the debug trace. + /* + This only prints information to the debug trace. - TODO: Print an informational message to the error log? - */ - static const char *const explain[] = { - // EVENT_SKIP_NOT, - "not skipped", - // EVENT_SKIP_IGNORE, - "skipped because event should be ignored", - // EVENT_SKIP_COUNT - "skipped because event skip counter was non-zero" - }; - DBUG_PRINT("info", ("OPTION_BEGIN: %d; IN_STMT: %d", - thd->options & OPTION_BEGIN ? 1 : 0, - rli->get_flag(Relay_log_info::IN_STMT))); - DBUG_PRINT("skip_event", ("%s event was %s", - ev->get_type_str(), explain[reason])); + TODO: Print an informational message to the error log? + */ + static const char *const explain[] = { + // EVENT_SKIP_NOT, + "not skipped", + // EVENT_SKIP_IGNORE, + "skipped because event should be ignored", + // EVENT_SKIP_COUNT + "skipped because event skip counter was non-zero" + }; + DBUG_PRINT("info", ("OPTION_BEGIN: %d; IN_STMT: %d", + thd->options & OPTION_BEGIN ? 1 : 0, + rli->get_flag(Relay_log_info::IN_STMT))); + DBUG_PRINT("skip_event", ("%s event was %s", + ev->get_type_str(), explain[reason])); #endif - } - else - exec_res= ev->apply_event(rli); DBUG_PRINT("info", ("apply_event error = %d", exec_res)); if (exec_res == 0) @@ -2078,7 +2072,7 @@ static int exec_relay_log_event(THD* thd delete ev; DBUG_RETURN(1); } - exec_res= apply_event_and_update_pos(ev, thd, rli, TRUE); + exec_res= apply_event_and_update_pos(ev, thd, rli); /* Format_description_log_event should not be deleted because it will be === modified file 'sql/slave.h' --- a/sql/slave.h 2008-12-10 09:02:25 +0000 +++ b/sql/slave.h 2009-09-09 15:27:23 +0000 @@ -190,8 +190,7 @@ int purge_relay_logs(Relay_log_info* rli void set_slave_thread_options(THD* thd); void set_slave_thread_default_charset(THD *thd, Relay_log_info const *rli); void rotate_relay_log(Master_info* mi); -int apply_event_and_update_pos(Log_event* ev, THD* thd, Relay_log_info* rli, - bool skip); +int apply_event_and_update_pos(Log_event* ev, THD* thd, Relay_log_info* rli); pthread_handler_t handle_slave_io(void *arg); pthread_handler_t handle_slave_sql(void *arg); === modified file 'sql/sql_binlog.cc' --- a/sql/sql_binlog.cc 2009-05-06 12:03:24 +0000 +++ b/sql/sql_binlog.cc 2009-09-09 15:27:23 +0000 @@ -56,17 +56,21 @@ void mysql_client_binlog_statement(THD* Format_description_event. */ my_bool have_fd_event= TRUE; - if (!thd->rli_fake) + int err; + Relay_log_info *rli; + + rli= thd->rli_fake; + if (!rli) { - thd->rli_fake= new Relay_log_info; + rli= thd->rli_fake= new Relay_log_info; #ifdef HAVE_valgrind - thd->rli_fake->is_fake= TRUE; + rli->is_fake= TRUE; #endif have_fd_event= FALSE; } - if (thd->rli_fake && !thd->rli_fake->relay_log.description_event_for_exec) + if (rli && !rli->relay_log.description_event_for_exec) { - thd->rli_fake->relay_log.description_event_for_exec= + rli->relay_log.description_event_for_exec= new Format_description_log_event(4); have_fd_event= FALSE; } @@ -78,16 +82,16 @@ void mysql_client_binlog_statement(THD* /* Out of memory check */ - if (!(thd->rli_fake && - thd->rli_fake->relay_log.description_event_for_exec && + if (!(rli && + rli->relay_log.description_event_for_exec && buf)) { my_error(ER_OUTOFMEMORY, MYF(0), 1); /* needed 1 bytes */ goto end; } - thd->rli_fake->sql_thd= thd; - thd->rli_fake->no_storage= TRUE; + rli->sql_thd= thd; + rli->no_storage= TRUE; for (char const *strptr= thd->lex->comment.str ; strptr < thd->lex->comment.str + thd->lex->comment.length ; ) @@ -170,8 +174,7 @@ void mysql_client_binlog_statement(THD* } ev= Log_event::read_log_event(bufptr, event_len, &error, - thd->rli_fake->relay_log. - description_event_for_exec); + rli->relay_log.description_event_for_exec); DBUG_PRINT("info",("binlog base64 err=%s", error)); if (!ev) @@ -209,9 +212,27 @@ void mysql_client_binlog_statement(THD* reporting. */ #if !defined(MYSQL_CLIENT) && defined(HAVE_REPLICATION) - if (apply_event_and_update_pos(ev, thd, thd->rli_fake, FALSE)) + err= ev->apply_event(rli); +#else + err= 0; +#endif + if (ev->get_type_code() == FORMAT_DESCRIPTION_EVENT) + { + /* + Install this as the format description event to use for any + following events. + */ + delete rli->relay_log.description_event_for_exec; + rli->relay_log.description_event_for_exec= + static_cast<Format_description_log_event *>(ev); + } + else { delete ev; + } + ev= 0; + if (err) + { /* TODO: Maybe a better error message since the BINLOG statement now contains several events. @@ -219,17 +240,6 @@ void mysql_client_binlog_statement(THD* my_error(ER_UNKNOWN_ERROR, MYF(0), "Error executing BINLOG statement"); goto end; } -#endif - - /* - Format_description_log_event should not be deleted because it - will be used to read info about the relay log's format; it - will be deleted when the SQL thread does not need it, - i.e. when this thread terminates. - */ - if (ev->get_type_code() != FORMAT_DESCRIPTION_EVENT) - delete ev; - ev= 0; } } @@ -238,7 +248,7 @@ void mysql_client_binlog_statement(THD* my_ok(thd); end: - thd->rli_fake->clear_tables_to_lock(); + rli->clear_tables_to_lock(); my_free(buf, MYF(MY_ALLOW_ZERO_PTR)); DBUG_VOID_RETURN; }

1 0

[Maria-developers] Updated (by Guest): mysqlbinlog: remove undesired effects of format description binlog statement (50)
by worklog-noreply＠askmonty.org 09 Sep '09

