Details
-
Bug
-
Status: Closed (View Workflow)
-
Critical
-
Resolution: Fixed
-
10.1.20
-
10.2.14, 10.1.32
Description
After upgrading to 10.1.20, replication now intermittently stops (randomly and not all slaves of a master at the same time) with the following errors:
2017-10-01 20:38:20 140079966804736 [ERROR] Master 'trades': Error reading packet from server: bogus data in log event; the first event 'mysql-bin.038184' at 156161357, the last event read from 'mysql-bin.038185' at 60018003, the last byte read from 'mysql-bin.038185' at 60018022. (server_errno=1236)
|
2017-10-01 20:38:20 140079966804736 [ERROR] Master 'trades': Slave I/O: Got fatal error 1236 from master when reading data from binary log: 'bogus data in log event; the first event 'mysql-bin.038184' at 156161357, the last event read from 'mysql-bin.038185' at 60018003, the last byte read from 'mysql-bin.038185' at 60018022.', Internal MariaDB error code: 1236
|
We can fix this by issuing a START SLAVE and everything works without issue but the regular replication failures occurs.
It looks like upstream bug : https://bugs.mysql.com/bug.php?id=84752
With setting slave_compressed_protocol=1 and change the sync_binlog value to off, it's working fine. This is not easily reproducible. As per the comment in upstream bug, able to reproduce only with high load on master server and multiple slaves (like 5 or 6)
Attachments
Issue Links
- relates to
-
-
- Closed
-
- links to