October 2009 - developers - lists.mariadb.org

[Maria-developers] bzr commit into MariaDB 5.1, with Maria 1.5:maria branch (monty:2742)
by Michael Widenius 06 Oct '09

06 Oct '09

#At lp:maria based on revid:knielsen@knielsen-hq.org-20091005122657-02wli3dgv82t3125 2742 Michael Widenius 2009-10-06 [merge] Merge with bug fix modified: storage/maria/ma_extra.c storage/maria/ma_locking.c storage/myisam/mi_extra.c === modified file 'storage/maria/ma_extra.c' --- a/storage/maria/ma_extra.c 2009-02-19 09:01:25 +0000 +++ b/storage/maria/ma_extra.c 2009-10-06 06:13:56 +0000 @@ -323,9 +323,16 @@ int maria_extra(MARIA_HA *info, enum ha_ error= my_errno; info->lock_type= F_UNLCK; } - if (share->kfile.file >= 0) - _ma_decrement_open_count(info); + /* + We don't need to call _mi_decrement_open_count() if we are + dropping the table, as the files will be removed anyway. If we + are aborted before the files is removed, it's better to not + call it as in that case the automatic repair on open will add + the missing index entries + */ pthread_mutex_lock(&share->intern_lock); + if (share->kfile.file >= 0 && function != HA_EXTRA_PREPARE_FOR_DROP) + _ma_decrement_open_count(info); if (info->trn) { _ma_remove_table_from_trnman(share, info->trn); === modified file 'storage/maria/ma_locking.c' --- a/storage/maria/ma_locking.c 2009-02-09 21:52:42 +0000 +++ b/storage/maria/ma_locking.c 2009-10-06 06:13:56 +0000 @@ -484,7 +484,7 @@ int _ma_decrement_open_count(MARIA_HA *i { uint old_lock=info->lock_type; share->global_changed=0; - lock_error=maria_lock_database(info,F_WRLCK); + lock_error= my_disable_locking ? 0 : maria_lock_database(info, F_WRLCK); /* Its not fatal even if we couldn't get the lock ! */ if (share->state.open_count > 0) { @@ -499,7 +499,7 @@ int _ma_decrement_open_count(MARIA_HA *i MYF(MY_NABP)); } } - if (!lock_error) + if (!lock_error && !my_disable_locking) lock_error=maria_lock_database(info,old_lock); } return test(lock_error || write_error); === modified file 'storage/myisam/mi_extra.c' --- a/storage/myisam/mi_extra.c 2009-09-03 14:05:38 +0000 +++ b/storage/myisam/mi_extra.c 2009-10-06 06:13:56 +0000 @@ -286,9 +286,19 @@ int mi_extra(MI_INFO *info, enum ha_extr info->lock_type = F_UNLCK; } if (share->kfile >= 0) + { + /* + We don't need to call _mi_decrement_open_count() if we are + dropping the table, as the files will be removed anyway. If we + are aborted before the files is removed, it's better to not + call it as in that case the automatic repair on open will add + the missing index entries + */ + if (function != HA_EXTRA_PREPARE_FOR_DROP) _mi_decrement_open_count(info); - if (share->kfile >= 0 && my_close(share->kfile,MYF(0))) - error=my_errno; + if (my_close(share->kfile,MYF(0))) + error=my_errno; + } { LIST *list_element ; for (list_element=myisam_open_list ;

1 0

[Maria-developers] bzr commit into MariaDB 5.1, with Maria 1.5:maria branch (monty:2742)
by Michael Widenius 06 Oct '09

