March 2022 - developers - lists.mariadb.org

Re: [Maria-developers] 880e92a48ba: MDEV-27760 event may non stop replicate in circular semisync setup
by Sergei Golubchik 06 Mar '22

06 Mar '22

Hi, Andrei, On Mar 01, Andrei Elkin wrote: > > On Feb 28, Andrei wrote: > >> revision-id: 880e92a48ba (mariadb-10.6.6-16-g880e92a48ba) > >> parent(s): 4030a9fb2eb > >> author: Andrei > >> committer: Andrei > >> timestamp: 2022-02-21 14:58:13 +0200 > >> message: > >> > >> MDEV-27760 event may non stop replicate in circular semisync setup > >> > >> diff --git a/sql/slave.cc b/sql/slave.cc > >> index c0eef02ca7a..a1f4f4f0081 100644 > >> --- a/sql/slave.cc > >> +++ b/sql/slave.cc > >> @@ -6186,13 +6186,13 @@ static int queue_event(Master_info* mi, const uchar *buf, ulong event_len) > >> bool is_rows_event= false; > >> /* > >> The flag has replicate_same_server_id semantics and is raised to accept > >> - a same-server-id event on the semisync slave, for both the gtid and legacy > >> - connection modes. > >> - Such events can appear as result of this server recovery so the event > >> - was created there and replicated elsewhere right before the crash. At recovery > >> - it could be evicted from the server's binlog. > >> - */ > >> - bool do_accept_own_server_id= false; > >> + a same-server-id event group by the gtid strict mode semisync slave. > >> + Own server-id events can appear as result of this server crash-recovery: > >> + the transaction was created on this server then being master, got replicated > >> + elsewhere right before the crash before commit; > >> + finally at recovery the transaction gets evicted from the server's binlog. > >> + */ > >> + static bool do_accept_own_server_id= false; > > > > No, sorry, it cannot be static. If you need to preserve it between > > events, you can keep it, for example, in Master_info. > > The var is set by Gtid events to serve as a marker of the whole current Gtid-headed > group of event. > Only next Gtid event may change the current value. > > Why can't it? Why it cannot be static? Because it's not a property of the server, it's a property of a specific replication stream. Technically one can have a multi-source and do_accept_own_server_id could have different values for different masters. > > So, the question now is, why do you need to preserve the value of > > do_accept_own_server_id between queue_event invocations? > > The flag as said above applies to the whole group of events, so it's > preserved at least through so many > queue_event() calls as many event in the group (I count Gtid itself in). As far as I can see, you can calculate it for every event just the same. > >> @@ -6281,6 +6281,8 @@ static int queue_event(Master_info* mi, const uchar *buf, ulong event_len) > >> dbug_rows_event_count = 0; > >> };); > >> #endif > >> + s_id= uint4korr(buf + SERVER_ID_OFFSET); > >> + > >> mysql_mutex_lock(&mi->data_lock); > >> > >> switch (buf[EVENT_TYPE_OFFSET]) { > >> @@ -6722,6 +6724,11 @@ static int queue_event(Master_info* mi, const uchar *buf, ulong event_len) > >> > >> ++mi->events_queued_since_last_gtid; > >> inc_pos= event_len; > >> + > >> + do_accept_own_server_id= > >> + (s_id == global_system_variables.server_id && rpl_semi_sync_slave_enabled > >> + && mi->using_gtid != Master_info::USE_GTID_NO && opt_gtid_strict_mode) ? > >> + true : false; Regards, Sergei VP of MariaDB Server Engineering and security(a)mariadb.org

1 0

[Maria-developers] MDEV-27776 EXPLAIN is about to get new columns, EXPLAIN EXTENDED|PARTITIONS will be deprecated
by Oleg Smirnov 03 Mar '22

03 Mar '22

* EXPLAIN SELECT prints "partitions", "filtered" columns and produces the warning with rewritten query * EXPLAIN EXTENDED and EXPLAIN PARTITIONS produce the same output as EXPLAIN SELECT and two warnings, the second warning is about EXPLAIN EXTENDED| PARTITIONS being deprecated (same as in MySQL 5.7) -- Best regards, ------------------------- *Oleg Smirnov* Sr. Software Engineer MariaDB Server Team

1 0

Re: [Maria-developers] 5d8838a79d5: MDEV-24920: Merge "old" SQL variable to "old_mode" sql variable
by Sergei Golubchik 02 Mar '22

