[00:18] at is also doing it. I don't see how that got missed before :-/ [00:20] I mean how I missed it before [00:28] at uploaded === locutusofborg_ is now known as locutusofborg [06:59] No more Postfix prompt on upgrade \o/ === alan_g_ is now known as alan_g [11:39] good morning [13:13] yay for no more postfix! i was just writing up that folks might see a Postfix prompt when doing an upgrade of cloud images [16:11] Is anybody looking into the daily builds of the desktop ISO for amd64? The latest one I see is from April 9. [16:12] bdmurray ^ [16:26] GunnarHj: I think that's a known issue that has been affecting ISO builds for all flavors for a while. For instance, see https://cdimage.ubuntu.com/kubuntu/daily-live/pending/ [16:26] and https://people.canonical.com/~ubuntu-archive/cd-build-logs/ubuntu/jammy/daily-live-20220414.log [16:28] Thanks jbicha. But then my question still stands. [16:29] There was an attempt to fix it at the cloud level this morning, which, err, seems to have mostly taken out the entire cloud's networking [16:29] Err, assuming we're talking about the same thing (observable as livefs builds failing trying to talk to people.canonical.com) [16:31] yeah, I think this log shows the issue better: https://launchpad.net/~ubuntu-cdimage/+livefs/ubuntu/jammy/ubuntu/+build/338594 [16:32] cjwatson: I'm merely talking about the result. As the kubuntu page which jbicha posted. Or http://cdimage.ubuntu.com/daily-live/pending/ [16:33] Definitely being worked on, but hard to give a projection right now. [16:33] cjwatson: Thanks, that answered my question. [16:34] That said, I filed a ticket the other day with an alternate workaround (which would have the effect of making builder resets significantly less reliable in the current state, but if the builder managed to reset at all then it would have correct metadata). Depending on how the current firefighting goes I may see about lighting a fire under that one. [16:35] lol at the mixed metaphors [16:35] Deliberately :) [16:35] (Very roughly: the cloud metadata service issue that causes this also has the effect that the builder's usual SSH keys don't get installed, so we can try sshing to the builder after resetting it as a test for whether it came up properly, and if not we can disable it instead.) [16:36] Canonical staff can see that side of things as https://portal.admin.canonical.com/C149289 [17:21] xnox: are you around? I ran into LP: #1969145 [17:22] Launchpad bug 1969145 in oss4 (Ubuntu) "oss4 autopkgtest regression: unrecognized command-line option ‘-mfloat-abi=hard’" [High, New] https://launchpad.net/bugs/1969145 === rs200974 is now known as rs20097 === cookie is now known as ckie [22:47] jbicha: le sigh, i bet we only should specify that on armhf; and not on all arches. and like the package is buggy; and only worked whilst compilers were not strict about such things [22:48] also pondering why we specify that at all, given that all our arches default to hard float abi i think. [22:49] jbicha: clearly this is not an issue with glib; so just ask release team to force-skiptest glib; and we can look into fixing oss4 builds later === TheMaster is now known as Unit193