[08:39:12] <dcausse>	 relforge is failing with "this action would add [6] total shards, but this cluster currently has [2000]/[2000] maximum shards open"
[08:41:38] <dcausse>	 somehow it has plenty of wikis there e.g. lfnwiki_content
[08:43:28] <dcausse>	 created on jun 13 with an empty mapping (so most likely autocreated), might be some testing I suppose
[08:43:37] <dcausse>	 going to delete them
[09:53:15] <dcausse>	 lunch
[12:02:57] <gehel>	 inflatador: I'm cancelling our 1:1, conflicting meeting
[12:43:49] <inflatador>	 gehel ACK
[12:43:49] <inflatador>	 <o/
[12:59:02] <inflatador>	 on wiki, is there normally a lag between updating a template and it actually displaying on the pages that use that template?
[12:59:21] <inflatador>	 nm, looks like there is
[13:01:25] <gehel>	 inflatador: you can do a "null edit" on a page to force it to re-render
[13:02:21] <gehel>	 https://www.mediawiki.org/wiki/Help:Dummy_edit#A_null_edit
[13:47:43] <dcausse>	 pfischer: I've done another round of review, should be close to be testable imo, please lemme know if you have questions
[13:50:40] <pfischer>	 Sure, I’ll have a look right away. Thank you!
[13:58:19] <pfischer>	 dcausse: Regarding https://gerrit.wikimedia.org/r/c/wikidata/query/rdf/+/1033386/comment/7457a5f9_25a340ea/ - Should the pipeline fail if there is no side-output for a subgraph?
[14:01:33] <topranks>	 inflatador: thanks for the help to prep for today's switch upgrade (T365993)
[14:01:33] <stashbot>	 T365993: Upgrade EVPN switches Eqiad row E-F to JunOS 22.2 - lsw1-e1-eqiad - https://phabricator.wikimedia.org/T365993
[14:02:30] <topranks>	 unfortunately we have to postpone the works until next week
[14:02:38] <topranks>	 so I guess elastic1104, elastic1089 & elastic1090 can be re-pooled for now 
[14:02:39] <inflatador>	 topranks and thank you for your comms. You make it easy for me ;)
[14:02:50] <topranks>	 apologies for the inconvenience 
[14:03:03] <topranks>	 np :) 
[14:03:09] <dcausse>	 pfischer: yes Map.apply should fail for us I think
[14:03:09] <inflatador>	 np, we have plenty of capacity. Hope everything went well!
[14:03:51] <topranks>	 yeah just need to move it out to make life easier for some of the teams 
[14:04:25] <topranks>	 we're still doing E2 tomorrow which has elastic1091 & elastic1092
[14:06:34] <inflatador>	 ACK, I have those on my list to ban today, along w/wdqs1018 & 1020
[14:08:11] <topranks>	 great, thanks
[14:17:14] <inflatador>	 ryankemper no rush, but https://gerrit.wikimedia.org/r/c/operations/puppet/+/1051369 is ready for review if/when you have time
[14:26:47] <gehel>	 pfischer: your opinion on T367391#9943556 would be welcomed!
[14:26:48] <stashbot>	 T367391:  Setup a test project to validate upload to the Gitlab package registry - https://phabricator.wikimedia.org/T367391
[16:07:01] <pfischer>	 gehel: looking
[16:26:17] <inflatador>	 workout, back in ~40
[17:11:49] <inflatador>	 back
[17:14:11] <dcausse>	 dinner
[17:48:42] <inflatador>	 lunch, back in time for pairing in ~45
[18:32:08] <inflatador>	 back
[18:34:43] <inflatador>	 ebernhardson ryankemper I'm in pairing if y'all feel like joining
[18:40:33] <ryankemper>	 inflatador: https://phabricator.wikimedia.org/T263073
[18:42:33] <inflatador>	 hieradata/common/profile/services_proxy/envoy.yaml
[21:13:43] <inflatador>	 ryankemper posted a follow-up on T368972 . TLDR it does look like envoy can do this, but no one else seems to be using it that way
[21:13:44] <stashbot>	 T368972: Use Envoy instead of LVS to route internal federation traffic for WDQS - https://phabricator.wikimedia.org/T368972
[21:15:31] <ryankemper>	 inflatador: it would need some way to get etcd state to determine what hosts are depooled though right? i.e. we need some mechanism whereby if a host is depooled from pybal that it's also not going to be included in envoy's pool
[21:19:48] <inflatador>	 ryankemper Ideally yes, although the health checks will catch short outages. Longer outages would require a puppet patch to remove the downed host
[21:25:25] <inflatador>	 the bigger problem with this approach as that we now have 2 places to configure LB config
[21:26:10] <inflatador>	 err..."2 LBs to configure"
[21:27:49] <cdanis>	 ryankemper: inflatador: btw the serviceops internship project that just concluded was an Envoy control plane that reads etcd confctl
[21:29:30] <cdanis>	 well, that was the intent anyway, I don't know what milestones were actually reached :) https://gitlab.wikimedia.org/repos/sre/sophroid
[21:31:29] <inflatador>	 cdanis yeah, I saw that presentation and was actually thinking of that, although I didn't wanna open that can of worms quite yet ;)
[21:31:35] <cdanis>	 ah ok :) nvm
[21:33:00] <inflatador>	 I'd love to end up there at some point though. I guess we can ask around and see if non-k8s envoy is in scope