06 Oct '09

#At lp:maria based on revid:monty@askmonty.org-20091006061356-w62ct9y58lfyl1oa 2742 Michael Widenius 2009-10-06 [merge] Merge with trunk modified: mysql-test/lib/mtr_cases.pm mysql-test/mysql-test-run.pl === modified file 'mysql-test/lib/mtr_cases.pm' --- a/mysql-test/lib/mtr_cases.pm 2009-09-07 20:50:10 +0000 +++ b/mysql-test/lib/mtr_cases.pm 2009-10-05 12:26:57 +0000 @@ -560,8 +560,12 @@ sub collect_one_suite($) sub optimize_cases { my ($cases)= @_; + my @new_cases= (); + foreach my $tinfo ( @$cases ) { + push @new_cases, $tinfo; + # Skip processing if already marked as skipped next if $tinfo->{skip}; @@ -615,6 +619,13 @@ sub optimize_cases { $tinfo->{'skip'}= 1; $tinfo->{'comment'}= "Doesn't support --binlog-format='$test_binlog_format'"; + # This test was added as a replication combination, but it is not + # actually ever possible to run it, as it is not made for this + # combination. + # So delete it from the list, rather than confuse the user with a + # message that this test is skipped (it is not really, just run + # with other combinations). + pop(@new_cases); next; } } @@ -683,6 +694,7 @@ sub optimize_cases { } } } + @$cases= @new_cases; } === modified file 'mysql-test/mysql-test-run.pl' --- a/mysql-test/mysql-test-run.pl 2009-09-26 02:25:56 +0000 +++ b/mysql-test/mysql-test-run.pl 2009-10-05 12:26:57 +0000 @@ -126,13 +126,13 @@ my $path_config_file; # The ge # executables will be used by the test suite. our $opt_vs_config = $ENV{'MTR_VS_CONFIG'}; -my $DEFAULT_SUITES= "binlog,federated,main,maria,ndb,rpl,rpl_ndb,innodb"; +my $DEFAULT_SUITES= "binlog,federated,main,maria,rpl,innodb"; my $opt_suites; our $opt_usage; our $opt_list_options; our $opt_suites; -our $opt_suites_default= "main,backup,backup_engines,binlog,rpl,rpl_ndb,ndb"; # Default suites to run +our $opt_suites_default= "main,backup,backup_engines,binlog,rpl"; # Default suites to run our $opt_script_debug= 0; # Script debugging, enable with --script-debug our $opt_verbose= 0; # Verbose output, enable with --verbose our $exe_mysql;

1 0

[Maria-developers] bzr commit into MariaDB 5.1, with Maria 1.5:maria branch (monty:2741)
by Michael Widenius 06 Oct '09

06 Oct '09

#At lp:maria based on revid:monty@askmonty.org-20091002115611-mp80efi9m4h4qyi6 2741 Michael Widenius 2009-10-06 Don't call _ma_decrement_open_count() for ma_extra(HA_EXTRA_PREPARE_FOR_DROP). Don't call _mi_decrement_open_count() for mi_extra(HA_EXTRA_PREPARE_FOR_DROP). This ensures that if we empty the key cache and a drop table fails later, the index will be automaticly rebuilt modified: storage/maria/ma_extra.c storage/maria/ma_locking.c storage/myisam/mi_extra.c per-file messages: storage/maria/ma_extra.c Don't call _ma_decrement_open_count() for ma_extra(HA_EXTRA_PREPARE_FOR_DROP). This ensures that if we empty the key cache and a drop table fails later, the index will be automaticly rebuilt storage/maria/ma_locking.c Simple optimization: Don't call maria_lock_database() if locking is disabled storage/myisam/mi_extra.c Don't call _mi_decrement_open_count() for mi_extra(HA_EXTRA_PREPARE_FOR_DROP). Simplify code to remove if === modified file 'storage/maria/ma_extra.c' --- a/storage/maria/ma_extra.c 2009-02-19 09:01:25 +0000 +++ b/storage/maria/ma_extra.c 2009-10-06 06:13:56 +0000 @@ -323,9 +323,16 @@ int maria_extra(MARIA_HA *info, enum ha_ error= my_errno; info->lock_type= F_UNLCK; } - if (share->kfile.file >= 0) - _ma_decrement_open_count(info); + /* + We don't need to call _mi_decrement_open_count() if we are + dropping the table, as the files will be removed anyway. If we + are aborted before the files is removed, it's better to not + call it as in that case the automatic repair on open will add + the missing index entries + */ pthread_mutex_lock(&share->intern_lock); + if (share->kfile.file >= 0 && function != HA_EXTRA_PREPARE_FOR_DROP) + _ma_decrement_open_count(info); if (info->trn) { _ma_remove_table_from_trnman(share, info->trn); === modified file 'storage/maria/ma_locking.c' --- a/storage/maria/ma_locking.c 2009-02-09 21:52:42 +0000 +++ b/storage/maria/ma_locking.c 2009-10-06 06:13:56 +0000 @@ -484,7 +484,7 @@ int _ma_decrement_open_count(MARIA_HA *i { uint old_lock=info->lock_type; share->global_changed=0; - lock_error=maria_lock_database(info,F_WRLCK); + lock_error= my_disable_locking ? 0 : maria_lock_database(info, F_WRLCK); /* Its not fatal even if we couldn't get the lock ! */ if (share->state.open_count > 0) { @@ -499,7 +499,7 @@ int _ma_decrement_open_count(MARIA_HA *i MYF(MY_NABP)); } } - if (!lock_error) + if (!lock_error && !my_disable_locking) lock_error=maria_lock_database(info,old_lock); } return test(lock_error || write_error); === modified file 'storage/myisam/mi_extra.c' --- a/storage/myisam/mi_extra.c 2009-09-03 14:05:38 +0000 +++ b/storage/myisam/mi_extra.c 2009-10-06 06:13:56 +0000 @@ -286,9 +286,19 @@ int mi_extra(MI_INFO *info, enum ha_extr info->lock_type = F_UNLCK; } if (share->kfile >= 0) + { + /* + We don't need to call _mi_decrement_open_count() if we are + dropping the table, as the files will be removed anyway. If we + are aborted before the files is removed, it's better to not + call it as in that case the automatic repair on open will add + the missing index entries + */ + if (function != HA_EXTRA_PREPARE_FOR_DROP) _mi_decrement_open_count(info); - if (share->kfile >= 0 && my_close(share->kfile,MYF(0))) - error=my_errno; + if (my_close(share->kfile,MYF(0))) + error=my_errno; + } { LIST *list_element ; for (list_element=myisam_open_list ;

