Comment 11 for bug 1042946
Revision history for this message
| Aleksandr Khomyakov (toairo) wrote Re: Memory leak on a filtered slave | #11 |
© 2004
Canonical Ltd.
•
Terms of use
•
Data privacy
•
Contact Launchpad Support
•
Blog
•
Careers
•
System status
•
1b1ed1a
(Get the code!)
now i not sure, because need more statistics.
yesterday i downgrade all leaked instance to 5.5.24, because this version showed me that does not leak too much how 05.05.27.
today, I see that the server is still leaked, but slower.
1 case: server with full replication, without rules.
- with 5.5.27 leak average 350Mb per 24h, after downgrade to 5.5.24- leak about 100Mb per 24h.
case #2: with part replication table from master, but most of the load is not of replication
- with 5.5.27 leak 1Gb per 24h, after downgrade to 5.5.24 - 150Mb per 24h
3 case: replication rules same as case #2, but in case #2 more
- with 5.5.27 about 500Gb per 24h
- with 5.5.20: I tried downgrade 5.5.20 before 5.5.24 - and not leaked, but on 5.5.20 worked about 24hours, need try again this version.
continue to monitor the memory.
about valgrind result on test instance.
most of time i not result about leaks, but today night after 4hours work mysql, i get this result (see. attach). but nothing new. in valgrind result from Alex, same trace.