[13:51:31] vgutierrez: nudge re https://gerrit.wikimedia.org/r/c/operations/puppet/+/1182652 - I'd like to avoid a merge conflict and to bring back beta cherry-picks to "only" 4 historic hacks, nothing recent :) [13:51:38] Beta currently depends on this patch to function. [14:01:14] that looks good [15:46:50] FYI, in a couple of minutes, I'll be restarting etcd on a single non-PyBal conf node in codfw for [0]. [15:46:50] * this will result in a brief disruption for clients connecting to that node (note: codfw is the read replica cluster, so conftool etc. writes will not be affected). [15:46:50] * as usual, once things are stable, I'll be following this with a rolling restart of all codfw-associated confds. [15:46:50] [0] https://phabricator.wikimedia.org/T352245 [15:46:50] cc: elukey arnoldokoth [15:47:19] \o/ ack [15:47:56] vgutierrez: I'll ping you here post-restart to check Liberica [15:48:45] Ack. Thank you. [15:53:57] moritzm: I'm going to do this now :) [15:55:51] ack! [15:55:55] I'm around [15:57:12] restarted [15:57:18] kicking various tires :) [16:01:05] alright, I think we're good [16:02:27] lol [16:02:45] vgutierrez: if you could take a quick look at Liberica in codfw-associated sites (eqsin, ulsfo, codfw) at your convenience, that would be awesome [16:03:27] Nov 10 15:56:20 lvs5004 libericad[518250]: time=2025-11-10T15:56:20.188Z level=WARN msg="ErrorCodeEventIndexCleared detected" key=/conftool/v1/pools/eqsin/ncredir/nginx [16:03:27] Nov 10 15:56:20 lvs5004 libericad[518250]: time=2025-11-10T15:56:20.188Z level=WARN msg="etcd watcher finished unexpectedly" service=ncredirlb_80 error="etcd error code 401. Retryable: false" [16:03:31] the regular 401s but all good [16:04:01] awesome, so exactly what we expend for some long-running idle watches that got interrupted [16:04:04] thank you! [16:04:11] *expect [16:05:49] alright, I'll get started on those confd restarts [16:05:52] thanks, all! [17:02:12] swfrench-wmf: BTW.. quick way of checking impacted instances: https://grafana.wikimedia.org/goto/OY5yZlzDg?orgId=1 [17:06:38] vgutierrez: ah, thank you! this is quite handy [19:43:32] anyone care to stamp https://gerrit.wikimedia.org/r/c/operations/puppet/+/1203462 ? [19:46:36] TIL. I think can be helpful [19:48:14] it's nice, I've also started using fzf recently which we already have in that list [19:49:11] (for those who try to run it, it is actually batcat and not bad, the executable) [19:49:18] er s/bad/bat [20:01:03] moritzm: sigh, you are right [20:01:05] fixing [20:02:06] yeah sorry, I for one forgot about the buster hosts completely. I guess a conditional for bullseye and above is all we need there I guess [20:02:33] yep [20:03:59] https://gerrit.wikimedia.org/r/c/operations/puppet/+/1203512 [20:04:10] thx! [20:08:26] hopefully these are history in a few weeks... [20:08:48] I can't even ssh to them anymore 😅 had to use cumin to see if puppet had indeed failed