1 0

[Maria-developers] bzr commit into MariaDB 5.1, with Maria 1.5:maria branch (knielsen:2752)
by knielsen＠knielsen-hq.org 05 Oct '09

05 Oct '09

#At lp:maria 2752 knielsen(a)knielsen-hq.org 2009-10-05 MBug#443092: test cases mysql and mysql_upgrade fails with wildcard DNS When DNS is not working and returns some address even for non-existing host names, some test cases fail with wrong error. Fix by disabling these tests on such systems. added: mysql-test/include/have_working_dns.inc mysql-test/r/have_working_dns.require modified: mysql-test/mysql-test-run.pl mysql-test/t/mysql.test mysql-test/t/mysql_upgrade.test === added file 'mysql-test/include/have_working_dns.inc' --- a/mysql-test/include/have_working_dns.inc 1970-01-01 00:00:00 +0000 +++ b/mysql-test/include/have_working_dns.inc 2009-10-05 13:22:23 +0000 @@ -0,0 +1,11 @@ +# +# Check if we have a working DNS. +# Some 'wildcard dns' return some address even for non-existing hosts. This +# makes it hard to test connections to such host names. +# The actual check for working DNS is done in Perl, and the result available +# in an environment variable. +# +--require r/have_working_dns.require +disable_query_log; +eval SELECT LENGTH('$HAVE_BROKEN_DNS') = 0 AS 'have_working_dns'; +enable_query_log; === modified file 'mysql-test/mysql-test-run.pl' --- a/mysql-test/mysql-test-run.pl 2009-09-29 19:02:48 +0000 +++ b/mysql-test/mysql-test-run.pl 2009-10-05 13:22:23 +0000 @@ -2003,6 +2003,15 @@ sub environment_setup { $ENV{'MYSQL_TMP_DIR'}= $opt_tmpdir; $ENV{'MYSQLTEST_VARDIR'}= $opt_vardir; + # + # Some stupid^H^H^H^H^H^Hignorant network providers set up "wildcard DNS" + # servers that return some given web server address for any lookup of a + # non-existent host name. This confuses test cases that want to test the + # behaviour when connecting to a non-existing host, so we need to be able + # to disable those tests when DNS is broken. + # + $ENV{HAVE_BROKEN_DNS}= defined(gethostbyname('invalid_hostname')); + # ---------------------------------------------------- # Setup env for NDB # ---------------------------------------------------- === added file 'mysql-test/r/have_working_dns.require' --- a/mysql-test/r/have_working_dns.require 1970-01-01 00:00:00 +0000 +++ b/mysql-test/r/have_working_dns.require 2009-10-05 13:22:23 +0000 @@ -0,0 +1,2 @@ +have_working_dns +1 === modified file 'mysql-test/t/mysql.test' --- a/mysql-test/t/mysql.test 2009-07-31 00:51:25 +0000 +++ b/mysql-test/t/mysql.test 2009-10-05 13:22:23 +0000 @@ -1,3 +1,4 @@ +-- source include/have_working_dns.inc # This test should work in embedded server after we fix mysqltest -- source include/not_embedded.inc # === modified file 'mysql-test/t/mysql_upgrade.test' --- a/mysql-test/t/mysql_upgrade.test 2009-07-28 19:59:38 +0000 +++ b/mysql-test/t/mysql_upgrade.test 2009-10-05 13:22:23 +0000 @@ -1,3 +1,4 @@ +-- source include/have_working_dns.inc # Can't run test of external client with embedded server -- source include/not_embedded.inc

1 0

[Maria-developers] [Branch ~maria-captains/maria/5.1] Rev 2741: MBug#443014: Too many 'skipped' messages in mysql-test-run
by noreply＠launchpad.net 05 Oct '09

