[00:01] -queuebot:#ubuntu-release- Builds: Ubuntu Server Subiquity amd64 [Bionic Final] has been updated (20180425.1) [00:04] slangasek: Knock yourself out. [00:05] And did you spot the bug or not? ;) [00:07] What bug? [00:09] tsimonq2: I mean, I've spotted the bug but I don't have a good fix for it. The issue is that SUPPORTED_SEEDS = ["all"] which encompasses everything and trumps the NON_LTS_SEEDS [00:18] tsimonq2: I think I might try reordering the list and special-casing 'all' [00:18] slangasek: Thanks. [00:19] (just got home and I'm packing for my flight tomorrow) [00:41] tsimonq2: https://code.launchpad.net/~vorlon/ubuntu-archive-publishing/lubuntu-next-support-fixup/+merge/344358 untested [00:46] mwhudson: hey! Can you help with testing the bionic subiquity images? [00:46] sil2100: sure, although in some sense i'm a very bad person to test them [00:47] sil2100: http://iso.qa.ubuntu.com/qatracker/milestones/389/builds/171072/testcases ? [00:47] uh http://iso.qa.ubuntu.com/qatracker/milestones/389/builds/171072/downloads has the wrong link [00:48] Oh [00:48] Let me try fixing that [00:48] bionic-live-amd64.iso.zsync [00:48] s/live/live-server/ i think [00:49] * sil2100 fixes that [00:50] infinity: https://code.launchpad.net/~vorlon/ubuntu-archive-publishing/lubuntu-next-support-fixup/+merge/344358 is an untested stab at the lubuntu support metadata. I don't see a trivial way to test this [00:52] mwhudson: should be fixed [00:53] sil2100: looks better thanks [01:09] Ok, I go take a few hours or sleep, but I'll start off earlier to help out with testing [01:09] o/ [01:25] oh phew the installed system boot was just a bit slow for some reason in this test [02:05] Tick tick tick. (Exciting) [02:08] slangasek: no it's tjaalton you want for linux-oem === Ursinha_ is now known as Ursinha === seyeongkim_ is now known as seyeongkim === sergiusens_ is now known as sergiusens === WillMoogle_ is now known as WillMoogle === jamespage_ is now known as jamespage === essembe is now known as sbeattie [03:35] -queuebot:#ubuntu-release- Unapproved: java-atk-wrapper (bionic-proposed/main) [0.33.3-20 => 0.33.3-20ubuntu0.1] (core) [03:36] -queuebot:#ubuntu-release- Builds: Ubuntu Kylin Desktop amd64 [Bionic Final] has been updated (20180426) [03:36] -queuebot:#ubuntu-release- Builds: Ubuntu Kylin Desktop i386 [Bionic Final] has been updated (20180426) [03:41] ^ um, the iso cronjob isn't still on, is it? [03:46] slangasek: ping ^ since you might still be around [03:46] it was turned off well before; just checked and it's off. someone must have manually requested that rebuild [03:46] despite it being a no-op [03:47] handsome_feng: ^^ was that you? [03:47] yeah, I diffed the manifest's and they were the same [03:47] were there test results already reported from the previous build? I can axe it from the iso tracker [03:48] slangasek: You mean rebuild the iso? [03:49] handsome_feng: yes [03:50] slangasek: It's zhangchao in our team, we found some thing wrong in the iso, so we rebuild it [03:50] handsome_feng: could you be more specific about what you think was wrong? [03:51] handsome_feng: ok but you rebuilt it against an archive that has not changed [03:53] sorry to change the topic but I'm going to bed… [03:53] We found that there are some permissions related errors in the x86_64 live system, we thought it maybe someting wrong when build the iso [03:54] handsome_feng: what permissions where? let us help you debug this, please [03:54] handsome_feng: I think there are some concerns about Ubuntu Kylin's request to offer 5 years of support for this LTS so please discuss that here too [03:55] good night all [03:56] jbicha: good night [04:09] slangasek: https://paste.ubuntu.com/p/x9cwtjBGGh/ [04:15] handsome_feng: when these dconf errors happen, can you please show the uid of the at-spi-bus-launcher process (e.g.: 'ps awxu | grep 3948') and the permissions on /run/user/999/ and /run/user/999/dconf (if it exists)? [04:16] crossbuild-essential-amd64/amd64 unsatisfiable Depends: gcc-x86-64-linux-gnu (>= 7.3) [04:16] handsome_feng: also: what is the effect of these permission