Memory leak @ read_view_open_now (read0read.c:152/166) in Percona Server 5.5.30-30.2 GA
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Percona Server moved to https://jira.percona.com/projects/PS |
Invalid
|
Undecided
|
Unassigned | ||
5.1 |
Invalid
|
Undecided
|
Unassigned | ||
5.5 |
Fix Released
|
Critical
|
Alexey Kopytov | ||
5.6 |
Invalid
|
Undecided
|
Unassigned |
Bug Description
After upgrade to Percona 5.5.30-30.2 server started to use more RAM. After 4-5h it uses all ram and swap and then oom-killer kills mysql process.
Server is running only mysql.
Version before upgrade: Percona 5.5.30-30.1.
New version installed from rpm. OS: CentOS release 6.4., x86_64.
With 5.5.30-30.1 server was running fine, and there was no problem with ram.
There is 13Gb ram and 4Gb swap in the server.
Variables changed from defaults:
max_connections = 1024
skip-name-resolve
table_cache = 1024
thread_cache = 128
thread_cache_size = 128
innodb_
innodb_
innodb_
innodb_
innodb_table_locks = 0
innodb_flush_method = O_DIRECT
max_tmp_tables = 10M
tmp_table_size = 512M
max_heap_table_size = 512M
tmpdir = /dev/shm/
key_buffer_size = 1G
sort_buffer = 2M
join_buffer = 32M
query-cache-size = 256M
mysqladmin status:
Uptime: 15382 Threads: 190 Questions: 11064491 Slow queries: 226905 Opens: 681 Flush tables: 1 Open tables: 591 Queries per second avg: 719.314
Output of show engine innodb status is attached.
Related branches
- Laurynas Biveinis (community): Approve
-
Diff: 222 lines (+68/-13)9 files modifiedPercona-Server/mysql-test/t/percona_bug1170103.test (+23/-0)
Percona-Server/storage/innobase/handler/ha_innodb.cc (+4/-0)
Percona-Server/storage/innobase/include/srv0srv.h (+5/-0)
Percona-Server/storage/innobase/read/read0read.c (+6/-0)
Percona-Server/storage/innobase/row/row0sel.c (+2/-2)
Percona-Server/storage/innobase/srv/srv0srv.c (+14/-0)
Percona-Server/storage/innobase/trx/trx0sys.c (+2/-0)
Percona-Server/storage/innobase/trx/trx0trx.c (+1/-0)
build/percona-server.spec (+11/-11)
- Laurynas Biveinis (community): Approve
-
Diff: 424 lines (+113/-73)7 files modifiedPercona-Server/storage/innobase/include/read0read.h (+17/-4)
Percona-Server/storage/innobase/include/trx0purge.h (+4/-0)
Percona-Server/storage/innobase/include/trx0trx.h (+1/-3)
Percona-Server/storage/innobase/read/read0read.cc (+79/-51)
Percona-Server/storage/innobase/row/row0sel.cc (+1/-1)
Percona-Server/storage/innobase/trx/trx0purge.cc (+7/-2)
Percona-Server/storage/innobase/trx/trx0trx.cc (+4/-12)
Changed in percona-xtradb-cluster: | |
milestone: | none → 5.5.31-25 |
status: | New → Triaged |
no longer affects: | percona-xtradb-cluster |
Do you use jemalloc or any other non-default memory allocation library preloaded?
These settings are too high I'd say:
tmp_table_size = 512M
max_heap_table_size = 512M
Please, upload the output of:
show global status;
to this report.