Memory leak on a filtered slave (Valgrind: malloc in Log_event::read_log_event)
I have recently been upgrading various MySQL instances to 5.5.25a. I have one specific machine that experiencing a memory leak. I fear this report may lack any actionable meat but here's what I can tell you:
* The machine that is leaking is replicating an all innodb stream. It has been doing so for years on previous versions with out issue.
* While there are several databases in the binary logs feeding it, it is only capturing about a dozen specific tables from a single database using Replicate_Do_Table
* No other Replicate_* options are filled in
* This machines master also has other slaves that are not using any of the aforementioned filtering and none of them are experiencing this leak.
* No applications talk to this instance. It's role is just to replicate with log-slave-updates in order to generate a smaller bin log to push over the wire to slaves in remote datacenters
* I've read over http://
* It has an innodb_
* I have attempted to kill the replication threads going in and out to see if this might indicate something session related to one of those. This had no effect in freeing any memory. The event scheduler thread is running, however there are no actual events to be run here. Similarly stopping that thread does not free any memory.
* My only current recourse is the periodically restart the mysql process :(
Please let me know if there's anything else that would be of use to actually track down the root cause of this bug.
- Memory leak on a filtered slave
+ Memory leak on a filtered slave (Valgrind: malloc in