09 Sep '09

----------------------------------------------------------------------- WORKLOG TASK -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- TASK...........: mysqlbinlog: remove undesired effects of format description binlog statement CREATION DATE..: Mon, 17 Aug 2009, 14:12 SUPERVISOR.....: Monty IMPLEMENTOR....: COPIES TO......: CATEGORY.......: Server-RawIdeaBin TASK ID........: 50 (http://askmonty.org/worklog/?tid=50) VERSION........: Server-9.x STATUS.........: Un-Assigned PRIORITY.......: 60 WORKED HOURS...: 2 ESTIMATE.......: 0 (hours remain) ORIG. ESTIMATE.: 0 PROGRESS NOTES: -=-=(Guest - Wed, 09 Sep 2009, 11:50)=-=- High Level Description modified. --- /tmp/wklog.50.old.11584 2009-09-09 11:50:12.000000000 +0300 +++ /tmp/wklog.50.new.11584 2009-09-09 11:50:12.000000000 +0300 @@ -5,7 +5,11 @@ This will cause an error when one tires to apply mysqlbinlog output to a 5.0.x server. -2. When one applies "format description binlog statement" at the slave, it -will change the slave's server_id when applied. +2. When one executes a BINLOG statement (containing "format description binlog +statement" or other events), this will cause subsequent statements run in +the same session to be binlogged with the server id of the last event +executed, not with the real server id. This can cause problems like infinite +recursive application in a circular replication topology. This WL is to fix these issues. + -=-=(Knielsen - Tue, 08 Sep 2009, 15:07)=-=- Low Level Design modified. --- /tmp/wklog.50.old.19208 2009-09-08 15:07:09.000000000 +0300 +++ /tmp/wklog.50.new.19208 2009-09-08 15:07:09.000000000 +0300 @@ -1 +1,34 @@ +I think the fix for point 2 is to replace the call to +apply_event_and_update_pos() in mysql_client_binlog_statement() with a direct +call to ev->apply_event(). However, I need to check two things to be sure this +is correct: + +1. The existing code does + + if (!ev->when) ev->when= my_time(0) + +Need to understand if this is needed/correct or not. + +2. The existing code does ev->update_pos(). I think this is redundant for +BINLOG statement (as it uses a fake rli structure), but I need to check to +make sure. + +Once this is done, point 1 may no longer be needed. The user can use +--base64-output=never when applying the output to a 5.0 server, and omit that +option when applying the output to a 5.1 server. There should be no need to +omit the format description event in the output when other BINLOG statements +for row events are present, as it no longer changes the server id. In fact the +format description event is required to be able to execute other BINLOG +statements, as its purpose is to define the binary format of the events +contained in these statements. + +Alternatively, we could implement that if the format description event of the +source binlog has server version < 5.1, and --base64-output=auto (default), +then the format description event is omitted (and should any BINLOG statement +be needed, unlikely as it is from a 5.0 server, we will need to throw an +error). + +The binlog format version for 5.0 and 5.1 is actually the same, hence the need +to look at server version to guess if the format description event can be +omitted. -=-=(Knielsen - Tue, 08 Sep 2009, 14:29)=-=- High-Level Specification modified. --- /tmp/wklog.50.old.17531 2009-09-08 14:29:30.000000000 +0300 +++ /tmp/wklog.50.new.17531 2009-09-08 14:29:30.000000000 +0300 @@ -12,3 +12,37 @@ One question that one needs to sort out before disabling server_id change is why it was put there in the first place? Should it be always removed? +First, the problem is not the format description log event. In fact all log +events have their own value of server_id, and every event in the BINLOG +statement is executed with its own id as server_id. + +It seems it was introduced deliberately in the patch for Bug#32407. This patch +makes sql thread and BINLOG statement share code path for executing the binary +event, even though the bug is really about something different (outputting +format description event to allow proper execution of other BINLOG +statements). + +Nevertheless, I think using the server_id in the events of a BINLOG statement +is wrong: + +1. It is different behaviour from executing normal statement-based events not +written using BINLOG. + +2. It causes the possibility for infinite cycle of replication in a cyclic +replication setup, if the server ids in BINLOG are different from all other +server ids in the cycle. + +3. The functionality of applying events with original server id from the event +is still available by pointing the slave thread to the binlog. The +functionality of applying row-based binlog events with server id of the server +executing the BINLOG statements is not otherwise available. + +In fact most of the code from the slave thread that is now also run when +executing BINLOG statements is redundant, or does the wrong thing (like +updating binlog position). + +Maybe there is a confusion on the purpose of the BINLOG statement. Is it to be +able to apply row-based events manually to a server (mysqlbinlog|mysql)? Or is +it to be able to simulate the effect of the slave sql thread? I think the +former is the most important one. + -=-=(Knielsen - Mon, 07 Sep 2009, 10:06)=-=- Research worklog, bugs, and existing server code. Worked 2 hours and estimate 0 hours remain (original estimate increased by 2 hours). -=-=(Psergey - Mon, 17 Aug 2009, 14:13)=-=- Dependency created: 39 now depends on 50 -=-=(Psergey - Mon, 17 Aug 2009, 14:13)=-=- High-Level Specification modified. --- /tmp/wklog.50.old.11389 2009-08-17 14:13:05.000000000 +0300 +++ /tmp/wklog.50.new.11389 2009-08-17 14:13:05.000000000 +0300 @@ -1 +1,14 @@ +First item +---------- +AFAIU what needs to be done is: +1. record a source server version (it is in the first binlog event). +2. don't emit a BINLOG statement if the recorded version number is 5.0.x or + below. + +and we'll get a 5.0-applicable binlog. + +Second item +----------- +One question that one needs to sort out before disabling server_id change is +why it was put there in the first place? Should it be always removed? DESCRIPTION: According to complaints in BUG#46640: 1. mysqlbinlog will emit a 5.0-incompatible "format description binlog statement" even when reading a binary log that was produced by 5.0. This will cause an error when one tires to apply mysqlbinlog output to a 5.0.x server. 2. When one executes a BINLOG statement (containing "format description binlog statement" or other events), this will cause subsequent statements run in the same session to be binlogged with the server id of the last event executed, not with the real server id. This can cause problems like infinite recursive application in a circular replication topology. This WL is to fix these issues. HIGH-LEVEL SPECIFICATION: First item ---------- AFAIU what needs to be done is: 1. record a source server version (it is in the first binlog event). 2. don't emit a BINLOG statement if the recorded version number is 5.0.x or below. and we'll get a 5.0-applicable binlog. Second item ----------- One question that one needs to sort out before disabling server_id change is why it was put there in the first place? Should it be always removed? First, the problem is not the format description log event. In fact all log events have their own value of server_id, and every event in the BINLOG statement is executed with its own id as server_id. It seems it was introduced deliberately in the patch for Bug#32407. This patch makes sql thread and BINLOG statement share code path for executing the binary event, even though the bug is really about something different (outputting format description event to allow proper execution of other BINLOG statements). Nevertheless, I think using the server_id in the events of a BINLOG statement is wrong: 1. It is different behaviour from executing normal statement-based events not written using BINLOG. 2. It causes the possibility for infinite cycle of replication in a cyclic replication setup, if the server ids in BINLOG are different from all other server ids in the cycle. 3. The functionality of applying events with original server id from the event is still available by pointing the slave thread to the binlog. The functionality of applying row-based binlog events with server id of the server executing the BINLOG statements is not otherwise available. In fact most of the code from the slave thread that is now also run when executing BINLOG statements is redundant, or does the wrong thing (like updating binlog position). Maybe there is a confusion on the purpose of the BINLOG statement. Is it to be able to apply row-based events manually to a server (mysqlbinlog|mysql)? Or is it to be able to simulate the effect of the slave sql thread? I think the former is the most important one. LOW-LEVEL DESIGN: I think the fix for point 2 is to replace the call to apply_event_and_update_pos() in mysql_client_binlog_statement() with a direct call to ev->apply_event(). However, I need to check two things to be sure this is correct: 1. The existing code does if (!ev->when) ev->when= my_time(0) Need to understand if this is needed/correct or not. 2. The existing code does ev->update_pos(). I think this is redundant for BINLOG statement (as it uses a fake rli structure), but I need to check to make sure. Once this is done, point 1 may no longer be needed. The user can use --base64-output=never when applying the output to a 5.0 server, and omit that option when applying the output to a 5.1 server. There should be no need to omit the format description event in the output when other BINLOG statements for row events are present, as it no longer changes the server id. In fact the format description event is required to be able to execute other BINLOG statements, as its purpose is to define the binary format of the events contained in these statements. Alternatively, we could implement that if the format description event of the source binlog has server version < 5.1, and --base64-output=auto (default), then the format description event is omitted (and should any BINLOG statement be needed, unlikely as it is from a 5.0 server, we will need to throw an error). The binlog format version for 5.0 and 5.1 is actually the same, hence the need to look at server version to guess if the format description event can be omitted. ESTIMATED WORK TIME ESTIMATED COMPLETION DATE ----------------------------------------------------------------------- WorkLog (v3.5.9)

