[03:20] after 3 failed installs on 3 different isos, i have to question the integrity of subiquity [03:20] is there a boot flag to use the old debian installer instead [04:04] Obscenity: did you both verify the iso checksums, and the data written to the installer media? which images did you use? what failed exactly? [04:05] well i tried the jammy daily first, then the jammy beta, and then the old 21.10 stable [04:05] it faled duing the entering of my username when asked [04:10] failed how? and you didn't answer the first question [04:36] no, i cantm and it has a popup that the installer failed, and i can send a report to canonical, which i did each time too [04:39] well, verify the checksums for downloaded iso and installer image writtern and see whether this helps. [04:40] i cant [04:40] i do have an idea of what to change though, so ill try that [04:47] if you'll be looking for support with this later, be sure to explain "i can't". [04:48] yeah i got it [04:48] subiquity crashes if i enter a mirror, then enter the user information screen [04:48] so no mirror [04:49] and i cant because the image is stored remotely on a VM Host [04:49] thats why, hah [04:49] it gives me a UUID that is not in valid UUID format [04:49] thats about it [04:50] what is "it" that gives you a UUID? [04:51] you should talk to whomanaged the VM Host to have them verify integrity of the .iso [04:51] *who manages [04:52] that would probably take several days, but it works now if i just dont use a mirror [04:52] https://www.virtualizor.com/ [04:52] that is "it" [04:55] this may be bug 1883401 [04:55] Bug 1883401 in subiquity (Ubuntu) "crash due to bad custom mirror" [Low, Triaged] https://launchpad.net/bugs/1883401 [04:56] but your descriptions is so imprecise, it's hard to tell [04:57] ooof, abandoned bug from 2020 [04:57] unlucky [04:58] at least i can add the mirror back to apt later [04:58] there is also bug 1874248 and bug 1860352 [04:58] Bug 1874248 in subiquity "stage-curthooks/001-configure-apt: SUCCESS: Status.FAIL" [Undecided, New] https://launchpad.net/bugs/1874248 [04:58] Bug 1860352 in subiquity "User supplied mirror server not verified, no errors reported" [Undecided, New] https://launchpad.net/bugs/1860352 [05:00] 1860352 makes sense since https-transport is not installed by default, which is pretty insane if you ask me [05:00] i don't see why you're saying the first bug report was 'abandoned'. it was just considered low importance, possibly because the error is a result of a bad environmental configuration [05:03] i just dont know what Triaged means so I assumed it was kicked aside [05:04] package apt-transport-https is no longer needed for adding https support to apt since apt 1.5 [05:05] huh, when i dont add it i always get [Ign] or [Err] in the apt output [05:05] "Triaged" status is the state where a bug report is under examination. [05:08] it is unclear which ubuntu release or exact output you're referring to. [05:11] since as long as i remember, at least from 10.04 to 18.04 if not newer [05:11] im going to try it when im logged in and ill let you know [05:11] the output i mean [05:27] huh, totally works in 22.04 === brlin_ is now known as brlin [07:34] good morning [07:40] mirespace: o/ [07:46] hi utkarsh2102 o/ [11:25] Walex: What directory would you pass to debootstrap for a side-by-side install though? [13:14] sdeziel: on the squid vs squid-openssl, I am not sure if it is just regarding historical reasons. I will dig around though :) squid-openssl was in main in impish though [13:20] Hey folks, if we have a patch that we'd like to see backported to focal's systemd, what would the appropriate process be? I will file a bug, but I'm not sure if those are regularly reviewed. [13:23] in any case, it all starts with a bug [13:23] I think bugs with patches get a special tag even [13:24] Odd_Bloke: did you check the recent systemd bugs, to see if there's something similar? https://bugs.launchpad.net/ubuntu/+source/systemd/+bugs?orderby=importance&start=0 [13:24] and then ping in ubuntu-devel perhaps, as systemd is not under server [13:27] lotuspsychje: Yup, no mention of it! [13:27] ahasenack: Thanks! It's Foundations still, presumably? [13:27] you presume correctly [14:26] sdeziel: hey. I don't think it's historical, TBH. I also agree that we should probably stick to the openssl version because that seems to be the preferred crypto suite. as athos said, squid-openssl was in main and has been demoted to universe; I'm trying to find more info on why this happened [14:28] athos: oh, I had not seen this re squid-openssl in impish [14:29] sergiodj: cool, let me know if I can help in any way [14:30] squid-openssl has been added quite recently. upstream is still working on the openssl support TBH, so that's something we should consider before promoting the package to main again [14:30] sdeziel: thanks for bringing this up [14:39] yeah, I'd be curious to know why it was demoted that recently? [16:45] athos: sergiodj: https://launchpad.net/ubuntu/jammy/amd64/squid-openssl [16:45] the "publishing" page for binaries [17:23] https://github.com/openzfs/zfs/issues/326 [17:23] Issue 326 in openzfs/zfs "Support fallocate(2)" [Closed] [18:07] ahasenack: thanks. I couldn't find an explanation on the demotion in the page, though [18:46] sdeziel: so, the reason for the demotion was that squid-openssl is not an rdep for anything & it's not in a seed [18:47] sergiodj: ah, I see, thx [18:48] with that in mind, and knowing that upstream is still working on the patches to support openssl 3, I think a good plan ahead is to wait until upstream finishes working on the PR, backport its missing bits to squid on Jammy, and then work on getting squid-openssl into main [18:49] sergiodj: and realistically, that'd be for Jammy+1 right? [18:50] sdeziel: Jammy+1 would have an MIR, but I believe we should/could also aim for Jammy given that squid is an important package [18:50] sergiodj: that'd be nice indeed :) [18:50] I will write something about squid-openssl in the release notes [18:50] as suggested by ahasenack [18:57] sergiodj: if the plan is to have the OpenSSL one in main, will there be a demotion of the GnuTLS one? (I'd be in favor of minimal maintenance IMHO) [18:58] sdeziel: the best outcome would be to have a squid-gnutls package (in universe), and have squid link against openssl IMO. I will talk to the debian maintainer and see what he thinks [18:59] sergiodj: works for me [21:11] maybe also look at what is best supported upstream... === eitteabs is now known as sbeattie