2026-03-21T14:44:22 lkocman I forwarded you some npm problem with quiz build 2026-03-21T17:52:07 I'm just updating narwal7 to Leap 16 2026-03-21T17:52:42 zypper dup is still running, but terribly slow since kanidm-unixd-clients was updated 2026-03-21T17:53:58 a new ssh connection I opened after that package update greets me with 2026-03-21T17:54:01 I have no name!@narwal7 (static.o.o):~ > 2026-03-21T17:54:25 which sounds funny - but also indicates that name resolution is somehow broken currently 2026-03-21T17:56:41 try if removing it from /etc/nsswitch.conf speeds it up 2026-03-21T17:58:17 indeed, that helps :-) 2026-03-21T18:05:44 note to myself: if you revert this change before rebooting, you don't need to do a highstate via the saltmaster to be able to login 2026-03-21T18:06:03 (you may laugh nau ;-) 2026-03-21T18:06:38 anyway - update done, everything seems to work 2026-03-21T18:08:11 we did not have this on the machine we did last time 2026-03-21T18:11:55 actually I'm not completely sure - IIRC we also had problems to open an additional ssh connection. This time it took $timeout (probably 30) seconds to get one, maybe we just weren't patient enough last time 2026-03-21T18:13:58 I'll upgrade narwal8 now, let's see if it shows similar symptoms 2026-03-21T18:19:28 if you have a minute, can you have a look at the just-rebased MR 2558 please? 2026-03-21T18:20:39 lgtm 2026-03-21T18:21:13 :-) 2026-03-21T18:21:27 I also updated https://progress.opensuse.org/issues/189810 now, you can track there if you want 2026-03-21T18:26:16 I'm not sure how useful it is to add "$host upgraded" comments - they are not too useful, and a salt \* grains.get osrelease is probably easier to read ;-) (and always up-to-date) 2026-03-21T18:28:33 I'd only add "surprises" to the ticket 2026-03-21T18:29:10 speaking about surprises - on both narwals, zypper dup gives me (for several, but not all packages) 2026-03-21T18:29:13 error: Key 70af9e8139db7c82 (SuSE Package Signing Key ) expired on 2024-09-20 08:21:47 2026-03-21T18:29:52 I did a highstate before, so the salt-managed keys should be up to date 2026-03-21T18:36:58 zypper dup is now also terribly slow on narwal8. I didn't watch exactly when this started, but the kanidm package looks like a good candidate 2026-03-21T18:38:06 restarting kanidm-unixd helps :-) 2026-03-21T18:39:57 are you actually updating kanidm, because last time leap 16 had an older version 2026-03-21T18:41:43 well, actually it's a downgrade (1.9.1 -> 1.7.1 according to /v/l/zypp/history) 2026-03-21T18:43:52 right 2026-03-21T18:45:39 according to zypp history, one of the last packages that was quickly installed was kanidm-unixd-clients 2026-03-21T18:57:22 narwal8 completed, now over to narwal5 aka static_master 2026-03-21T19:15:58 narwal5 done, with the same kanidm-unixd slowdown 2026-03-21T19:21:06 if needed we can add the workaround to upgrade-os 2026-03-21T19:21:53 the problem is that ideally you need to restart kanidm-unixd in the middle of zypper dup 2026-03-21T19:22:04 doing it after zypper dup is possible, but wastes lots of time 2026-03-21T19:22:18 I mean to just patch nsswitch temporarily 2026-03-21T19:23:08 sadly that temporarily breaks ssh logins and sudo 2026-03-21T19:23:20 which you don't need for the duration of dup 2026-03-21T19:24:16 indeed 2026-03-21T19:25:00 if you update the script, please also extend the "after reboot" section with a mention of salt-call saltutil.refresh_grains 2026-03-21T21:01:12 all jekyll pages are now also served by narwal[5-8] :-) 2026-03-21T21:07:30 nice 2026-03-21T21:07:43 now we can upgrade jekyll without downtime :) 2026-03-21T21:07:59 right 2026-03-21T21:15:24 https://get.opensuse.org/tumbleweed/ is 404 2026-03-21T21:15:35 andhttps://get.opensuse.org/ returns static.o.o 2026-03-21T21:19:07 Mar 21 21:17:58 narwal5 nginx[7853]: nginx: [emerg] a duplicate default server for [::]:80 in /etc/nginx/vhosts.d/static.opensuse.org.conf:18 2026-03-21T21:19:08 nice[tm] 2026-03-21T21:20:04 ah because both web_static and web_jekyll have default_server 2026-03-21T21:20:16 I suggest removing that default from both 2026-03-21T21:27:19 good idea, MR 2737 created 2026-03-21T21:45:24 for the records: we have default_server in several other roles (calendar, discourse, mailman etc.), but it's unlikely that we'll deploy two of them on one server