1 0

[Maria-developers] Updated (by Guest): mysqlbinlog: remove undesired effects of format description binlog statement (50)
by worklog-noreply＠askmonty.org 09 Sep '09

09 Sep '09

----------------------------------------------------------------------- WORKLOG TASK -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- TASK...........: mysqlbinlog: remove undesired effects of format description binlog statement CREATION DATE..: Mon, 17 Aug 2009, 14:12 SUPERVISOR.....: Monty IMPLEMENTOR....: COPIES TO......: CATEGORY.......: Server-RawIdeaBin TASK ID........: 50 (http://askmonty.org/worklog/?tid=50) VERSION........: Server-9.x STATUS.........: Un-Assigned PRIORITY.......: 60 WORKED HOURS...: 2 ESTIMATE.......: 0 (hours remain) ORIG. ESTIMATE.: 0 PROGRESS NOTES: -=-=(Guest - Wed, 09 Sep 2009, 11:50)=-=- High Level Description modified. --- /tmp/wklog.50.old.11584 2009-09-09 11:50:12.000000000 +0300 +++ /tmp/wklog.50.new.11584 2009-09-09 11:50:12.000000000 +0300 @@ -5,7 +5,11 @@ This will cause an error when one tires to apply mysqlbinlog output to a 5.0.x server. -2. When one applies "format description binlog statement" at the slave, it -will change the slave's server_id when applied. +2. When one executes a BINLOG statement (containing "format description binlog +statement" or other events), this will cause subsequent statements run in +the same session to be binlogged with the server id of the last event +executed, not with the real server id. This can cause problems like infinite +recursive application in a circular replication topology. This WL is to fix these issues. + -=-=(Knielsen - Tue, 08 Sep 2009, 15:07)=-=- Low Level Design modified. --- /tmp/wklog.50.old.19208 2009-09-08 15:07:09.000000000 +0300 +++ /tmp/wklog.50.new.19208 2009-09-08 15:07:09.000000000 +0300 @@ -1 +1,34 @@ +I think the fix for point 2 is to replace the call to +apply_event_and_update_pos() in mysql_client_binlog_statement() with a direct +call to ev->apply_event(). However, I need to check two things to be sure this +is correct: + +1. The existing code does + + if (!ev->when) ev->when= my_time(0) + +Need to understand if this is needed/correct or not. + +2. The existing code does ev->update_pos(). I think this is redundant for +BINLOG statement (as it uses a fake rli structure), but I need to check to +make sure. + +Once this is done, point 1 may no longer be needed. The user can use +--base64-output=never when applying the output to a 5.0 server, and omit that +option when applying the output to a 5.1 server. There should be no need to +omit the format description event in the output when other BINLOG statements +for row events are present, as it no longer changes the server id. In fact the +format description event is required to be able to execute other BINLOG +statements, as its purpose is to define the binary format of the events +contained in these statements. + +Alternatively, we could implement that if the format description event of the +source binlog has server version < 5.1, and --base64-output=auto (default), +then the format description event is omitted (and should any BINLOG statement +be needed, unlikely as it is from a 5.0 server, we will need to throw an +error). + +The binlog format version for 5.0 and 5.1 is actually the same, hence the need +to look at server version to guess if the format description event can be +omitted. -=-=(Knielsen - Tue, 08 Sep 2009, 14:29)=-=- High-Level Specification modified. --- /tmp/wklog.50.old.17531 2009-09-08 14:29:30.000000000 +0300 +++ /tmp/wklog.50.new.17531 2009-09-08 14:29:30.000000000 +0300 @@ -12,3 +12,37 @@ One question that one needs to sort out before disabling server_id change is why it was put there in the first place? Should it be always removed? +First, the problem is not the format description log event. In fact all log +events have their own value of server_id, and every event in the BINLOG +statement is executed with its own id as server_id. + +It seems it was introduced deliberately in the patch for Bug#32407. This patch +makes sql thread and BINLOG statement share code path for executing the binary +event, even though the bug is really about something different (outputting +format description event to allow proper execution of other BINLOG +statements). + +Nevertheless, I think using the server_id in the events of a BINLOG statement +is wrong: + +1. It is different behaviour from executing normal statement-based events not +written using BINLOG. + +2. It causes the possibility for infinite cycle of replication in a cyclic +replication setup, if the server ids in BINLOG are different from all other +server ids in the cycle. + +3. The functionality of applying events with original server id from the event +is still available by pointing the slave thread to the binlog. The +functionality of applying row-based binlog events with server id of the server +executing the BINLOG statements is not otherwise available. + +In fact most of the code from the slave thread that is now also run when +executing BINLOG statements is redundant, or does the wrong thing (like +updating binlog position). + +Maybe there is a confusion on the purpose of the BINLOG statement. Is it to be +able to apply row-based events manually to a server (mysqlbinlog|mysql)? Or is +it to be able to simulate the effect of the slave sql thread? I think the +former is the most important one. + -=-=(Knielsen - Mon, 07 Sep 2009, 10:06)=-=- Research worklog, bugs, and existing server code. Worked 2 hours and estimate 0 hours remain (original estimate increased by 2 hours). -=-=(Psergey - Mon, 17 Aug 2009, 14:13)=-=- Dependency created: 39 now depends on 50 -=-=(Psergey - Mon, 17 Aug 2009, 14:13)=-=- High-Level Specification modified. --- /tmp/wklog.50.old.11389 2009-08-17 14:13:05.000000000 +0300 +++ /tmp/wklog.50.new.11389 2009-08-17 14:13:05.000000000 +0300 @@ -1 +1,14 @@ +First item +---------- +AFAIU what needs to be done is: +1. record a source server version (it is in the first binlog event). +2. don't emit a BINLOG statement if the recorded version number is 5.0.x or + below. + +and we'll get a 5.0-applicable binlog. + +Second item +----------- +One question that one needs to sort out before disabling server_id change is +why it was put there in the first place? Should it be always removed? DESCRIPTION: According to complaints in BUG#46640: 1. mysqlbinlog will emit a 5.0-incompatible "format description binlog statement" even when reading a binary log that was produced by 5.0. This will cause an error when one tires to apply mysqlbinlog output to a 5.0.x server. 2. When one executes a BINLOG statement (containing "format description binlog statement" or other events), this will cause subsequent statements run in the same session to be binlogged with the server id of the last event executed, not with the real server id. This can cause problems like infinite recursive application in a circular replication topology. This WL is to fix these issues. HIGH-LEVEL SPECIFICATION: First item ---------- AFAIU what needs to be done is: 1. record a source server version (it is in the first binlog event). 2. don't emit a BINLOG statement if the recorded version number is 5.0.x or below. and we'll get a 5.0-applicable binlog. Second item ----------- One question that one needs to sort out before disabling server_id change is why it was put there in the first place? Should it be always removed? First, the problem is not the format description log event. In fact all log events have their own value of server_id, and every event in the BINLOG statement is executed with its own id as server_id. It seems it was introduced deliberately in the patch for Bug#32407. This patch makes sql thread and BINLOG statement share code path for executing the binary event, even though the bug is really about something different (outputting format description event to allow proper execution of other BINLOG statements). Nevertheless, I think using the server_id in the events of a BINLOG statement is wrong: 1. It is different behaviour from executing normal statement-based events not written using BINLOG. 2. It causes the possibility for infinite cycle of replication in a cyclic replication setup, if the server ids in BINLOG are different from all other server ids in the cycle. 3. The functionality of applying events with original server id from the event is still available by pointing the slave thread to the binlog. The functionality of applying row-based binlog events with server id of the server executing the BINLOG statements is not otherwise available. In fact most of the code from the slave thread that is now also run when executing BINLOG statements is redundant, or does the wrong thing (like updating binlog position). Maybe there is a confusion on the purpose of the BINLOG statement. Is it to be able to apply row-based events manually to a server (mysqlbinlog|mysql)? Or is it to be able to simulate the effect of the slave sql thread? I think the former is the most important one. LOW-LEVEL DESIGN: I think the fix for point 2 is to replace the call to apply_event_and_update_pos() in mysql_client_binlog_statement() with a direct call to ev->apply_event(). However, I need to check two things to be sure this is correct: 1. The existing code does if (!ev->when) ev->when= my_time(0) Need to understand if this is needed/correct or not. 2. The existing code does ev->update_pos(). I think this is redundant for BINLOG statement (as it uses a fake rli structure), but I need to check to make sure. Once this is done, point 1 may no longer be needed. The user can use --base64-output=never when applying the output to a 5.0 server, and omit that option when applying the output to a 5.1 server. There should be no need to omit the format description event in the output when other BINLOG statements for row events are present, as it no longer changes the server id. In fact the format description event is required to be able to execute other BINLOG statements, as its purpose is to define the binary format of the events contained in these statements. Alternatively, we could implement that if the format description event of the source binlog has server version < 5.1, and --base64-output=auto (default), then the format description event is omitted (and should any BINLOG statement be needed, unlikely as it is from a 5.0 server, we will need to throw an error). The binlog format version for 5.0 and 5.1 is actually the same, hence the need to look at server version to guess if the format description event can be omitted. ESTIMATED WORK TIME ESTIMATED COMPLETION DATE ----------------------------------------------------------------------- WorkLog (v3.5.9)