05 Oct '09

------------------------------------------------------------ revno: 2741 committer: knielsen(a)knielsen-hq.org branch nick: work-5.1 timestamp: Mon 2009-10-05 14:26:57 +0200 message: MBug#443014: Too many 'skipped' messages in mysql-test-run Remove mysql-test-run.pl unnecessary and confusing 'skipped' messages. - In mysql-test-run.pl, we auto-generate combinations of replication tests. But this sometimes generates combinations that are meaningless, like running a test that requires row-based replication with statement-based. These superfluous combinationes should not be reported as skipped, they should just be deleted. - Remove ndb suites from default suites, as we do not support NDB in MariaDB. Keep skip messages resulting from running mysql-test-run.pl in special ways, eg. --mysqld=--binlog-format=statement. modified: mysql-test/lib/mtr_cases.pm mysql-test/mysql-test-run.pl -- lp:maria https://code.launchpad.net/~maria-captains/maria/5.1 Your team Maria developers is subscribed to branch lp:maria. To unsubscribe from this branch go to https://code.launchpad.net/~maria-captains/maria/5.1/+edit-subscription.

1 0

[Maria-developers] bzr commit into MariaDB 5.1, with Maria 1.5:maria branch (knielsen:2741)
by knielsen＠knielsen-hq.org 05 Oct '09

05 Oct '09

#At lp:maria 2741 knielsen(a)knielsen-hq.org 2009-10-05 MBug#443014: Too many 'skipped' messages in mysql-test-run Remove mysql-test-run.pl unnecessary and confusing 'skipped' messages. - In mysql-test-run.pl, we auto-generate combinations of replication tests. But this sometimes generates combinations that are meaningless, like running a test that requires row-based replication with statement-based. These superfluous combinationes should not be reported as skipped, they should just be deleted. - Remove ndb suites from default suites, as we do not support NDB in MariaDB. Keep skip messages resulting from running mysql-test-run.pl in special ways, eg. --mysqld=--binlog-format=statement. modified: mysql-test/lib/mtr_cases.pm mysql-test/mysql-test-run.pl === modified file 'mysql-test/lib/mtr_cases.pm' --- a/mysql-test/lib/mtr_cases.pm 2009-09-07 20:50:10 +0000 +++ b/mysql-test/lib/mtr_cases.pm 2009-10-05 12:26:57 +0000 @@ -560,8 +560,12 @@ sub collect_one_suite($) sub optimize_cases { my ($cases)= @_; + my @new_cases= (); + foreach my $tinfo ( @$cases ) { + push @new_cases, $tinfo; + # Skip processing if already marked as skipped next if $tinfo->{skip}; @@ -615,6 +619,13 @@ sub optimize_cases { $tinfo->{'skip'}= 1; $tinfo->{'comment'}= "Doesn't support --binlog-format='$test_binlog_format'"; + # This test was added as a replication combination, but it is not + # actually ever possible to run it, as it is not made for this + # combination. + # So delete it from the list, rather than confuse the user with a + # message that this test is skipped (it is not really, just run + # with other combinations). + pop(@new_cases); next; } } @@ -683,6 +694,7 @@ sub optimize_cases { } } } + @$cases= @new_cases; } === modified file 'mysql-test/mysql-test-run.pl' --- a/mysql-test/mysql-test-run.pl 2009-09-26 02:25:56 +0000 +++ b/mysql-test/mysql-test-run.pl 2009-10-05 12:26:57 +0000 @@ -126,13 +126,13 @@ my $path_config_file; # The ge # executables will be used by the test suite. our $opt_vs_config = $ENV{'MTR_VS_CONFIG'}; -my $DEFAULT_SUITES= "binlog,federated,main,maria,ndb,rpl,rpl_ndb,innodb"; +my $DEFAULT_SUITES= "binlog,federated,main,maria,rpl,innodb"; my $opt_suites; our $opt_usage; our $opt_list_options; our $opt_suites; -our $opt_suites_default= "main,backup,backup_engines,binlog,rpl,rpl_ndb,ndb"; # Default suites to run +our $opt_suites_default= "main,backup,backup_engines,binlog,rpl"; # Default suites to run our $opt_script_debug= 0; # Script debugging, enable with --script-debug our $opt_verbose= 0; # Verbose output, enable with --verbose our $exe_mysql;

1 0

[Maria-developers] bzr commit into MariaDB 5.1, with Maria 1.5:maria branch (knielsen:2741)
by knielsen＠knielsen-hq.org 05 Oct '09

