[08:25:37] <Emperor>	 trixie> I couldn't get the installer to go on sretest2010 yesterday, complaining about not finding the right kernel version, which I'm guessing means the install image needs updating to reflect the point release (few more notes on the bottom of T394357 )
[08:25:38] <stashbot>	 T394357: Q4:rack/setup/install sretest2010 Config J 1P test host - https://phabricator.wikimedia.org/T394357
[08:40:46] <hashar>	 codesearch does not respond and I have filed https://phabricator.wikimedia.org/T404163 for it
[08:40:54] <hashar>	 I am not quite sure who has access to it
[08:47:41] <vgutierrez>	 hashar: https://www.mediawiki.org/wiki/Codesearch#Administration
[08:48:30] <hashar>	 vgutierrez: thanks! I have wrote my message in the wrong channel, that was intended for #wikimedia-cloud-admin :)
[08:48:45] <hashar>	 (I also lack global root but that is another story :]  )
[08:48:51] <vgutierrez>	 ☕? :)
[13:24:29] <andrewbogott>	 moritzm: are you available to build a fresh Trixie install image? Or is something else happening with the kernel package issue?
[13:28:44] <moritzm>	 not sure what you mean? 
[13:29:06] <moritzm>	 the trixie d-i image got updated for the 13.1 point release already
[13:30:00] <andrewbogott>	 as of ~24 hours ago the installation failed with 'No kernel modules were found', Emperor reports the same issue. Maybe it's been fixed since we last tried?
[13:30:08] <andrewbogott>	 (or maybe that's unrelated to the build)
[13:30:49] <andrewbogott>	 I can re-run the install if you want to connect and see what I'm seeing
[13:31:19] <moritzm>	 that's typically the error when d-i has an old image (like 13.0) boots up with that kernel, but that at run time attempts to install the 13.0 udebs which are no longer on the Debian mirrors
[13:31:45] <moritzm>	 this was most likely fixed by yesterdays rebuild, so please retry
[13:31:55] <andrewbogott>	 ok, will retry, thanks!
[13:41:23] <andrewbogott>	 moritzm: same behavior just now.cloudcephmon2004-dev.mgmt.codfw.wmnet if you want to investigate.
[13:42:29] <andrewbogott>	 resulting from  "sudo cookbook --no-locks sre.hosts.reimage --os trixie cloudcephmon2004-dev"
[14:03:06] <moritzm>	 in a meeting, will check later
[14:13:06] <andrewbogott>	 thanks!
[15:00:07] <moritzm>	 andrewbogott: I found the issue, it was actually unrelated to the point update
[15:00:29] <moritzm>	 there's a setting which was needed while trixie d-i was in development (when I did the early prep work)
[15:00:49] <andrewbogott>	 thank you! Should I try again now or are you still in progress?
[15:01:04] <moritzm>	 and I had missed to revert it and now that development for next Debian releases started to gain traction, it's now visible again
[15:01:07] <moritzm>	 https://gerrit.wikimedia.org/r/1187012
[15:01:37] <moritzm>	 I'll deploy and will let you know when you can kick it off again
[15:04:27] <andrewbogott>	 great, thanks moritzm 
[15:10:12] <moritzm>	 andrewbogott: patch is merged and I forced a Puppet run, it should work now
[15:10:23] <andrewbogott>	 I'll give it a whirl
[15:21:24] <andrewbogott>	 not finished installing yet, but it's definitely past the spot that failed before
[15:25:12] <moritzm>	 excellent
[16:03:06] <andrewbogott>	 install completed, thanks again moritzm 
[16:42:12] <moritzm>	 great!
[16:51:54] <jasmine_>	 hi folks, just a reminder that serviceops will be running the DC Switchover live test tomorrow during the late UTC window. 
[20:53:08] <swfrench-wmf>	 (cross-posting in a couple of places) we've recently made some changes that temporarily have rest-gateway serving active/passive (i.e., all traffic is served from eqiad, currently).
[20:53:08] <swfrench-wmf>	 I'll be keeping an eye on things throughout the day, but just in case something happens while I'm not available, there are details in [0] covering how to return rest-gateway to "active/active-like" behavior.
[20:53:08] <swfrench-wmf>	 [0] https://phabricator.wikimedia.org/T400131#11169897
[21:06:42] <sukhe>	 nicely done!