1 0

Re: [Maria-developers] Rev 2721: Merge xtradb-7 -> MariaDB in file:///home/psergey/bzr-new/mysql-5.1-maria-contd2/
by Sergey Petrunya 08 Sep '09

08 Sep '09

Hi Peter, On Tue, Sep 08, 2009 at 06:33:00PM +0200, Peter Laursen wrote: > @Sergey > . > does this mean that we can now check out code, compile and XtraDB will be > functional? > (cc: to Khusboo as she is assigned as build-engineer for the Windows MariaDB > planned to be provided by Webyog) No, not yet: I haven't pushed this yet, because the result of the merge has certain issues. There will be more commits and then push. Besides the merge, we have also two known windows problems, as mentioned here: http://askmonty.org/wiki/index.php/MariaDB_5.1_TODO#Bugfixes We're working on these also. Do you want me to notify you when lp:maria is ready for windows packaging? BR Sergey -- Sergey Petrunia, Software Developer Monty Program AB, http://askmonty.org Blog: http://s.petrunia.net/blog

2 2

[Maria-developers] Rev 2722: Post-merge fix in file:///home/psergey/bzr-new/mysql-5.1-maria-contd2/
by Sergey Petrunya 08 Sep '09

08 Sep '09

At file:///home/psergey/bzr-new/mysql-5.1-maria-contd2/ ------------------------------------------------------------ revno: 2722 revision-id: psergey(a)askmonty.org-20090908162138-wtd6r54dkcp4hg5a parent: psergey(a)askmonty.org-20090908160458-fglojzacu8inqb9q committer: Sergey Petrunya <psergey(a)askmonty.org> branch nick: mysql-5.1-maria-contd2 timestamp: Tue 2009-09-08 20:21:38 +0400 message: Post-merge fix === removed directory 'mysql-test.percona'

1 0

[Maria-developers] Updated (by Knielsen): mysqlbinlog: remove undesired effects of format description binlog statement (50)
by worklog-noreply＠askmonty.org 08 Sep '09

