[09:51:29] Emperor, hashar o/ would you be available to assist (if needed) for https://phabricator.wikimedia.org/T412951 in about an hour? [09:51:43] otherwise we can do it tomorrow [09:53:17] (hour and a half) [09:53:21] hi, I am not available in an hour, I have lunch with my kids [09:53:57] from the task title, I am not sure what I can bring? [10:00:24] elukey: I'm trying to catch up after last week's TOIL, what time tomorrow would it be? My calendar is clear other than 12:00-12:30 UTC tomorrow [10:02:40] hashar: okok makes sense! It was just an extra set of eyes to validate the new backend for scap deployments, you can freely skip. In case I can warn you afterwards so you know that happened [10:02:43] I have crawled through the task and its parent. I remember Ahmon presenting to our team the patches he has done for Swift. Looks like we are moving to Ceph [10:03:08] if I got it right, a verification needed is to ensure the MW full images upload properly [10:03:38] the full rebuild and resulting deployment can be triggered using `scap` https://wikitech.wikimedia.org/wiki/MediaWiki_On_Kubernetes#Full_image_rebuild_and_deployment [10:04:05] which would be from the deployment server: scap sync-world --k8s-only --k8s-confirm-diff -D full_image_build:true [10:04:46] yep yep that part is clear, I thought that having you around to validate would have been a stronger signal that the new backend worked [10:04:55] but it is not mandatory, thanks :) [10:05:04] ahhh [10:05:24] Emperor: I'd use the MW infra window if possible, so every day at around 11 UTC [10:05:35] well I must tell I am merely an end user of scap this days and I don't have much details about mw-k8s [10:05:52] elukey: let's go with 11 UTC tomorrow then, I might have got to e.g. the questions on that ticket before then... [10:07:20] I'll be happy to join tomorrow (after 10UTC, I have a meeting before that) [10:08:02] hashar: not sure about it, buut I'll ping you as FYI so you are aware. If anything goes bad afterwards, during the next deployments, the more people know what we did the better :) [10:08:33] yeah you can loop me in [10:08:55] okok <3 [10:09:09] I can work on other things while it happens and interrupt [10:09:16] it is not like I am out of pet projects :b [10:09:32] :D [10:09:47] Emperor: okok! [10:10:47] <-- inbox somewhat on fire right now, even with all their usual filtering &c [10:17:04] Okok we can do it on Wed, I have a conflicting meeting tomorrow at 11:30 UTC [10:17:21] I'll send a meeting invite so we can sync [10:30:13] 👍 [17:18:14] topranks and I are going to enable the v6 glue for ns1 [17:18:35] once we update the zone file on our end, we will update the registrar immediately afterwards [17:19:21] it's a big change in some ways so we will keep an eye out pages [17:19:26] *for [17:26:39] topranks: anything else I am forgetting? [17:27:09] sukhe: no I don't think so [17:27:39] just a matter of publishing the AAAA our side then have the registrar add the glue AAAA record in .org [17:27:53] yeah sounds good [17:28:34] unfortunately we will need to wait for that a bit; in hindsight starting earlier in the day and emailing them is probably a better idea :P [17:28:43] something to keep in mind for ns0/2 [17:30:56] how you mean? [17:31:46] topranks: usually Markmonitor takes some time to reply. and the change doesn't go live until they add the glue, so it may be a while before we can actually test if something was broken [17:31:55] that may be late for you (not for me, it's noon here) [17:32:07] we'll see some requests before they do anything [17:34:20] well we might... basically some resolvers might do a refresh-before-expiry thing to the IP they already know, and ask for both A/AAAA [17:38:22] topranks: https://w.wiki/Hny8