[12:47:01] PXE seems to be failing for all bookworm installs [12:47:23] probably something related to a new point release or configuration? [12:47:36] but heads up for those about to reimage se servers [12:51:12] there is a tftp alert, that could be another lead [12:54:53] there hasn't been a Debian point release, so that can be ruled out. what tftp alert do you mean? [12:55:50] oh I've asked on foundations, lets continue here! [12:58:36] https://alerts.wikimedia.org/?q=alertname%3DTFTP%20service&q=team%3Dsre&q=%40receiver%3Dirc-spam [12:58:43] ^ moritz [12:59:26] however, install was failing for eqiad servers and only for bookworm, not bullseye [12:59:52] I got it also failing for codfw bookworm [13:06:52] had a loiok: on install2004 atftpd failed to restart, but since the package only ships a legacy sysvinit init script, the failure isn't correctly alerted [13:07:09] the failing restart is also caused by being a crappy init script: [13:07:27] it didn't terminate quickly enough for the restart, which led to a failing start [13:08:10] I bounced it manually now, but we should really fix this by adding a proper systemd unit via Puppet or by updating the deb, I'll open a task [13:08:29] arnaudb: you can retry, it should work now [13:09:01] lets retry then :) thanks [13:12:11] this is actually fixed in bookworm, atftpd ships a proper systemd unit in that version, so no task needed, this will be fixed when we move off Bullseye for the install servers [14:43:51] https://sal.toolforge.org/log/4mdBBo8BGiVuUzOdiPDR → lets downtime it for 7 more days [14:45:54] thanks!