05 Oct '09

#At lp:maria 2741 knielsen(a)knielsen-hq.org 2009-10-05 MBug#443014: Too many 'skipped' messages in mysql-test-run Remove mysql-test-run.pl unnecessary and confusing 'skipped' messages. - In mysql-test-run.pl, we auto-generate combinations of replication tests. But this sometimes generates combinations that are meaningless, like running a test that requires row-based replication with statement-based. These superfluous combinationes should not be reported as skipped, they should just be deleted. - Remove ndb suites from default suites, as we do not support NDB in MariaDB. Keep skip messages resulting from running mysql-test-run.pl in special ways, eg. --mysqld=--binlog-format=statement. modified: mysql-test/lib/mtr_cases.pm mysql-test/mysql-test-run.pl === modified file 'mysql-test/lib/mtr_cases.pm' --- a/mysql-test/lib/mtr_cases.pm 2009-09-07 20:50:10 +0000 +++ b/mysql-test/lib/mtr_cases.pm 2009-10-05 11:06:49 +0000 @@ -560,10 +560,13 @@ sub collect_one_suite($) sub optimize_cases { my ($cases)= @_; - foreach my $tinfo ( @$cases ) + my $i= 0; + while ($i < @$cases) { + my $tinfo= $cases->[$i]; + # Skip processing if already marked as skipped - next if $tinfo->{skip}; + goto next_iteration if $tinfo->{skip}; # ======================================================= # If a special binlog format was selected with @@ -615,6 +618,13 @@ sub optimize_cases { $tinfo->{'skip'}= 1; $tinfo->{'comment'}= "Doesn't support --binlog-format='$test_binlog_format'"; + # This test was added as a replication combination, but it is not + # actually ever possible to run it, as it is not made for this + # combination. + # So delete it from the list, rather than confuse the user with a + # message that this test is skipped (it is not really, just run + # with other combinations). + splice(@$cases, $i, 1); next; } } @@ -682,6 +692,9 @@ sub optimize_cases { if ( $default_engine =~ /^innodb/i ); } } + + next_iteration: + $i++; } } === modified file 'mysql-test/mysql-test-run.pl' --- a/mysql-test/mysql-test-run.pl 2009-09-26 02:25:56 +0000 +++ b/mysql-test/mysql-test-run.pl 2009-10-05 11:06:49 +0000 @@ -126,13 +126,13 @@ my $path_config_file; # The ge # executables will be used by the test suite. our $opt_vs_config = $ENV{'MTR_VS_CONFIG'}; -my $DEFAULT_SUITES= "binlog,federated,main,maria,ndb,rpl,rpl_ndb,innodb"; +my $DEFAULT_SUITES= "binlog,federated,main,maria,rpl,innodb"; my $opt_suites; our $opt_usage; our $opt_list_options; our $opt_suites; -our $opt_suites_default= "main,backup,backup_engines,binlog,rpl,rpl_ndb,ndb"; # Default suites to run +our $opt_suites_default= "main,backup,backup_engines,binlog,rpl"; # Default suites to run our $opt_script_debug= 0; # Script debugging, enable with --script-debug our $opt_verbose= 0; # Verbose output, enable with --verbose our $exe_mysql;

2 1

[Maria-developers] Rev 2731: BUG#39249 Maria:query cache returns out of date results in file:///Users/bell/maria/bzr/work-maria-5.1-querycache/
by sanja＠askmonty.org 05 Oct '09

