Hello Icinga users,
Just a post to let you know of my troubles.
We have a significant Icinga installation (4400 hosts, 62000 services).
It runs since 2018, 3 years ago.
Nothing was made to handle purging, and we can see now essentially one big table : icinga_statehistory (70 GB on drive).
I have tester a select count(*) correponding to the delete that would be done by activating IDO cleanup (should be more advertized) and keeping one year of history.
Select It lasted 40min.
I scheduled an operation for 1-2h for the real delete.
I’m here writing this 3 hours after start, not knowing when it will end…
Icingaweb dashboards are not updated anymore.
At least notifications work.
I though about doing it by shorter interval, but at least 40 only for the select, I thought that it will be faster to do it at once…
Later… My delete lasted 5h15.
Now I try to do an optimize on the table, but innodb_tmpdir is on /tmp which is too small…