[MDEV-6608] Slave loses master binlog filename when master crashes in the middle of writing an event group Created: 2014-08-19 Updated: 2022-09-12 Resolved: 2022-09-12 |
|
| Status: | Closed |
| Project: | MariaDB Server |
| Component/s: | OTHER |
| Affects Version/s: | 10.0.13 |
| Fix Version/s: | N/A |
| Type: | Bug | Priority: | Major |
| Reporter: | Kristian Nielsen | Assignee: | Andrei Elkin |
| Resolution: | Won't Fix | Votes: | 0 |
| Labels: | replication | ||
| Description |
|
If the master crashes in the middle of writing an event group to the binlog, The slave code is written to handle this by recognising the master restart on But there is a bug in the update of master binlog position. The update of the The result is that the slave ends up in a state where it has processed events I checked MySQL, it seems they avoid this problem during binlog recovery at Here is an MTR test case. The test case fails by --sync-with-master not
|
| Comments |
| Comment by Kristian Nielsen [ 2014-08-25 ] |
|
I made a patch that solves the problem like the MySQL solution, by truncating But on second thought, I really do not like this approach, I think it is Generally, it is a very bad idea to destroy data during crash recovery. The In this particular case, some (possibly temporary) problem reading some event And even if recovery would not be possible from a damaged-but-not-truncated So I think I want to drop that patch, and instead try an approach that leaves |
| Comment by Sergei Golubchik [ 2022-09-12 ] |
|
10.0 was EOLed in March 2019 |