05 Oct '09

At file:///Users/bell/maria/bzr/work-maria-5.1-querycache/ ------------------------------------------------------------ revno: 2731 revision-id: sanja(a)askmonty.org-20090923065953-l23fsjfy8lx3hkm4 parent: knielsen(a)knielsen-hq.org-20090907131358-6bq1e3qfcgi30hu8 committer: sanja(a)askmonty.org branch nick: work-maria-5.1-querycache timestamp: Wed 2009-09-23 09:59:53 +0300 message: BUG#39249 Maria:query cache returns out of date results BUG#41098 Query Cache returns wrong result with concurent insert === modified file 'BUILD/SETUP.sh' --- a/BUILD/SETUP.sh 2009-05-11 15:31:30 +0000 +++ b/BUILD/SETUP.sh 2009-09-23 06:59:53 +0000 @@ -172,6 +172,7 @@ max_no_embedded_configs="$SSL_LIBRARY --with-plugins=max" +max_no_qc_configs="$SSL_LIBRARY --with-plugins=max --without-query-cache" max_no_ndb_configs="$SSL_LIBRARY --with-plugins=max-no-ndb --with-embedded-server --with-libevent" max_configs="$SSL_LIBRARY --with-plugins=max --with-embedded-server -with-libevent" # Disable NDB in maria max builds === added file 'BUILD/compile-pentium-debug-max-no-qc' --- a/BUILD/compile-pentium-debug-max-no-qc 1970-01-01 00:00:00 +0000 +++ b/BUILD/compile-pentium-debug-max-no-qc 2009-09-23 06:59:53 +0000 @@ -0,0 +1,9 @@ +#! /bin/sh + +path=`dirname $0` +. "$path/SETUP.sh" + +extra_flags="$pentium_cflags $debug_cflags" +extra_configs="$pentium_configs $debug_configs $max_no_qc_configs" + +. "$path/FINISH.sh" === modified file 'sql/mysql_priv.h' --- a/sql/mysql_priv.h 2009-09-03 14:56:46 +0000 +++ b/sql/mysql_priv.h 2009-09-23 06:59:53 +0000 @@ -933,7 +933,6 @@ #define query_cache_abort(A) #define query_cache_end_of_result(A) -#define query_cache_invalidate_by_MyISAM_filename_ref NULL #define query_cache_maybe_disabled(T) 1 #define query_cache_is_cacheable_query(L) 0 #endif /*HAVE_QUERY_CACHE*/ === modified file 'storage/maria/ha_maria.cc' --- a/storage/maria/ha_maria.cc 2009-06-29 21:03:30 +0000 +++ b/storage/maria/ha_maria.cc 2009-09-23 06:59:53 +0000 @@ -28,6 +28,7 @@ #include <my_bit.h> #include "ha_maria.h" #include "trnman_public.h" +#include "trnman.h" C_MODE_START #include "maria_def.h" @@ -907,6 +908,8 @@ if (!(file= maria_open(name, mode, test_if_locked | HA_OPEN_FROM_SQL_LAYER))) return (my_errno ? my_errno : -1); + file->s->chst_invalidator= query_cache_invalidate_by_MyISAM_filename_ref; + if (test_if_locked & (HA_OPEN_IGNORE_IF_LOCKED | HA_OPEN_TMP_TABLE)) VOID(maria_extra(file, HA_EXTRA_NO_WAIT_LOCK, 0)); @@ -3227,6 +3230,9 @@ */ *engine_data= 0; + if (file->s->now_transactional && file->s->have_versioning) + return (file->trn->trid != file->s->state.last_change_trn); + /* If a concurrent INSERT has happened just before the currently processed SELECT statement, the total size of the table is unknown. === modified file 'storage/maria/ma_delete.c' --- a/storage/maria/ma_delete.c 2009-01-12 11:12:00 +0000 +++ b/storage/maria/ma_delete.c 2009-09-23 06:59:53 +0000 @@ -118,13 +118,6 @@ mi_sizestore(lastpos, info->cur_row.lastpos); VOID(_ma_writeinfo(info,WRITEINFO_UPDATE_KEYFILE)); allow_break(); /* Allow SIGHUP & SIGINT */ - if (info->invalidator != 0) - { - DBUG_PRINT("info", ("invalidator... '%s' (delete)", - share->open_file_name.str)); - (*info->invalidator)(share->open_file_name.str); - info->invalidator=0; - } DBUG_RETURN(0); err: === modified file 'storage/maria/ma_locking.c' --- a/storage/maria/ma_locking.c 2009-02-09 21:52:42 +0000 +++ b/storage/maria/ma_locking.c 2009-09-23 06:59:53 +0000 @@ -214,7 +214,6 @@ VOID(_ma_test_if_changed(info)); info->lock_type=lock_type; - info->invalidator=share->invalidator; share->w_locks++; share->tot_locks++; break; @@ -269,7 +268,6 @@ } if (check_keybuffer) VOID(_ma_test_if_changed(info)); - info->invalidator=share->invalidator; } else if (lock_type == F_WRLCK && info->lock_type == F_RDLCK) { === modified file 'storage/maria/ma_state.c' --- a/storage/maria/ma_state.c 2008-12-27 02:05:16 +0000 +++ b/storage/maria/ma_state.c 2009-09-23 06:59:53 +0000 @@ -318,6 +318,13 @@ DBUG_ASSERT(!info->s->base.born_transactional); share->state.state= *info->state; info->state= &share->state.state; +#ifdef HAVE_QUERY_CACHE + DBUG_PRINT("info", ("invalidator... '%s' (status update)", + info->s->data_file_name.str)); + DBUG_ASSERT(info->s->chst_invalidator != NULL); + (*info->s->chst_invalidator)((const char *)info->s->data_file_name.str); +#endif + } info->append_insert_at_end= 0; } @@ -469,6 +476,8 @@ tables->state_start.checksum); history->trid= trn->commit_trid; + share->state.last_change_trn= trn->commit_trid; + if (history->next) { /* Remove not visible states */ === modified file 'storage/maria/ma_update.c' --- a/storage/maria/ma_update.c 2008-10-31 23:14:58 +0000 +++ b/storage/maria/ma_update.c 2009-09-23 06:59:53 +0000 @@ -187,13 +187,6 @@ */ VOID(_ma_writeinfo(info, WRITEINFO_UPDATE_KEYFILE)); allow_break(); /* Allow SIGHUP & SIGINT */ - if (info->invalidator != 0) - { - DBUG_PRINT("info", ("invalidator... '%s' (update)", - share->open_file_name.str)); - (*info->invalidator)(share->open_file_name.str); - info->invalidator=0; - } DBUG_RETURN(0); err: === modified file 'storage/maria/ma_write.c' --- a/storage/maria/ma_write.c 2009-02-19 09:01:25 +0000 +++ b/storage/maria/ma_write.c 2009-09-23 06:59:53 +0000 @@ -295,13 +295,6 @@ info->cur_row.lastpos= filepos; VOID(_ma_writeinfo(info, WRITEINFO_UPDATE_KEYFILE)); - if (info->invalidator != 0) - { - DBUG_PRINT("info", ("invalidator... '%s' (update)", - share->open_file_name.str)); - (*info->invalidator)(share->open_file_name.str); - info->invalidator=0; - } /* Update status of the table. We need to do so after each row write === modified file 'storage/maria/maria_def.h' --- a/storage/maria/maria_def.h 2009-02-19 09:01:25 +0000 +++ b/storage/maria/maria_def.h 2009-09-23 06:59:53 +0000 @@ -83,6 +83,7 @@ pgcache_page_no_t first_bitmap_with_space; ulonglong auto_increment; TrID create_trid; /* Minum trid for file */ + TrID last_change_trn; /* selfdescriptive */ ulong update_count; /* Updated for each write lock */ ulong status; double *rec_per_key_part; @@ -337,7 +338,7 @@ /* Mapings to read/write the data file */ size_t (*file_read)(MARIA_HA *, uchar *, size_t, my_off_t, myf); size_t (*file_write)(MARIA_HA *, const uchar *, size_t, my_off_t, myf); - invalidator_by_filename invalidator; /* query cache invalidator */ + invalidator_by_filename chst_invalidator; /* query cache invalidator */ my_off_t key_del_current; /* delete links for index pages */ ulong this_process; /* processid */ ulong last_process; /* For table-change-check */ @@ -507,7 +508,6 @@ uint int_nod_flag; /* -""- */ uint32 int_keytree_version; /* -""- */ int (*read_record)(MARIA_HA *, uchar*, MARIA_RECORD_POS); - invalidator_by_filename invalidator; /* query cache invalidator */ ulonglong last_auto_increment; /* auto value at start of statement */ ulong this_unique; /* uniq filenumber or thread */ ulong last_unique; /* last unique number */ === modified file 'storage/myisam/ha_myisam.cc' --- a/storage/myisam/ha_myisam.cc 2009-06-29 21:03:30 +0000 +++ b/storage/myisam/ha_myisam.cc 2009-09-23 06:59:53 +0000 @@ -660,6 +660,9 @@ if (!(file=mi_open(name, mode, test_if_locked | HA_OPEN_FROM_SQL_LAYER))) return (my_errno ? my_errno : -1); + + file->s->chst_invalidator= query_cache_invalidate_by_MyISAM_filename_ref; + if (!table->s->tmp_table) /* No need to perform a check for tmp table */ { if ((my_errno= table2myisam(table, &keyinfo, &recinfo, &recs))) === modified file 'storage/myisam/mi_locking.c' --- a/storage/myisam/mi_locking.c 2009-09-03 14:05:38 +0000 +++ b/storage/myisam/mi_locking.c 2009-09-23 06:59:53 +0000 @@ -329,6 +329,12 @@ #endif info->s->state.state= *info->state; info->state= &info->s->state.state; +#ifdef HAVE_QUERY_CACHE + DBUG_PRINT("info", ("invalidator... '%s' (status update)", + info->filename)); + DBUG_ASSERT(info->s->chst_invalidator != NULL); + (*info->s->chst_invalidator)((const char *)info->filename); +#endif } info->append_insert_at_end= 0; === modified file 'storage/myisam/myisamdef.h' --- a/storage/myisam/myisamdef.h 2009-04-25 09:04:38 +0000 +++ b/storage/myisam/myisamdef.h 2009-09-23 06:59:53 +0000 @@ -190,7 +190,10 @@ const uchar *record, my_off_t pos); size_t (*file_read) (MI_INFO *, uchar *, size_t, my_off_t, myf); size_t (*file_write) (MI_INFO *, const uchar *, size_t, my_off_t, myf); - invalidator_by_filename invalidator; /* query cache invalidator */ + /* query cache invalidator for merged tables */ + invalidator_by_filename invalidator; + /* query cache invalidator for changing state */ + invalidator_by_filename chst_invalidator; ulong this_process; /* processid */ ulong last_process; /* For table-change-check */ ulong last_version; /* Version on start */

