diff options
| author | Andrei <andrei.elkin@mariadb.com> | 2022-05-12 19:24:26 +0300 |
|---|---|---|
| committer | Sergei Golubchik <serg@mariadb.org> | 2022-05-13 09:45:32 +0200 |
| commit | 726bd8c968e1b60a72f2a798a8fc269b78239cdc (patch) | |
| tree | 58f4f7c6becc57e690cb68bb1f75843de5d0ea92 | |
| parent | a5dc12eefd4bea1c3f77d02c55d0d459b4ae0566 (diff) | |
| download | mariadb-git-726bd8c968e1b60a72f2a798a8fc269b78239cdc.tar.gz | |
MDEV-28550 improper handling of replication event group that contains
GTID_LIST_EVENT or INCIDENT_EVENT.
It's legal to have either of the two inside a group. E.g
Gtid_event, Gtid_log_list_event, Query_1, ... Xid_log_event
is permitted.
However, the slave IO thread treated both
as the terminal even when the group represents a DDL query.
That causes a premature Gtid state update so the slave IO would think
the whole group has been collected while in fact Query_1 etc are yet to process.
Fixed with correcting a condition to compute the terminal event
of the group.
Tested with rpl_mysqlbinlog_slave_consistency (of 10.9) and
rpl_gtid_errorlog.test.
| -rw-r--r-- | sql/slave.cc | 5 |
1 files changed, 3 insertions, 2 deletions
diff --git a/sql/slave.cc b/sql/slave.cc index ca3397e5e7b..ac9addee7a4 100644 --- a/sql/slave.cc +++ b/sql/slave.cc @@ -6586,8 +6586,9 @@ dbug_gtid_accept: mi->using_gtid != Master_info::USE_GTID_NO && mi->events_queued_since_last_gtid > 0 && ( (mi->last_queued_gtid_standalone && - !Log_event::is_part_of_group((Log_event_type)(uchar) - buf[EVENT_TYPE_OFFSET])) || + (LOG_EVENT_IS_QUERY((Log_event_type)(uchar) + buf[EVENT_TYPE_OFFSET]) || + (uchar)buf[EVENT_TYPE_OFFSET] == INCIDENT_EVENT)) || (!mi->last_queued_gtid_standalone && ((uchar)buf[EVENT_TYPE_OFFSET] == XID_EVENT || ((uchar)buf[EVENT_TYPE_OFFSET] == QUERY_EVENT && /* QUERY_COMPRESSED_EVENT would never be commmit or rollback */ |