08 Sep '09

----------------------------------------------------------------------- WORKLOG TASK -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- TASK...........: mysqlbinlog: remove undesired effects of format description binlog statement CREATION DATE..: Mon, 17 Aug 2009, 14:12 SUPERVISOR.....: Monty IMPLEMENTOR....: COPIES TO......: CATEGORY.......: Server-RawIdeaBin TASK ID........: 50 (http://askmonty.org/worklog/?tid=50) VERSION........: Server-9.x STATUS.........: Un-Assigned PRIORITY.......: 60 WORKED HOURS...: 2 ESTIMATE.......: 0 (hours remain) ORIG. ESTIMATE.: 0 PROGRESS NOTES: -=-=(Knielsen - Tue, 08 Sep 2009, 15:07)=-=- Low Level Design modified. --- /tmp/wklog.50.old.19208 2009-09-08 15:07:09.000000000 +0300 +++ /tmp/wklog.50.new.19208 2009-09-08 15:07:09.000000000 +0300 @@ -1 +1,34 @@ +I think the fix for point 2 is to replace the call to +apply_event_and_update_pos() in mysql_client_binlog_statement() with a direct +call to ev->apply_event(). However, I need to check two things to be sure this +is correct: + +1. The existing code does + + if (!ev->when) ev->when= my_time(0) + +Need to understand if this is needed/correct or not. + +2. The existing code does ev->update_pos(). I think this is redundant for +BINLOG statement (as it uses a fake rli structure), but I need to check to +make sure. + +Once this is done, point 1 may no longer be needed. The user can use +--base64-output=never when applying the output to a 5.0 server, and omit that +option when applying the output to a 5.1 server. There should be no need to +omit the format description event in the output when other BINLOG statements +for row events are present, as it no longer changes the server id. In fact the +format description event is required to be able to execute other BINLOG +statements, as its purpose is to define the binary format of the events +contained in these statements. + +Alternatively, we could implement that if the format description event of the +source binlog has server version < 5.1, and --base64-output=auto (default), +then the format description event is omitted (and should any BINLOG statement +be needed, unlikely as it is from a 5.0 server, we will need to throw an +error). + +The binlog format version for 5.0 and 5.1 is actually the same, hence the need +to look at server version to guess if the format description event can be +omitted. -=-=(Knielsen - Tue, 08 Sep 2009, 14:29)=-=- High-Level Specification modified. --- /tmp/wklog.50.old.17531 2009-09-08 14:29:30.000000000 +0300 +++ /tmp/wklog.50.new.17531 2009-09-08 14:29:30.000000000 +0300 @@ -12,3 +12,37 @@ One question that one needs to sort out before disabling server_id change is why it was put there in the first place? Should it be always removed? +First, the problem is not the format description log event. In fact all log +events have their own value of server_id, and every event in the BINLOG +statement is executed with its own id as server_id. + +It seems it was introduced deliberately in the patch for Bug#32407. This patch +makes sql thread and BINLOG statement share code path for executing the binary +event, even though the bug is really about something different (outputting +format description event to allow proper execution of other BINLOG +statements). + +Nevertheless, I think using the server_id in the events of a BINLOG statement +is wrong: + +1. It is different behaviour from executing normal statement-based events not +written using BINLOG. + +2. It causes the possibility for infinite cycle of replication in a cyclic +replication setup, if the server ids in BINLOG are different from all other +server ids in the cycle. + +3. The functionality of applying events with original server id from the event +is still available by pointing the slave thread to the binlog. The +functionality of applying row-based binlog events with server id of the server +executing the BINLOG statements is not otherwise available. + +In fact most of the code from the slave thread that is now also run when +executing BINLOG statements is redundant, or does the wrong thing (like +updating binlog position). + +Maybe there is a confusion on the purpose of the BINLOG statement. Is it to be +able to apply row-based events manually to a server (mysqlbinlog|mysql)? Or is +it to be able to simulate the effect of the slave sql thread? I think the +former is the most important one. + -=-=(Knielsen - Mon, 07 Sep 2009, 10:06)=-=- Research worklog, bugs, and existing server code. Worked 2 hours and estimate 0 hours remain (original estimate increased by 2 hours). -=-=(Psergey - Mon, 17 Aug 2009, 14:13)=-=- Dependency created: 39 now depends on 50 -=-=(Psergey - Mon, 17 Aug 2009, 14:13)=-=- High-Level Specification modified. --- /tmp/wklog.50.old.11389 2009-08-17 14:13:05.000000000 +0300 +++ /tmp/wklog.50.new.11389 2009-08-17 14:13:05.000000000 +0300 @@ -1 +1,14 @@ +First item +---------- +AFAIU what needs to be done is: +1. record a source server version (it is in the first binlog event). +2. don't emit a BINLOG statement if the recorded version number is 5.0.x or + below. + +and we'll get a 5.0-applicable binlog. + +Second item +----------- +One question that one needs to sort out before disabling server_id change is +why it was put there in the first place? Should it be always removed? DESCRIPTION: According to complaints in BUG#46640: 1. mysqlbinlog will emit a 5.0-incompatible "format description binlog statement" even when reading a binary log that was produced by 5.0. This will cause an error when one tires to apply mysqlbinlog output to a 5.0.x server. 2. When one applies "format description binlog statement" at the slave, it will change the slave's server_id when applied. This WL is to fix these issues. HIGH-LEVEL SPECIFICATION: First item ---------- AFAIU what needs to be done is: 1. record a source server version (it is in the first binlog event). 2. don't emit a BINLOG statement if the recorded version number is 5.0.x or below. and we'll get a 5.0-applicable binlog. Second item ----------- One question that one needs to sort out before disabling server_id change is why it was put there in the first place? Should it be always removed? First, the problem is not the format description log event. In fact all log events have their own value of server_id, and every event in the BINLOG statement is executed with its own id as server_id. It seems it was introduced deliberately in the patch for Bug#32407. This patch makes sql thread and BINLOG statement share code path for executing the binary event, even though the bug is really about something different (outputting format description event to allow proper execution of other BINLOG statements). Nevertheless, I think using the server_id in the events of a BINLOG statement is wrong: 1. It is different behaviour from executing normal statement-based events not written using BINLOG. 2. It causes the possibility for infinite cycle of replication in a cyclic replication setup, if the server ids in BINLOG are different from all other server ids in the cycle. 3. The functionality of applying events with original server id from the event is still available by pointing the slave thread to the binlog. The functionality of applying row-based binlog events with server id of the server executing the BINLOG statements is not otherwise available. In fact most of the code from the slave thread that is now also run when executing BINLOG statements is redundant, or does the wrong thing (like updating binlog position). Maybe there is a confusion on the purpose of the BINLOG statement. Is it to be able to apply row-based events manually to a server (mysqlbinlog|mysql)? Or is it to be able to simulate the effect of the slave sql thread? I think the former is the most important one. LOW-LEVEL DESIGN: I think the fix for point 2 is to replace the call to apply_event_and_update_pos() in mysql_client_binlog_statement() with a direct call to ev->apply_event(). However, I need to check two things to be sure this is correct: 1. The existing code does if (!ev->when) ev->when= my_time(0) Need to understand if this is needed/correct or not. 2. The existing code does ev->update_pos(). I think this is redundant for BINLOG statement (as it uses a fake rli structure), but I need to check to make sure. Once this is done, point 1 may no longer be needed. The user can use --base64-output=never when applying the output to a 5.0 server, and omit that option when applying the output to a 5.1 server. There should be no need to omit the format description event in the output when other BINLOG statements for row events are present, as it no longer changes the server id. In fact the format description event is required to be able to execute other BINLOG statements, as its purpose is to define the binary format of the events contained in these statements. Alternatively, we could implement that if the format description event of the source binlog has server version < 5.1, and --base64-output=auto (default), then the format description event is omitted (and should any BINLOG statement be needed, unlikely as it is from a 5.0 server, we will need to throw an error). The binlog format version for 5.0 and 5.1 is actually the same, hence the need to look at server version to guess if the format description event can be omitted. ESTIMATED WORK TIME ESTIMATED COMPLETION DATE ----------------------------------------------------------------------- WorkLog (v3.5.9)

1 0

[Maria-developers] Updated (by Knielsen): mysqlbinlog: remove undesired effects of format description binlog statement (50)
by worklog-noreply＠askmonty.org 08 Sep '09

