[05:51:17] We are going to take over the mediawiki-config repo for the x1 failover. Please coordinate with us before merging/Deploying something [06:05:06] The failover is done, deployments and merges on mediawiki-config can resume as usual [06:09:39] nginx is the only submodule left in operations/puppet :) [06:12:14] nice! (to both) [06:12:22] \o/ [06:28:02] should we also move nginx to operations/puppet? Shouldn't be difficult [06:28:10] so we'll be submodules free finally [06:28:18] I would love it, personally [06:31:40] I have tested the procedure that Joe proposed (rm submodule, copy to environments/prod, merge, copy to modules/etc.) 4 times, no glitch [06:32:21] just wondering if there is some reason outside my knowledge that we want nginx kept as a submodule but I can't imagine why that would be [07:08:50] elukey: +1 on doing that, I doubt anyone is using that outside of us, maybe just put in on next SRE meeting's agenda to give anyone a chance to object [07:24:21] I've just updated our buster d-i image to rc3, if anyone is installing a buster host today and runs into any issues, please let me know [09:57:44] FYI all the YARN NodeManager Node-State [09:58:04] unknowns are due to icinga still catching up with the removal of a per host check [09:58:16] should go away after the next puppet runs on icinga1001 [10:45:04] @all i plan to deploy the hira search order change today at 11:30 UTC. please vocide any concerns now if you have them https://gerrit.wikimedia.org/r/c/operations/puppet/+/511686 [10:46:57] sounds good [10:50:39] I'm looking at the snapshot differences from the puppet compiler [10:50:46] https://puppet-compiler.wmflabs.org/compiler1002/17020/snapshot1006.eqiad.wmnet/ [10:51:04] meh it doesn't matter, just the order [10:51:12] fine by me then [10:51:31] ack [11:00:03] jbond42: +1 good for me [11:30:13] thanks all pushing now [20:35:07] andrewbogott: o/ [20:35:09] https://phabricator.wikimedia.org/T225128#5304749 [20:35:18] is there anything special I should know here? [20:35:44] i want to role spare::system to cloudvirtan so we can move them back to analytics prod [20:35:51] don't want to accidnetally trigger any alerts [20:36:04] Those boxes currently have two networks hooked up, for control plane and virt plane. You probably only need one if you're going to manage them as bare metal. [20:36:32] I would just downtime all the existing hostnames for a year or two, assuming that they'll be renamed/rebuilt before the downtimes expire. [20:36:45] ok, they will with new hostnames. [20:36:48] I don't think they'll trigger any alerts that aren't associated with the hostnames directly. [20:36:50] ok great. [21:01:32] anyone able to look at the error from I think our Debian mirror on sodium? Subject: [ftpsync@sodium] ERROR: rsync errors [21:03:14] cdanis: yea, i can look [21:04:23] too many deletions for a --max-delete-limit .. but what did it want to delete [21:09:28] it's various installers from sid that get removed [21:11:04] i think it will fix itself in <12h, next time we get pushed to. it deleted the max number of files it was allowed and the remaining ones should just be deleted next time [21:12:15] ftpsync = rsync but they are pushing to us (Debian mirror, as opposed to the Ubuntu mirror where we pull) [21:16:27] ahh that makes sense [21:16:39] I was thinking it was probably something having to do with it being close to buster release time [21:16:47] thanks :) [21:18:23] yw. it deleted 40000 and skipped 2254 or so.. should be fine next run [21:25:00] busterday soon [21:28:57] https://disney.fandom.com/wiki/Buster_(Toy_Story) [21:38:23] heh