All,
I do have a problem with database growing rapidly. I think it is due to auto vacuum not being able to purge table history_uint.
Background: This is zabbix 6.0.12 and Postgres 13. Housekeeping is set to clear history after 30 days.
Autovacuum works on at least some other tables:
Manually analysing vacuum produces the following
I guess that this is not expected result.
What could be done to correct the situation and stop growth of the database?
I do have a problem with database growing rapidly. I think it is due to auto vacuum not being able to purge table history_uint.
Background: This is zabbix 6.0.12 and Postgres 13. Housekeeping is set to clear history after 30 days.
Autovacuum works on at least some other tables:
Code:
relname | n_live_tup | n_dead_tup | dead_percent | last_vacuum | last_autovacuum | autovacuum_count -------------------+------------+------------+---------------------+-------------------------------+-------------------------------+------------------ ... history | 128642188 | 74456 | 0.05787836879764514 | 2023-01-23 11:33:10.218843+02 | 2023-01-21 00:52:28.789278+02 | 1 history_str | 773932 | 27738 | 3.584035806763385 | | 2023-01-22 07:22:45.220621+02 | 1 history_text | 5037827 | 10740 | 0.21318715390584075 | | 2023-01-23 08:14:05.414644+02 | 2 history_uint | 723448060 | 164961825 | 22.80216564545076 | | | 0 ...
Code:
# vacuum (verbose, analyze) history_uint; INFO: aggressively vacuuming "public.history_uint" ERROR: cannot freeze committed xmax 203701611 CONTEXT: while scanning block 420246 of relation "public.history_uint"
What could be done to correct the situation and stop growth of the database?