Tech WG/Ops Diary: Difference between revisions

(ES restart)
(2 intermediate revisions by the same user not shown)
Line 237: Line 237:
}}
}}


== 24-04-17 ==
== 24-04-17 Elastic Search issues ==


Manisha: Saw a warning in the admin panel:
Manisha: Saw a warning in the admin panel:
Line 256: Line 256:


Had to wait a tiny bit before [https://discord.com/channels/1049136631065628772/1049184335514828860/1230220033481773087 the warning in the admin panel] disappeared.
Had to wait a tiny bit before [https://discord.com/channels/1049136631065628772/1049184335514828860/1230220033481773087 the warning in the admin panel] disappeared.
== 24-04-30 Sidekiq scheduler queue issues ==
Manisha: Noticed this warning about sidekiq [[File:Sidekiq issues.png|frameless]]
And saw that we were hitting our swap limits.
Restarted services. Sidekiq status is active but we are still near our swap limits and the sidqkiq warning in the admin panel hasn't gone away. [https://discord.com/channels/1049136631065628772/1049184335514828860/1235046062171885568 Link to discussion on the technical-wg channel].
[https://github.com/mastodon/mastodon/issues/20710#issuecomment-1329680431 Maybe this github discussion would help?] but on checking the sidekiq dashboard, while the no. of processed requests did plummet today, I do see sidekiq running. The warning about the scheduler queue(s) is puzzling as there are scheduled jobs in the sidekiq container! hmmm
Something is definitely up. sidekiq's status is <code> active (exited)</code>. I restarted ES and that helped free up swap but restarting sideqik isn't helping and the warning isn't going away :(
Jonny: we are sorta always at the limit of RAM and swap, cost of the ES service. disk is full! cron job likely not working
Manisha: ran this <code> RAILS_ENV=production /home/mastodon/live/bin/tootctl media remove --remove-headers</code> as mastodon user. It freed up 10.5 GB and the sidekiq warning went away! :)