Bug #96443 | M-S, sql playback on slave hang | ||
---|---|---|---|
Submitted: | 7 Aug 2019 6:42 | Modified: | 17 Oct 2019 10:16 |
Reporter: | lieyia -- | Email Updates: | |
Status: | Can't repeat | Impact on me: | |
Category: | MySQL Server: Replication | Severity: | S2 (Serious) |
Version: | 5.7.21 | OS: | CentOS (CentOS release 6.6 (Final)) |
Assigned to: | CPU Architecture: | x86 |
[7 Aug 2019 6:42]
lieyia --
[7 Aug 2019 6:43]
lieyia --
my.cnf,mysql.err,stack
Attachment: info (application/octet-stream, text), 56.96 KiB.
[7 Aug 2019 6:46]
lieyia --
and the table has 200+ million records
[9 Aug 2019 2:10]
lieyia --
Some more information: 1. the data of slave is backed up from master by xtrabackup, 1.2TB 2. when i start slave after buf_dump_thread loading all ibuf(the result of "select * from performance_schema.events_stages_current" is empty), buf_dump_thread do not wait lock where sql hang; 3. when sql hang, i kill mysqld and start angain, slave playback normally after several times;
[9 Aug 2019 2:31]
lieyia --
info of no lock on buf_dump_thread
Attachment: info2 (application/octet-stream, text), 43.17 KiB.
[17 Oct 2019 10:16]
MySQL Verification Team
Hello lieyia, Thank you for the report and details. I tried to reproduce this issue on a moderate box with the provided schema and huge dummy data but on latest GA builds 5.7.27/5.7.28. MySQL 5.7.21 is nearly 2 year old release and since then many bugs fixed in the replication module, moreover we don't fix bugs in old versions, don't backport bug fixes, so need to check with latest version anyway. So, please, upgrade and inform us if problem still exists. If you can provide more information, feel free to add it to this bug and change the status back to 'Open'. Thank you for your interest in MySQL. regards, Umesh