failures? [04:16] is this only checked on amd64? I mean, this was true before for arm64 and ppc64el on these archs as well [04:17] doko: that package only exists on !amd64 [04:18] so do you intend for the amd64 package to pull in a foreign-arch foo-to-x86 compiler? [04:18] slangasek: the effect is that the peony, ukui-panel, ukui-menu launch failed [04:18] doko: or, it's an arch: all package; do you intend this only to be installed on !amd64? [04:19] no. but as I said, that was already the case with crossbuild-essential-arm64/arm64 [04:19] yes, the latter [04:19] doko: right - so it's possible the dependencies of arch: all packages are only enforced on amd64 (this is somewhat configurable in britney, I don't remember if that's the current value). We can consider overriding this [04:20] handsome_feng: when did this last work? [04:20] ahh, ok. so maybe no need to make it Arch:any [04:21] slangasek: the daily 0421 [04:24] we tested it in daily 0421. And it's ok. [04:27] slangasek: https://paste.ubuntu.com/p/3JNzF5m4V8/ [04:29] slangasek: and this error only happend in x86_64, the i386 works fine in live system. [04:29] handsome_feng: ok. can you also test the 20180425 image to see if it happens there? that will help us better bisect the delta between the images [04:30] 109 packages changed between 20180421 and 20180426 so that will take some time to dig through [04:32] handsome_feng: ok, so something has pre-created the /sur/user/999/dconf/user file as root. not good.. [04:33] slangasek: OK, we will test it ASAP, but it will take some time to download the iso [04:34] handsome_feng: yes, understood. I will keep digging into the changesets in the meantime. Also, the squashfs from the 20180424 build is still available from launchpad; if 20180425 is broken we should try to download that and manually add it to an ISO for testing [04:34] handsome_feng: https://launchpad.net/~ubuntu-cdimage/+livefs/ubuntu/bionic/ubuntukylin/+build/130378 [04:35] slangasek: Thanks [04:41] slangasek: dunno what's the criteria to unblock (never done that before) but the tests that had been run looked green so it's unblocked now [04:44] tjaalton: ok, and un-frozen as well now [04:59] great [05:11] -queuebot:#ubuntu-release- Builds: Xubuntu Desktop amd64 [Bionic Final] has been marked as ready [05:11] -queuebot:#ubuntu-release- Builds: Xubuntu Desktop i386 [Bionic Final] has been marked as ready [05:22] slangasek: 20180425 has the same problem [05:22] ok, at least that tells us which side the problem is on [05:23] handsome_feng: how do I reproduce this locally? with the 'try' or 'install' option? [05:23] slangasek: with try [05:25] This only occurs in live system, if we install the system directly, all goes fine. [05:26] handsome_feng: ok. I went to 'try' with 20180426 and I didn't see the problem in /run/user/999/dconf [05:26] this is with 20180426 amd64 [05:26] slangasek: OK , we will try that now [05:30] please accept gcc-snapshot [05:31] doko: I might leave it there for a little bit if that's ok, so I have something to trigger archive regeneration for the lubuntu cleanup I'm working on [05:33] ok [05:49] slangasek: we test the 20180426, and it also has the same problem.. [05:50] handsome_feng: ok. I am also testing 20180425.1 now to see if I can reproduce the problem here [06:00] infinity1, tsimonq2: alright, I've wrapped my head around this code now and have good output now - 62 packages having their support term lowered. Pushing. [06:01] handsome_feng: I cannot reproduce the problem on 20180425.1. Can you give me any more guidance on steps to reproduce this? [06:05] slangasek: ack, thabks. [06:05] *thanks [06:09] handsome_feng: did you fill in a bug for it already? [06:10] sil2100: good morning [06:12] slangasek: We've burn the 20180426 iso again. The bug disappears. [06:12] hmm [06:13] handsome_feng: do you think the previous burn didn't get done correctly? or do you think that we have a race condition in the boot? [06:14] Maybe the previous burn was wrong. We will test 20180425.1 again to confirm the problem. [06:15] handsome_feng: I would suggest also trying 20180426 several times to be