02 Mar '22

Hi, Rucha, On Mar 02, Rucha Deodhar wrote: > revision-id: 5d8838a79d5 (mariadb-10.6.1-352-g5d8838a79d5) > parent(s): f8b3c661231 > author: Rucha Deodhar > committer: Rucha Deodhar > timestamp: 2022-02-15 22:29:54 +0530 > message: > > MDEV-24920: Merge "old" SQL variable to "old_mode" sql variable > > Analysis: There are 2 server variables- "old_mode" and "old". "old" is no > longer needed as "old_mode" has replaced it (however still used in some places > in the code). "old_mode" and "old" has same purpose- emulate behavior from > previous MariaDB versions. So they can be merged to avoid confusion. > Fix: Deprecate "old" variable and create another mode for @@old_mode to mimic > behavior of previous "old" variable. Create specific modes for specifix task > that --old sql variable was doing earlier and use the new modes instead. > > diff --git a/mysql-test/main/mysqld--help.result b/mysql-test/main/mysqld--help.result > index 7cbfa52b846..261dbe2ffe0 100644 > --- a/mysql-test/main/mysqld--help.result > +++ b/mysql-test/main/mysqld--help.result > @@ -683,7 +683,8 @@ The following specify which files/extra groups are read (specified before remain > --old-mode=name Used to emulate old behavior from earlier MariaDB or > MySQL versions. Any combination of: > NO_DUP_KEY_WARNINGS_WITH_IGNORE, NO_PROGRESS_INFO, > - ZERO_DATE_TIME_CAST, UTF8_IS_UTF8MB3 > + ZERO_DATE_TIME_CAST, UTF8_IS_UTF8MB3, > + INDEX_HINT_MASK_JOIN, CHECKSUM_SLOW_NULLS I'm sorry, I still cannot understand what these new modes do :( I know that it was my idea to use "CHECKSUM_SLOW_NULLS", but now I'm, like, "what does it mean?". Let's discuss it on slack INDEX_HINT_MASK_JOIN - this seems to be easier, what about INGORE_INDEX_ONLY_FOR_JOIN ? That's a bit long, but shorter than NO_DUP_KEY_WARNINGS_WITH_IGNORE. Regards, Sergei VP of MariaDB Server Engineering and security(a)mariadb.org

1 0

Re: [Maria-developers] 880e92a48ba: MDEV-27760 event may non stop replicate in circular semisync setup
by Sergei Golubchik 01 Mar '22

01 Mar '22

Hi, Andrei, On Feb 28, Andrei wrote: > revision-id: 880e92a48ba (mariadb-10.6.6-16-g880e92a48ba) > parent(s): 4030a9fb2eb > author: Andrei > committer: Andrei > timestamp: 2022-02-21 14:58:13 +0200 > message: > > MDEV-27760 event may non stop replicate in circular semisync setup > > diff --git a/sql/slave.cc b/sql/slave.cc > index c0eef02ca7a..a1f4f4f0081 100644 > --- a/sql/slave.cc > +++ b/sql/slave.cc > @@ -6186,13 +6186,13 @@ static int queue_event(Master_info* mi, const uchar *buf, ulong event_len) > bool is_rows_event= false; > /* > The flag has replicate_same_server_id semantics and is raised to accept > - a same-server-id event on the semisync slave, for both the gtid and legacy > - connection modes. > - Such events can appear as result of this server recovery so the event > - was created there and replicated elsewhere right before the crash. At recovery > - it could be evicted from the server's binlog. > - */ > - bool do_accept_own_server_id= false; > + a same-server-id event group by the gtid strict mode semisync slave. > + Own server-id events can appear as result of this server crash-recovery: > + the transaction was created on this server then being master, got replicated > + elsewhere right before the crash before commit; > + finally at recovery the transaction gets evicted from the server's binlog. > + */ > + static bool do_accept_own_server_id= false; No, sorry, it cannot be static. If you need to preserve it between events, you can keep it, for example, in Master_info. yes, I remember, that's where it was and I asked why it's in Master_info. You could've answered, that you need to preserve the value between queue_event invocations - I didn't realize that at the time. So, the question now is, why do you need to preserve the value of do_accept_own_server_id between queue_event invocations? > /* > FD_q must have been prepared for the first R_a event > inside get_master_version_and_clock() > @@ -6281,6 +6281,8 @@ static int queue_event(Master_info* mi, const uchar *buf, ulong event_len) > dbug_rows_event_count = 0; > };); > #endif > + s_id= uint4korr(buf + SERVER_ID_OFFSET); > + > mysql_mutex_lock(&mi->data_lock); > > switch (buf[EVENT_TYPE_OFFSET]) { > @@ -6722,6 +6724,11 @@ static int queue_event(Master_info* mi, const uchar *buf, ulong event_len) > > ++mi->events_queued_since_last_gtid; > inc_pos= event_len; > + > + do_accept_own_server_id= > + (s_id == global_system_variables.server_id && rpl_semi_sync_slave_enabled > + && mi->using_gtid != Master_info::USE_GTID_NO && opt_gtid_strict_mode) ? > + true : false; this is a ridiculous pattern bool var = (bool expr) ? true : false; don't do that. > } > break; > /* > @@ -6909,7 +6916,6 @@ static int queue_event(Master_info* mi, const uchar *buf, ulong event_len) > */ > > mysql_mutex_lock(log_lock); > - s_id= uint4korr(buf + SERVER_ID_OFFSET); > /* > Write the event to the relay log, unless we reconnected in the middle > of an event group and now need to skip the initial part of the group that > @@ -6955,7 +6961,7 @@ static int queue_event(Master_info* mi, const uchar *buf, ulong event_len) > else > if ((s_id == global_system_variables.server_id && > !(mi->rli.replicate_same_server_id || > - (do_accept_own_server_id= rpl_semi_sync_slave_enabled))) || > + do_accept_own_server_id)) || > event_that_should_be_ignored(buf) || > /* > the following conjunction deals with IGNORE_SERVER_IDS, if set > Regards, Sergei VP of MariaDB Server Engineering and security(a)mariadb.org

1 0