0000212: pgpool runs out of memory - Pgpool-II Bug Tracker

ID	Project	Category	View Status	Date Submitted	Last Update

0000212	Pgpool-II	Bug	public	2016-06-29 21:49	2016-09-28 10:07

Reporter	lassetuominen	Assigned To	Muhammad Usama
Priority	normal	Severity	crash	Reproducibility	always
Status	closed	Resolution	open
Platform	Linux	OS	CentOS	OS Version	7 64-bit

Summary	0000212: pgpool runs out of memory
Description	Our setup has four db servers with one pgpool in replication mode with no loadbalancing or caching. pgpool server has 8 GB memory and db servers have 4 GB memory. After a couple of hours running our frontend application, pgpool has eaten all of the memory and crashes. Occasionally one of the backend databases might crash as well and there is a mismatch between databases. I ran pgpool using valgrind: valgrind --tool=memcheck --track-origins=yes --undef-value-errors=yes --trace-children=yes --leak-check=full pgpool -n -f /etc/pgpool-II-95/pgpool.conf > pgpool_valgrind.log 2>&1 & Please let me know if you need more information.
Steps To Reproduce	Could be possible to reproduce with pgbench.
Additional Information	PostgreSQL 9.5.3 pgpool-II 3.5.2 Somewhat relevant parameters of pgpool.conf: listen_backlog_multiplier = 2 serialize_accept = off authentication_timeout = 60 num_init_children = 4 max_pool = 4 child_life_time = 0 child_max_connections = 0 connection_life_time = 0 client_idle_limit = 0 connection_cache = off reset_query_list = 'ABORT; DISCARD ALL' replication_mode = on replicate_select = on insert_lock = on lobj_lock_table = '' replication_stop_on_mismatch = off failover_if_affected_tuples_mismatch = off relcache_expire = 300 relcache_size = 256 check_temp_table = off check_unlogged_table = off memory_cache_enabled = off
Tags	No tags attached.

lassetuominen 2016-06-29 21:49 reporter	valgrind_pgpool.txt (69,465 bytes) valgrind_pgpool.txt (69,465 bytes)

lassetuominen 2016-06-29 21:49 reporter	kernelmessages.txt (12,640 bytes) kernelmessages.txt (12,640 bytes)

lassetuominen 2016-06-29 21:50 reporter	memoryusage.png (41,777 bytes) memoryusage.png (41,777 bytes)

Muhammad Usama 2016-07-13 03:15 developer ~0000900	Hi I am not able to reproduce the memory issue using the pgbench and also the valgrind report you shared only shows a couple of KB memory unreachable. Can you please share some more information on how to reproduce a memory issue. Did you get this with pgbench or some other specific type of load? Also if it is possible can you plese share the output of following just before pgpool runs out of memory 1- free -m 2- vmstat -s 3- top 4- cat /proc/meminfo

t-ishii 2016-09-28 10:07 developer ~0001097	No response over 2 month. Issue closed.

Date Modified	Username	Field	Change
2016-06-29 21:49	lassetuominen	New Issue
2016-06-29 21:49	lassetuominen	File Added: valgrind_pgpool.txt
2016-06-29 21:49	lassetuominen	File Added: kernelmessages.txt
2016-06-29 21:50	lassetuominen	File Added: memoryusage.png
2016-07-01 15:35	t-ishii	Assigned To	=> Muhammad Usama
2016-07-01 15:35	t-ishii	Status	new => assigned
2016-07-13 03:15	Muhammad Usama	Note Added: 0000900
2016-07-13 03:15	Muhammad Usama	Status	assigned => feedback
2016-09-28 10:07	t-ishii	Note Added: 0001097
2016-09-28 10:07	t-ishii	Status	feedback => closed