[12:29:51] hi all i just came accross pypuppetdb which makes qureying the puppetdb via python a bit easier. the following snippiet queries the puppet db to find all managed puppet files which live under /etc/apt/sources.list.d/ which are not managed by apt::repository https://phabricator.wikimedia.org/P10000 (i also managed to grab paste 10000, so yay :) ) [12:30:05] volans: not sure if you have used this but may be of intrest ^^ [12:31:27] jbond42: yes known library [13:06:19] jbond42: TL;DR at the start we had an outdated puppetdb ourselves that was not compatible with newer versions of the library. But besides that, the voxpupuli group is mainly ruby-oriented, hence the library was mostly abandoned since 2017 (no active maintainer). In the last few months some activity has been restored and merge rights were given to some people, and there have been some new releases [13:06:25] and improvements. So I guess we could ... [13:06:27] ... start to re-consider it. FWIW is also the one used inside puppetboard, another project of theirs, with the same issues. [13:11:51] volans: ack thanks, for what its worth i am a member of voxpupuli, not a very active one, but it may help if we want to try and make some progress on some of theses things [13:13:22] lol, who knew! [13:13:33] I whish I had know that before :) [13:15:23] :) [13:23:05] jbond42: as a side-note you just become the maintainer of our puppetboard installation :-P [13:23:31] volans: lol Doh! [17:47:52] jbond42: something weird just happened with puppet-merge [17:48:00] /usr/local/bin/puppet-merge: 47: /usr/local/bin/puppet-merge: 3: parameter not set [17:48:19] and i think it didn't finish on all hosts maybe? [17:48:37] ottomata: yes sorry i just pushed a breaking fix, patch iminent [18:06:37] ottomata: the merge script should be working again but the output is a bit wrong [18:06:47] patch https://gerrit.wikimedia.org/r/#/c/operations/puppet/+/559924 [18:07:04] cdanis: can i bug you again to give that a quick sanity check [18:09:43] ty! [18:15:50] jbond42: ignore if you expect this [18:15:52] but i see [18:15:52] ERROR: puppet-merge on puppetmaster2001.codfw.wmnet (ops) failed [18:16:22] yes thats the thing im fixing now [18:16:25] k [18:19:17] heh I assumed same starting case on all puppetmasters, I guess can't always assume that ofc... oh and I suppose it always matters for labs-private as well [18:29:15] cdanis: i think i may have worked out the race condition, of to go eat dinner but thought you may be intrested https://phabricator.wikimedia.org/T241262 [18:30:22] but I thought that was why we provided the hash [18:33:02] ahh im not sure that happens in the refactor. the python script i think ends up using the hash but the shel script dosn't [18:33:11] 👀 [18:34:09] crap, you are right [18:35:13] go have dinner, I think I see an easy way to fix [18:36:14] grat thanks ping me if you need review [18:46:52] is puppet-merge fixed or [18:47:05] chaomodus: yes, aside from a few corner cases [18:48:49] oke doke [19:18:03] jbond42: okay i've fixed the issue, we now always pass a SHA to remote hosts https://gerrit.wikimedia.org/r/#/c/operations/puppet/+/559936 [19:22:02] thanks cdanis just checked looks good but missing a cat [19:22:05] oops thanks [19:22:14] next time I won't forget 😼 [19:22:18] :) [19:22:41] its what the internt is made off, cats all the way down :P [21:38:50] arrrrrggghhhhhh [21:39:35] https://phabricator.wikimedia.org/P10007 anyone want to solve the puzzle? :) [22:16:26] okay I believe things are much better in puppet-merge land :)