[13:20:36] topranks: o/ if you have a minute later on https://gerrit.wikimedia.org/r/c/operations/homer/public/+/784703 [13:29:16] elukey: +1 for the change, but it needs some homer additions for it to work. [13:29:31] give me a few minutes to prep the patch for that [13:30:15] sure thanks! [13:31:26] in the meantime I'll merge the puppet changes [14:45:05] hashar: _joe_ o/ gitlab ci templates are working fine for my CI purposes...but it sure does take a while to run the lints tests [14:45:23] each time it starts from a base buster image, apt installs conda and other deps, etc. [14:45:35] suuurrre would be nice to have a docker image with that stuff to start from! :D [14:46:36] they can mostly run in parallel, but each job takes about 3 minutes [14:47:27] mostly installing .deb packages [16:38:58] ottomata: you might take a look at https://gerrit.wikimedia.org/r/plugins/gitiles/integration/config/+/refs/heads/master/dockerfiles/ and think about how you could propose a new image there that would help you out. [16:39:44] bd808: ty. i think hashar told me not to use that because it was 'legacy' [16:39:49] heh [16:40:17] I'm using an image from there as part of Toolhub's production image build pipeline, so I hope not too legacy ;) [16:40:27] mostly just legacy in that gitlab is the future i thikn [16:40:44] but yeah that is probably the way to go while we wait for gitlab docker immage pipeline [16:40:58] the someday maybe but only kind of now future yeah [16:41:24] * bd808 has been burned as an early adopter here one too many times to rush to the hot stove again [16:41:45] ...differential :p ? [16:42:00] so far, gitlab is awesome. [16:42:55] We're going to replace mr1-eqiad, the mgmt network will be briefly unreachable and might trigger some alerting noise (even if parent is downtimed) [16:44:36] ottomata: yeah, differential was one source of scars that I'm still actually trying to replace from Toolforge. [16:46:01] The missing bit right now for making what you want "easy" is a way to push to a docker registry from gitlab. That is one of the blockers for moving pipelinelib things from jenkins to gitlab too, so likely to be solved eventually. [16:46:39] you could be cheeky and setup a pipelinelib project to publish your base image :) [16:47:43] Hello to anyone who understands about netbox! I need to use two new public IPs in Dallas and I suspect that I need to somehow tell netbox that those IPs are reserved before I declare them in the dns repo. Can anyone advise? https://gerrit.wikimedia.org/r/c/operations/dns/+/785182 [16:49:18] andrewbogott: Does this help? https://wikitech.wikimedia.org/wiki/DNS/Netbox#How_to_manually_allocate_a_special_purpose_IP_address_in_Netbox [16:50:15] (we're done with the mr1-eqiad replacement) [16:51:22] btullis: that probably helps! I will read once I'm done with this interview that's about to happen. ty [16:52:00] andrewbogott: You're welcome. [17:16:00] https://wikitech.wikimedia.org/w/api.php?origin=*&format=json&action=query&prop=info&list=prefixsearch&indexpageids=1&pssearch=Incident_documentation%2F2022-03&psprofile=classic&pslimit=max& [17:16:03] 6 results [17:16:15] but there should be 8 https://wikitech.wikimedia.org/wiki/Incident_status [17:16:25] I've purged them all, and the 7th was created several days ago [17:16:31] no dice.. [17:17:03] I switched to classic sometime last year as recommended as the default one seems lagged due to elastic difficulties of some sorts around wikitech [17:23:10] the default is still returning results from other years that don't match the prefix, e.g. https://wikitech.wikimedia.org/w/api.php?origin=*&format=json&action=query&prop=info&list=prefixsearch&indexpageids=1&pssearch=Incident_documentation%2F2022&pslimit=max& includes 2020- and 2021 and is too "smart". [17:27:44] * Krinkle realizes PEBKAC. [17:28:07] preexisting issues nonwithstanding, the problem is that the prefix is incorrect. These are redirects, not the actual pages. [17:28:13] nvm :)