Project for issues affecting Puppet configuration and general Puppet code.
Dec 26 2018
We are officially running puppet6 everywhere now :)
Dec 20 2018
Oct 3 2018
Sep 21 2018
I think i got them all, though i could have missed some.
Jul 31 2018
Puppet runs down to 15secs on mw3.
Just to have it somewhere, @Paladox managed to get it to 40 seconds
Jul 26 2018
You given the list of stuff to fix, fix it :)
Apr 4 2018
Your welcome :)
This is finally done! :)
Apr 3 2018
Whenever I'll upgrade (somewhere this evening) I will grant John access to RN as well.
Apr 2 2018
@Southparkfan We have changed from using puppetdb 2 to puppetdb 4 with https://github.com/miraheze/puppet/pull/640/files (puppetdb is disabled on puppet1 for the time being) this should help with keeping the heap down as wmf have found lower ram usage because of the upgrade. Though still needs 2gb of ram due to postgresql on there and puppetdb.
@Southparkfan puppetdb (which is java) and then we have postgresql on there too which uses the ram too. And then there's puppet when the nodes hit puppet when they run.
Before I approve or decline this, what is exactly using so much memory? Java code, Postgres, puppet, apache? Have attempts been made to reduce memory usage?
Approved by me.
1G and 1 core on puppet1 was always a push imho anyway.
Mar 24 2018
.git permissions messed up for some reason. Also in future, just clear out /config/, it saves a hell of a lot of time.
@John Error was
Well the purpose of my comment was for you to give us more information but that’s fine anyway.
Good point. I guess the purpose of this task is resolved then.
Without any information, it’s impossible to judge and you just have an open issued assigned to SPF serving no purpose for anyone?
Fixed by refreshing /srv/mediawiki, but could this reoccur again soon? (Moving to "Normal" since it's now fixed, but keeping open in case it's a more "permanent" issue)
mw1 Debian Puppet CRITICAL 2018-03-24 00:34:47 1d 8h 10m 52s 3/3 CRITICAL: Puppet has 1 failures. Last run 28 seconds ago with 1 failures. Failed resources (up to 3 shown): Exec[git_pull_MediaWiki config]
Mar 9 2018
Mar 8 2018
@labster puppet will likly to fail if there is not enough heap for puppetdb (which puppetdb could be restarted)
Remind me again why needing 2x more RAM counts as a software upgrade.
Assigned to @Southparkfan for 1) approval/review of an upgrade 2) to be around when the upgrade would happen
Mar 7 2018
So, it looks like PuppetDB and exported resources are working on test1. Therefore, it is now time to upgrade puppet1 to Debian 9 and Puppet4. But before, @Paladox says that PuppetDB needs more RAM, so we should probably upgrade puppet1. What do you think @Southparkfan ?
Jan 14 2018
Update, we have started testing this module now :).
Jan 8 2018
Hi, pull requests are accepted on GitHub. Puppet repo in phab is for redundancy, and having it closing the bugs as appropriate (which... didn't work as intended for now).
Jan 6 2018
You mean applying changes not pulling them?
Another update: The certificate issue has been resolved, only now puppet does not seem to be reading /etc/puppet properly, therefore it isn't pulling any changes.
Jan 4 2018
Just to have an update on this task:
Jan 1 2018
Realistic goal, we need this for PHP7 and the new MediaWiki upgrade, so this should be done before June.