2 1

[Maria-developers] Progress (by Knielsen): mysqlbinlog: remove undesired effects of format description binlog statement (50)
by worklog-noreply＠askmonty.org 05 Oct '09

05 Oct '09

----------------------------------------------------------------------- WORKLOG TASK -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- TASK...........: mysqlbinlog: remove undesired effects of format description binlog statement CREATION DATE..: Mon, 17 Aug 2009, 14:12 SUPERVISOR.....: Monty IMPLEMENTOR....: Knielsen COPIES TO......: CATEGORY.......: Server-Sprint TASK ID........: 50 (http://askmonty.org/worklog/?tid=50) VERSION........: Server-9.x STATUS.........: Code-Review PRIORITY.......: 60 WORKED HOURS...: 16 ESTIMATE.......: 4 (hours remain) ORIG. ESTIMATE.: 10 PROGRESS NOTES: -=-=(Knielsen - Mon, 05 Oct 2009, 14:38)=-=- Follow-up with MySQL developers on patch Worked 2 hours and estimate 4 hours remain (original estimate unchanged). -=-=(Knielsen - Tue, 29 Sep 2009, 10:12)=-=- Category updated. --- /tmp/wklog.50.old.3293 2009-09-29 10:12:26.000000000 +0300 +++ /tmp/wklog.50.new.3293 2009-09-29 10:12:26.000000000 +0300 @@ -1 +1 @@ -Server-RawIdeaBin +Server-Sprint -=-=(Knielsen - Tue, 29 Sep 2009, 10:12)=-=- Status updated. --- /tmp/wklog.50.old.3293 2009-09-29 10:12:26.000000000 +0300 +++ /tmp/wklog.50.new.3293 2009-09-29 10:12:26.000000000 +0300 @@ -1 +1 @@ -Un-Assigned +Code-Review -=-=(Knielsen - Mon, 28 Sep 2009, 16:00)=-=- Updated patch against MySQL 5.1.39 Fixes a couple of issues with the patch. Discussions with MySQL developer. Worked 4 hours and estimate 6 hours remain (original estimate unchanged). -=-=(Knielsen - Wed, 09 Sep 2009, 18:31)=-=- Discussed with MySQL developers on Bug#46640. Committed a suggested fix: https://lists.launchpad.net/maria-developers/msg00926.html Worked 8 hours and estimate 10 hours remain (original estimate increased by 18 hours). -=-=(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 ------------------------------------------------------------ -=-=(View All Progress Notes, 11 total)=-=- http://askmonty.org/worklog/index.pl?tid=50&nolimit=1 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] Progress (by Knielsen): mysqlbinlog: remove undesired effects of format description binlog statement (50)
by worklog-noreply＠askmonty.org 05 Oct '09

05 Oct '09

----------------------------------------------------------------------- WORKLOG TASK -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- TASK...........: mysqlbinlog: remove undesired effects of format description binlog statement CREATION DATE..: Mon, 17 Aug 2009, 14:12 SUPERVISOR.....: Monty IMPLEMENTOR....: Knielsen COPIES TO......: CATEGORY.......: Server-Sprint TASK ID........: 50 (http://askmonty.org/worklog/?tid=50) VERSION........: Server-9.x STATUS.........: Code-Review PRIORITY.......: 60 WORKED HOURS...: 16 ESTIMATE.......: 4 (hours remain) ORIG. ESTIMATE.: 10 PROGRESS NOTES: -=-=(Knielsen - Mon, 05 Oct 2009, 14:38)=-=- Follow-up with MySQL developers on patch Worked 2 hours and estimate 4 hours remain (original estimate unchanged). -=-=(Knielsen - Tue, 29 Sep 2009, 10:12)=-=- Category updated. --- /tmp/wklog.50.old.3293 2009-09-29 10:12:26.000000000 +0300 +++ /tmp/wklog.50.new.3293 2009-09-29 10:12:26.000000000 +0300 @@ -1 +1 @@ -Server-RawIdeaBin +Server-Sprint -=-=(Knielsen - Tue, 29 Sep 2009, 10:12)=-=- Status updated. --- /tmp/wklog.50.old.3293 2009-09-29 10:12:26.000000000 +0300 +++ /tmp/wklog.50.new.3293 2009-09-29 10:12:26.000000000 +0300 @@ -1 +1 @@ -Un-Assigned +Code-Review -=-=(Knielsen - Mon, 28 Sep 2009, 16:00)=-=- Updated patch against MySQL 5.1.39 Fixes a couple of issues with the patch. Discussions with MySQL developer. Worked 4 hours and estimate 6 hours remain (original estimate unchanged). -=-=(Knielsen - Wed, 09 Sep 2009, 18:31)=-=- Discussed with MySQL developers on Bug#46640. Committed a suggested fix: https://lists.launchpad.net/maria-developers/msg00926.html Worked 8 hours and estimate 10 hours remain (original estimate increased by 18 hours). -=-=(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 ------------------------------------------------------------ -=-=(View All Progress Notes, 11 total)=-=- http://askmonty.org/worklog/index.pl?tid=50&nolimit=1 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