08 Sep '09

----------------------------------------------------------------------- WORKLOG TASK -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- TASK...........: mysqlbinlog: remove undesired effects of format description binlog statement CREATION DATE..: Mon, 17 Aug 2009, 14:12 SUPERVISOR.....: Monty IMPLEMENTOR....: COPIES TO......: CATEGORY.......: Server-RawIdeaBin TASK ID........: 50 (http://askmonty.org/worklog/?tid=50) VERSION........: Server-9.x STATUS.........: Un-Assigned PRIORITY.......: 60 WORKED HOURS...: 2 ESTIMATE.......: 0 (hours remain) ORIG. ESTIMATE.: 0 PROGRESS NOTES: -=-=(Knielsen - Tue, 08 Sep 2009, 15:07)=-=- Low Level Design modified. --- /tmp/wklog.50.old.19208 2009-09-08 15:07:09.000000000 +0300 +++ /tmp/wklog.50.new.19208 2009-09-08 15:07:09.000000000 +0300 @@ -1 +1,34 @@ +I think the fix for point 2 is to replace the call to +apply_event_and_update_pos() in mysql_client_binlog_statement() with a direct +call to ev->apply_event(). However, I need to check two things to be sure this +is correct: + +1. The existing code does + + if (!ev->when) ev->when= my_time(0) + +Need to understand if this is needed/correct or not. + +2. The existing code does ev->update_pos(). I think this is redundant for +BINLOG statement (as it uses a fake rli structure), but I need to check to +make sure. + +Once this is done, point 1 may no longer be needed. The user can use +--base64-output=never when applying the output to a 5.0 server, and omit that +option when applying the output to a 5.1 server. There should be no need to +omit the format description event in the output when other BINLOG statements +for row events are present, as it no longer changes the server id. In fact the +format description event is required to be able to execute other BINLOG +statements, as its purpose is to define the binary format of the events +contained in these statements. + +Alternatively, we could implement that if the format description event of the +source binlog has server version < 5.1, and --base64-output=auto (default), +then the format description event is omitted (and should any BINLOG statement +be needed, unlikely as it is from a 5.0 server, we will need to throw an +error). + +The binlog format version for 5.0 and 5.1 is actually the same, hence the need +to look at server version to guess if the format description event can be +omitted. -=-=(Knielsen - Tue, 08 Sep 2009, 14:29)=-=- High-Level Specification modified. --- /tmp/wklog.50.old.17531 2009-09-08 14:29:30.000000000 +0300 +++ /tmp/wklog.50.new.17531 2009-09-08 14:29:30.000000000 +0300 @@ -12,3 +12,37 @@ One question that one needs to sort out before disabling server_id change is why it was put there in the first place? Should it be always removed? +First, the problem is not the format description log event. In fact all log +events have their own value of server_id, and every event in the BINLOG +statement is executed with its own id as server_id. + +It seems it was introduced deliberately in the patch for Bug#32407. This patch +makes sql thread and BINLOG statement share code path for executing the binary +event, even though the bug is really about something different (outputting +format description event to allow proper execution of other BINLOG +statements). + +Nevertheless, I think using the server_id in the events of a BINLOG statement +is wrong: + +1. It is different behaviour from executing normal statement-based events not +written using BINLOG. + +2. It causes the possibility for infinite cycle of replication in a cyclic +replication setup, if the server ids in BINLOG are different from all other +server ids in the cycle. + +3. The functionality of applying events with original server id from the event +is still available by pointing the slave thread to the binlog. The +functionality of applying row-based binlog events with server id of the server +executing the BINLOG statements is not otherwise available. + +In fact most of the code from the slave thread that is now also run when +executing BINLOG statements is redundant, or does the wrong thing (like +updating binlog position). + +Maybe there is a confusion on the purpose of the BINLOG statement. Is it to be +able to apply row-based events manually to a server (mysqlbinlog|mysql)? Or is +it to be able to simulate the effect of the slave sql thread? I think the +former is the most important one. + -=-=(Knielsen - Mon, 07 Sep 2009, 10:06)=-=- Research worklog, bugs, and existing server code. Worked 2 hours and estimate 0 hours remain (original estimate increased by 2 hours). -=-=(Psergey - Mon, 17 Aug 2009, 14:13)=-=- Dependency created: 39 now depends on 50 -=-=(Psergey - Mon, 17 Aug 2009, 14:13)=-=- High-Level Specification modified. --- /tmp/wklog.50.old.11389 2009-08-17 14:13:05.000000000 +0300 +++ /tmp/wklog.50.new.11389 2009-08-17 14:13:05.000000000 +0300 @@ -1 +1,14 @@ +First item +---------- +AFAIU what needs to be done is: +1. record a source server version (it is in the first binlog event). +2. don't emit a BINLOG statement if the recorded version number is 5.0.x or + below. + +and we'll get a 5.0-applicable binlog. + +Second item +----------- +One question that one needs to sort out before disabling server_id change is +why it was put there in the first place? Should it be always removed? DESCRIPTION: According to complaints in BUG#46640: 1. mysqlbinlog will emit a 5.0-incompatible "format description binlog statement" even when reading a binary log that was produced by 5.0. This will cause an error when one tires to apply mysqlbinlog output to a 5.0.x server. 2. When one applies "format description binlog statement" at the slave, it will change the slave's server_id when applied. This WL is to fix these issues. HIGH-LEVEL SPECIFICATION: First item ---------- AFAIU what needs to be done is: 1. record a source server version (it is in the first binlog event). 2. don't emit a BINLOG statement if the recorded version number is 5.0.x or below. and we'll get a 5.0-applicable binlog. Second item ----------- One question that one needs to sort out before disabling server_id change is why it was put there in the first place? Should it be always removed? First, the problem is not the format description log event. In fact all log events have their own value of server_id, and every event in the BINLOG statement is executed with its own id as server_id. It seems it was introduced deliberately in the patch for Bug#32407. This patch makes sql thread and BINLOG statement share code path for executing the binary event, even though the bug is really about something different (outputting format description event to allow proper execution of other BINLOG statements). Nevertheless, I think using the server_id in the events of a BINLOG statement is wrong: 1. It is different behaviour from executing normal statement-based events not written using BINLOG. 2. It causes the possibility for infinite cycle of replication in a cyclic replication setup, if the server ids in BINLOG are different from all other server ids in the cycle. 3. The functionality of applying events with original server id from the event is still available by pointing the slave thread to the binlog. The functionality of applying row-based binlog events with server id of the server executing the BINLOG statements is not otherwise available. In fact most of the code from the slave thread that is now also run when executing BINLOG statements is redundant, or does the wrong thing (like updating binlog position). Maybe there is a confusion on the purpose of the BINLOG statement. Is it to be able to apply row-based events manually to a server (mysqlbinlog|mysql)? Or is it to be able to simulate the effect of the slave sql thread? I think the former is the most important one. LOW-LEVEL DESIGN: I think the fix for point 2 is to replace the call to apply_event_and_update_pos() in mysql_client_binlog_statement() with a direct call to ev->apply_event(). However, I need to check two things to be sure this is correct: 1. The existing code does if (!ev->when) ev->when= my_time(0) Need to understand if this is needed/correct or not. 2. The existing code does ev->update_pos(). I think this is redundant for BINLOG statement (as it uses a fake rli structure), but I need to check to make sure. Once this is done, point 1 may no longer be needed. The user can use --base64-output=never when applying the output to a 5.0 server, and omit that option when applying the output to a 5.1 server. There should be no need to omit the format description event in the output when other BINLOG statements for row events are present, as it no longer changes the server id. In fact the format description event is required to be able to execute other BINLOG statements, as its purpose is to define the binary format of the events contained in these statements. Alternatively, we could implement that if the format description event of the source binlog has server version < 5.1, and --base64-output=auto (default), then the format description event is omitted (and should any BINLOG statement be needed, unlikely as it is from a 5.0 server, we will need to throw an error). The binlog format version for 5.0 and 5.1 is actually the same, hence the need to look at server version to guess if the format description event can be omitted. ESTIMATED WORK TIME ESTIMATED COMPLETION DATE ----------------------------------------------------------------------- WorkLog (v3.5.9)

