[MDEV-28435] rpl.rpl_mysqlbinlog_slave_consistency fails intermittently on tables comparison Created: 2022-04-28  Updated: 2022-05-03  Resolved: 2022-04-28

Status: Closed
Project: MariaDB Server
Component/s: Replication, Tests
Affects Version/s: 10.9
Fix Version/s: 10.9.1

Type: Bug Priority: Major
Reporter: Angelique Sklavounos (Inactive) Assignee: Brandon Nesterenko
Resolution: Fixed Votes: 0
Labels: None

Attachments: File mysqld.1.err.4     File mysqld.2.err.4    

 Description   

rpl.rpl_mysqlbinlog_slave_consistency fails intermittently (retry-pass or retry-fail [but not necessarily for the same test case]):

10.9 43fa8e0b8f3bae1ff8493cfd3adb3944
https://buildbot.askmonty.org/buildbot/builders/kvm-rpm-centos73-ppc64le/builds/14215/steps/mtr/logs/stdio

rpl.rpl_mysqlbinlog_slave_consistency 'stmt' w4 [ fail ]
 
        Test ended at 2022-04-26 22:41:54
 
 
 
CURRENT_TEST: rpl.rpl_mysqlbinlog_slave_consistency
 
mysqltest: In included file "/usr/share/mysql-test/suite/rpl/include/rpl_check_table_consistency.inc": 
included from /usr/share/mysql-test/suite/rpl/include/mysqlbinlog_slave_consistency.inc at line 147:
 
included from /usr/share/mysql-test/suite/rpl/t/rpl_mysqlbinlog_slave_consistency.test at line 249:
 
At line 36: "master had 3 tables but slave had 1 after binlog replay"

Searching through cross-ref, this same error is seen randomly with the various test cases (see below).
FWIW, I looked through most of the failures and could not see any cases fail where a --stop_position was specified.

Below are collated fail messages from different builders:

included from /buildbot/s390x-sles-15/build/mysql-test/suite/rpl/t/rpl_mysqlbinlog_slave_consistency.test at line 73:
 
At line 36: "master had 5 tables but slave had 0 after binlog replay"
 
 
 
included from /home/buildbot/aarch64-debian-9/build/mysql-test/suite/rpl/t/rpl_mysqlbinlog_slave_consistency.test at line 85:
 
At line 36: "master had 2 tables but slave had 0 after binlog replay"
 
 
 
included from /home/buildbot/s390x-ubuntu-2004/build/mysql-test/suite/rpl/t/rpl_mysqlbinlog_slave_consistency.test at line 97:
 
At line 36: "master had 4 tables but slave had 0 after binlog replay"
 
 
 
included from /buildbot/s390x-sles-15/build/mysql-test/suite/rpl/t/rpl_mysqlbinlog_slave_consistency.test at line 108:
 
At line 36: "master had 3 tables but slave had 0 after binlog replay"
 
 
 
included from /home/buildbot/s390x-ubuntu-2004/build/mysql-test/suite/rpl/t/rpl_mysqlbinlog_slave_consistency.test at line 120:
 
At line 36: "master had 1 tables but slave had 0 after binlog replay"
 
 
 
included from /home/buildbot/aarch64-centos-7/build/mysql-test/suite/rpl/t/rpl_mysqlbinlog_slave_consistency.test at line 131:
 
At line 36: "master had 2 tables but slave had 0 after binlog replay"
 
 
 
included from /home/buildbot/s390x-ubuntu-2004/build/mysql-test/suite/rpl/t/rpl_mysqlbinlog_slave_consistency.test at line 143:
 
At line 36: "master had 2 tables but slave had 0 after binlog replay"
 
 
 
included from /home/buildbot/s390x-ubuntu-2004/build/mysql-test/suite/rpl/t/rpl_mysqlbinlog_slave_consistency.test at line 165:
 
At line 36: "master had 4 tables but slave had 0 after binlog replay"
 
 
 
included from /buildbot/s390x-rhel-8/build/mysql-test/suite/rpl/t/rpl_mysqlbinlog_slave_consistency.test at line 177:
 
At line 36: "master had 1 tables but slave had 0 after binlog replay"
 
 
 
included from /usr/share/mysql/mysql-test/suite/rpl/t/rpl_mysqlbinlog_slave_consistency.test at line 189:
 
At line 36: "master had 1 tables but slave had 0 after binlog replay"
 
 
 
included from /buildbot/s390x-rhel-8/build/mysql-test/suite/rpl/t/rpl_mysqlbinlog_slave_consistency.test at line 237:
 
At line 36: "master had 1 tables but slave had 0 after binlog replay"
 
 
 
included from /usr/share/mysql-test/suite/rpl/t/rpl_mysqlbinlog_slave_consistency.test at line 249:
 
At line 36: "master had 3 tables but slave had 1 after binlog replay"
 
 
 
included from /buildbot/amd64-ubuntu-2004-fulltest/build/mysql-test/suite/rpl/t/rpl_mysqlbinlog_slave_consistency.test at line 261:
 
At line 36: "master had 2 tables but slave had 0 after binlog replay"
 
 
 
included from /buildbot/s390x-rhel-8/build/mysql-test/suite/rpl/t/rpl_mysqlbinlog_slave_consistency.test at line 344:
 
At line 36: "master had 1 tables but slave had 0 after binlog replay"

First seen on 63a479e2e033cfe578040e1331efc1192413ee7c.

 Comments   
Comment by Brandon Nesterenko [ 2022-04-28 ]

Hi Andrei! This is ready for review:

Patch 695854a

Comment by Brandon Nesterenko [ 2022-04-28 ]

Pushed into 10.9 as 51b28b2

Thanks angelique.sklavounos for the preliminary analysis that helped identify the cause of the failure quicker!

Generated at Thu Feb 08 10:00:43 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.