|
|
Which version of PostgreSQL are you using? Is the row count also increasing? |
|
|
|
- Old SOGo nightly build 4.0.4.20181204-1.
- Row count seems not increasing:
sogo=# select count(*) from sogo_cache_folder ;
count
-------
5301
sogo_cache_folder.c_uid has duplicate values.
Did a fresh backup of the sogo_cache_folder table, then drop it and restore, the size reduced to 13MB, but it's quickly growing to 1693MB in less than 30 minutes, still growing. |
|
|
|
To be precise:
- Row count does NOT increase.
- Table size is growing quickly.
|
|
|
|
Check your PostgreSQL setup. Tuneup the autovacuum configuration. |
|
|
|
Upgraded SOGo to the latest nightly build 5.1.0.20210414-1, but seems no suitable sope packages for this sogo version? It's 4.9.r1664.SHORTDATE and sogo doesn't work anymore, complaining:
2021-04-14 17:26:09.198 sogod[77230] EXCEPTION: <NSException: 0x55ca43abb4f0> NAME:NSInvalidArgumentException REASON:NSURL(instance) does not recognize queryComponents INFO:(null)
|
|
|
|
Server OS is Ubuntu 16.04.7 LTS (Xenial). |
|
|
|
You'll need to force an upgrade to the latest SOPE packages. The ones you have installed (ending with .SHORTDATE) have a bad versioning. |
|
|
|
Removing and reinstalling all sope + sogo packages fixes the does not recognize queryComponents error. |
|
|
|
Updated global autovauum settings. But is it a good idea to configure individual table level autovacuum settings for sogo.sogo_cache_folder by default? |
|
|
|
It depends on your global settings. Maybe you'll want more aggressive settings for the sogo_cache_folder table if you have disk space issues. |
|
|
|
Update after tuned autovacuum for 2 days:
postgresql.conf has settings like below, should be fast enough to trigger the autovacuum:
track_counts = on
autovacuum = on
autovacuum_max_workers = 3
autovacuum_naptime = 1min
autovacuum_vacuum_threshold = 50
autovacuum_analyze_threshold = 50
autovacuum_vacuum_scale_factor = 0.01
autovacuum_analyze_scale_factor = 0.01
- The situation: It still grows fast, 70GB in 2 days. Any idea?
sogo=# select relname,n_dead_tup,last_autovacuum,now() from pg_stat_user_tables;
relname | n_dead_tup | last_autovacuum | now
------------------------+------------+-------------------------------+-------------------------------
sogo_user_profile | 25 | 2021-04-15 07:09:14.391686+02 | 2021-04-16 03:41:04.018066+02
sogo_alarms_folder | 0 | | 2021-04-16 03:41:04.018066+02
sogo_quick_appointment | 119 | 2021-04-14 17:52:48.863971+02 | 2021-04-16 03:41:04.018066+02
sogo_acl | 0 | | 2021-04-16 03:41:04.018066+02
sogo_sessions_folder | 17 | 2021-04-15 21:45:22.963037+02 | 2021-04-16 03:41:04.018066+02
sogo_quick_contact | 13 | | 2021-04-16 03:41:04.018066+02
sogo_cache_folder | 22 | 2021-04-16 03:22:33.693857+02 | 2021-04-16 03:41:04.018066+02
sogo_store | 47 | 2021-04-14 17:52:16.922995+02 | 2021-04-16 03:41:04.018066+02
sogo_folder_info | 0 | | 2021-04-16 03:41:04.018066+02
(9 rows)
sogo=# SELECT pg_size_pretty( pg_total_relation_size('sogo_cache_folder') );
pg_size_pretty
----------------
70 GB
(1 row)
|
|