1 0

[Maria-developers] Updated (by Knielsen): mysqlbinlog: remove undesired effects of format description binlog statement (50)
by worklog-noreply＠askmonty.org 08 Sep '09

08 Sep '09

----------------------------------------------------------------------- WORKLOG TASK -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- TASK...........: mysqlbinlog: remove undesired effects of format description binlog statement CREATION DATE..: Mon, 17 Aug 2009, 14:12 SUPERVISOR.....: Monty IMPLEMENTOR....: COPIES TO......: CATEGORY.......: Server-RawIdeaBin TASK ID........: 50 (http://askmonty.org/worklog/?tid=50) VERSION........: Server-9.x STATUS.........: Un-Assigned PRIORITY.......: 60 WORKED HOURS...: 2 ESTIMATE.......: 0 (hours remain) ORIG. ESTIMATE.: 0 PROGRESS NOTES: -=-=(Knielsen - Tue, 08 Sep 2009, 14:29)=-=- High-Level Specification modified. --- /tmp/wklog.50.old.17531 2009-09-08 14:29:30.000000000 +0300 +++ /tmp/wklog.50.new.17531 2009-09-08 14:29:30.000000000 +0300 @@ -12,3 +12,37 @@ One question that one needs to sort out before disabling server_id change is why it was put there in the first place? Should it be always removed? +First, the problem is not the format description log event. In fact all log +events have their own value of server_id, and every event in the BINLOG +statement is executed with its own id as server_id. + +It seems it was introduced deliberately in the patch for Bug#32407. This patch +makes sql thread and BINLOG statement share code path for executing the binary +event, even though the bug is really about something different (outputting +format description event to allow proper execution of other BINLOG +statements). + +Nevertheless, I think using the server_id in the events of a BINLOG statement +is wrong: + +1. It is different behaviour from executing normal statement-based events not +written using BINLOG. + +2. It causes the possibility for infinite cycle of replication in a cyclic +replication setup, if the server ids in BINLOG are different from all other +server ids in the cycle. + +3. The functionality of applying events with original server id from the event +is still available by pointing the slave thread to the binlog. The +functionality of applying row-based binlog events with server id of the server +executing the BINLOG statements is not otherwise available. + +In fact most of the code from the slave thread that is now also run when +executing BINLOG statements is redundant, or does the wrong thing (like +updating binlog position). + +Maybe there is a confusion on the purpose of the BINLOG statement. Is it to be +able to apply row-based events manually to a server (mysqlbinlog|mysql)? Or is +it to be able to simulate the effect of the slave sql thread? I think the +former is the most important one. + -=-=(Knielsen - Mon, 07 Sep 2009, 10:06)=-=- Research worklog, bugs, and existing server code. Worked 2 hours and estimate 0 hours remain (original estimate increased by 2 hours). -=-=(Psergey - Mon, 17 Aug 2009, 14:13)=-=- Dependency created: 39 now depends on 50 -=-=(Psergey - Mon, 17 Aug 2009, 14:13)=-=- High-Level Specification modified. --- /tmp/wklog.50.old.11389 2009-08-17 14:13:05.000000000 +0300 +++ /tmp/wklog.50.new.11389 2009-08-17 14:13:05.000000000 +0300 @@ -1 +1,14 @@ +First item +---------- +AFAIU what needs to be done is: +1. record a source server version (it is in the first binlog event). +2. don't emit a BINLOG statement if the recorded version number is 5.0.x or + below. + +and we'll get a 5.0-applicable binlog. + +Second item +----------- +One question that one needs to sort out before disabling server_id change is +why it was put there in the first place? Should it be always removed? DESCRIPTION: According to complaints in BUG#46640: 1. mysqlbinlog will emit a 5.0-incompatible "format description binlog statement" even when reading a binary log that was produced by 5.0. This will cause an error when one tires to apply mysqlbinlog output to a 5.0.x server. 2. When one applies "format description binlog statement" at the slave, it will change the slave's server_id when applied. This WL is to fix these issues. HIGH-LEVEL SPECIFICATION: First item ---------- AFAIU what needs to be done is: 1. record a source server version (it is in the first binlog event). 2. don't emit a BINLOG statement if the recorded version number is 5.0.x or below. and we'll get a 5.0-applicable binlog. Second item ----------- One question that one needs to sort out before disabling server_id change is why it was put there in the first place? Should it be always removed? First, the problem is not the format description log event. In fact all log events have their own value of server_id, and every event in the BINLOG statement is executed with its own id as server_id. It seems it was introduced deliberately in the patch for Bug#32407. This patch makes sql thread and BINLOG statement share code path for executing the binary event, even though the bug is really about something different (outputting format description event to allow proper execution of other BINLOG statements). Nevertheless, I think using the server_id in the events of a BINLOG statement is wrong: 1. It is different behaviour from executing normal statement-based events not written using BINLOG. 2. It causes the possibility for infinite cycle of replication in a cyclic replication setup, if the server ids in BINLOG are different from all other server ids in the cycle. 3. The functionality of applying events with original server id from the event is still available by pointing the slave thread to the binlog. The functionality of applying row-based binlog events with server id of the server executing the BINLOG statements is not otherwise available. In fact most of the code from the slave thread that is now also run when executing BINLOG statements is redundant, or does the wrong thing (like updating binlog position). Maybe there is a confusion on the purpose of the BINLOG statement. Is it to be able to apply row-based events manually to a server (mysqlbinlog|mysql)? Or is it to be able to simulate the effect of the slave sql thread? I think the former is the most important one. ESTIMATED WORK TIME ESTIMATED COMPLETION DATE ----------------------------------------------------------------------- WorkLog (v3.5.9)

1 0

[Maria-developers] Updated (by Knielsen): mysqlbinlog: remove undesired effects of format description binlog statement (50)
by worklog-noreply＠askmonty.org 08 Sep '09

