Replication servers crashes when starting slave

[Steffen Boehme]

We run different mysql databases in master slave setups where our main db has one master and 10 slaves.
The master is version 5.5.20, slaves are different ...
(see attached files for more info).

We got now problems on 2 of our slaves, which started crashing during the normal work.
When trying to restart the databases with replication enabled they crashed immediately, when started with the skip-slave-start option the server started up well, but after restarting the slave SQL_THREAD the server crashed again.
This happened only to 2 of our slaves, not to the rest.

I tried a CHANGE MASTER TO to get fresh copies of the binlogs, but this didn't help.
Right now I downgraded one of the slaves to version 5.5.17 (the hardy version where the OS is Lucid) and restarted it with replication enabled and this works ...

I will attach several crash logs, the server variables and status and hope this will help to check out, whats the problem.

Thanks!
Steffen

[Alexey Kopytov (akopytov)]

Thank you for the bug report. This is a known regression with the userstat feature in Percona Server 5.5.24, see bug #1008278.

The workaround is to disable userstat on 5.5.24 slaves. We will release the fix in Percona Server 5.5.25.