08 Sep '09

----------------------------------------------------------------------- WORKLOG TASK -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- TASK...........: mysqlbinlog: remove undesired effects of format description binlog statement CREATION DATE..: Mon, 17 Aug 2009, 14:12 SUPERVISOR.....: Monty IMPLEMENTOR....: COPIES TO......: CATEGORY.......: Server-RawIdeaBin TASK ID........: 50 (http://askmonty.org/worklog/?tid=50) VERSION........: Server-9.x STATUS.........: Un-Assigned PRIORITY.......: 60 WORKED HOURS...: 2 ESTIMATE.......: 0 (hours remain) ORIG. ESTIMATE.: 0 PROGRESS NOTES: -=-=(Knielsen - Tue, 08 Sep 2009, 14:29)=-=- High-Level Specification modified. --- /tmp/wklog.50.old.17531 2009-09-08 14:29:30.000000000 +0300 +++ /tmp/wklog.50.new.17531 2009-09-08 14:29:30.000000000 +0300 @@ -12,3 +12,37 @@ One question that one needs to sort out before disabling server_id change is why it was put there in the first place? Should it be always removed? +First, the problem is not the format description log event. In fact all log +events have their own value of server_id, and every event in the BINLOG +statement is executed with its own id as server_id. + +It seems it was introduced deliberately in the patch for Bug#32407. This patch +makes sql thread and BINLOG statement share code path for executing the binary +event, even though the bug is really about something different (outputting +format description event to allow proper execution of other BINLOG +statements). + +Nevertheless, I think using the server_id in the events of a BINLOG statement +is wrong: + +1. It is different behaviour from executing normal statement-based events not +written using BINLOG. + +2. It causes the possibility for infinite cycle of replication in a cyclic +replication setup, if the server ids in BINLOG are different from all other +server ids in the cycle. + +3. The functionality of applying events with original server id from the event +is still available by pointing the slave thread to the binlog. The +functionality of applying row-based binlog events with server id of the server +executing the BINLOG statements is not otherwise available. + +In fact most of the code from the slave thread that is now also run when +executing BINLOG statements is redundant, or does the wrong thing (like +updating binlog position). + +Maybe there is a confusion on the purpose of the BINLOG statement. Is it to be +able to apply row-based events manually to a server (mysqlbinlog|mysql)? Or is +it to be able to simulate the effect of the slave sql thread? I think the +former is the most important one. + -=-=(Knielsen - Mon, 07 Sep 2009, 10:06)=-=- Research worklog, bugs, and existing server code. Worked 2 hours and estimate 0 hours remain (original estimate increased by 2 hours). -=-=(Psergey - Mon, 17 Aug 2009, 14:13)=-=- Dependency created: 39 now depends on 50 -=-=(Psergey - Mon, 17 Aug 2009, 14:13)=-=- High-Level Specification modified. --- /tmp/wklog.50.old.11389 2009-08-17 14:13:05.000000000 +0300 +++ /tmp/wklog.50.new.11389 2009-08-17 14:13:05.000000000 +0300 @@ -1 +1,14 @@ +First item +---------- +AFAIU what needs to be done is: +1. record a source server version (it is in the first binlog event). +2. don't emit a BINLOG statement if the recorded version number is 5.0.x or + below. + +and we'll get a 5.0-applicable binlog. + +Second item +----------- +One question that one needs to sort out before disabling server_id change is +why it was put there in the first place? Should it be always removed? DESCRIPTION: According to complaints in BUG#46640: 1. mysqlbinlog will emit a 5.0-incompatible "format description binlog statement" even when reading a binary log that was produced by 5.0. This will cause an error when one tires to apply mysqlbinlog output to a 5.0.x server. 2. When one applies "format description binlog statement" at the slave, it will change the slave's server_id when applied. This WL is to fix these issues. HIGH-LEVEL SPECIFICATION: First item ---------- AFAIU what needs to be done is: 1. record a source server version (it is in the first binlog event). 2. don't emit a BINLOG statement if the recorded version number is 5.0.x or below. and we'll get a 5.0-applicable binlog. Second item ----------- One question that one needs to sort out before disabling server_id change is why it was put there in the first place? Should it be always removed? First, the problem is not the format description log event. In fact all log events have their own value of server_id, and every event in the BINLOG statement is executed with its own id as server_id. It seems it was introduced deliberately in the patch for Bug#32407. This patch makes sql thread and BINLOG statement share code path for executing the binary event, even though the bug is really about something different (outputting format description event to allow proper execution of other BINLOG statements). Nevertheless, I think using the server_id in the events of a BINLOG statement is wrong: 1. It is different behaviour from executing normal statement-based events not written using BINLOG. 2. It causes the possibility for infinite cycle of replication in a cyclic replication setup, if the server ids in BINLOG are different from all other server ids in the cycle. 3. The functionality of applying events with original server id from the event is still available by pointing the slave thread to the binlog. The functionality of applying row-based binlog events with server id of the server executing the BINLOG statements is not otherwise available. In fact most of the code from the slave thread that is now also run when executing BINLOG statements is redundant, or does the wrong thing (like updating binlog position). Maybe there is a confusion on the purpose of the BINLOG statement. Is it to be able to apply row-based events manually to a server (mysqlbinlog|mysql)? Or is it to be able to simulate the effect of the slave sql thread? I think the former is the most important one. ESTIMATED WORK TIME ESTIMATED COMPLETION DATE ----------------------------------------------------------------------- WorkLog (v3.5.9)

1 0

Re: [Maria-developers] MariaDB release help
by Kristian Nielsen 08 Sep '09

08 Sep '09

Hakan Kuecuekyilmaz <hakan(a)askmonty.org> writes: > Where do you need some help for our release? Looking at http://askmonty.org/wiki/index.php/MariaDB_5.1_TODO, seems it is currently mostly the mysql-5.1.38 merge. One thing that would be useful is to try some additional testing on top of what is already in Buildbot, as there is quite a lot missing... some of this might be better done after the merge is pushed though, but you could start preparing. Maybe you could check what additional test suites/tools are already available that could be used? Stuff like unit tests, mysql-test-run --ps-protocol, additional test suites... One thing in particular is to prepare to test the actual release. Releases are built in a two-step process, first build a source tarball, and then build the binaries from those sources. Buildbot builds directly from a bzr checkout. And Binaries are run from the installed location, while Buildbot runs the tests from the build directory. It would be really embarrassing if some major part of our binaries were released completely broken due to some file or something not being included and no testing at all done :-(. So we could use some tools to test stuff like this. Eventually we want this in Buildbot of course. Concrete suggestion: 1. Testing on installed mysqld. Run `BUILD/compile-dist && make dist`. Unpack those sources and run `./configure <options> && make && make install`. Run `make clean` (to catch bad references into build dir). Then run the testsuite on the installed mysqld. 2. What we used to call "smoke test" at MySQL: Build a binary package, then install it on some machine and see that it can start and stop. Basically testing the package stuff (directory structure, mysqld_safe, ...) which is not covered by the test suite. See http://askmonty.org/wiki/index.php/MariaDB::PackageBuild for how packages are built currently. > My idea to work on and look into it: Currently we have around 10 test > failures on Mac/PPC. If you want, I can start looking into those > failures. Where should we report failing tests? I would like to open > a bug for each test failure. Currently we seem to be using Launchpad for bugs. However, it does not sound all that useful to blindly report one bug per failure. Probably some of them are related with same root cause. If you can identify those root causes, then one bug each makes sense (maybe this is what you meant). Otherwise a single bug probably, just noting that someone needs to look into it when time is available. - Kristian.

2 1