sure the problem is really gone [06:15] slangasek: OK, We will do it [06:17] slangasek: morning o/ [06:17] phew [06:19] slangasek: any blockers reported in the meantime? [06:20] sil2100:Good morning! and we didn't fill a bug for it, we just found this problem when we test the iso this morning [06:22] sil2100: none that I know of [06:22] -queuebot:#ubuntu-release- Unapproved: python-pip (bionic-proposed/universe) [9.0.1-2 => 9.0.1-2.1] (no packageset) [06:23] -queuebot:#ubuntu-release- Unapproved: accepted python-pip [source] (bionic-proposed) [9.0.1-2.1] [06:36] infinity1: ubuntu-studio looks ok for both amd64 and i386 [06:37] -queuebot:#ubuntu-release- Unapproved: yapf (bionic-proposed/universe) [0.20.1-1 => 0.20.1-1ubuntu1] (no packageset) [06:38] -queuebot:#ubuntu-release- Unapproved: accepted yapf [source] (bionic-proposed) [0.20.1-1ubuntu1] [06:39] -queuebot:#ubuntu-release- Unapproved: pywinrm (bionic-proposed/universe) [0.0.3-1 => 0.3.0-1ubuntu1] (no packageset) [06:40] -queuebot:#ubuntu-release- Unapproved: accepted pywinrm [source] (bionic-proposed) [0.3.0-1ubuntu1] [06:43] mwhudson: I'll give subiquity a quick spin as well, so that you don't have to feel that your bias affected the test results ;) [06:44] slangasek: do you know if it's still ok to process package removals? [06:45] sil2100: sure [06:45] sil2100: I mean, if you remove the package and somebody disagrees, there will be sadness because there's no time for it to be readded [06:45] -queuebot:#ubuntu-release- Unapproved: pyroma (bionic-proposed/universe) [2.0.2-1 => 2.0.2-1ubuntu1] (no packageset) [06:46] -queuebot:#ubuntu-release- Unapproved: accepted pyroma [source] (bionic-proposed) [2.0.2-1ubuntu1] [06:48] -queuebot:#ubuntu-release- New sync: python-requests-ntlm (bionic-proposed/primary) [1.1.0-1] [06:50] xnox: the fallback casper message on exit is nice, but I must say I find it sometimes confusing [06:52] xnox: maybe 'confusing' being the wrong term, but since I see it everytime before plymouth kicks in, it's confusing as at one time the user is told to just reboot and then suddenly "nope, just press enter to do that" [06:57] -queuebot:#ubuntu-release- Unapproved: woo (bionic-proposed/universe) [1.0+dfsg1-2 => 1.0+dfsg1-2ubuntu1] (no packageset) [06:58] -queuebot:#ubuntu-release- Unapproved: accepted woo [source] (bionic-proposed) [1.0+dfsg1-2ubuntu1] [06:59] it seems no-one updated slideshow translations, can someone find out what needs to be fixed/documented where for the process? https://code.launchpad.net/~ubiquity-slideshow/ubiquity-slideshow-ubuntu/html [07:01] I've tested Finnish and French, the second (snap) and last slides are untranslated due to string changes during the cycle [07:02] (the middle string on the last slide is translated though) [07:03] -queuebot:#ubuntu-release- New: accepted python-requests-ntlm [sync] (bionic-proposed) [1.1.0-1] [07:03] it's documented here https://wiki.ubuntu.com/NonLanguagePackTranslationDeadline but maybe some process document needs to be updated as well [07:05] -queuebot:#ubuntu-release- Unapproved: capstone (bionic-proposed/universe) [3.0.4-3 => 3.0.4-5] (no packageset) (sync) [07:05] -queuebot:#ubuntu-release- Unapproved: accepted capstone [sync] (bionic-proposed) [3.0.4-5] [07:09] anybody got quick access to an ISO to boot? - clicking quit from ubiquity doesnt launch into the live session - appears to hang (standard bios boot) [07:11] -queuebot:#ubuntu-release- New binary: python-ntlm-auth [amd64] (bionic-proposed/none) [1.1.0-1] (no packageset) [07:12] -queuebot:#ubuntu-release- New: accepted python-ntlm-auth [amd64] (bionic-proposed) [1.1.0-1] [07:12] filed bug #1767048 [07:12] bug 1767048 in ubiquity-slideshow-ubuntu (Ubuntu) "Slideshow translations were not updated for 18.04 LTS" [Undecided,New] https://launchpad.net/bugs/1767048 [07:12] fossfreedom: xubuntu? or all? [07:12] fossfreedom, where do you press "quit" and which flavour? [07:15] fossfreedom: correct, clicking Quit from the keyboard selection menu does not seem to do anything [07:15] (Ubuntu main flavor for me) [07:15] it just hangs and becomes unresponsive [07:16] Mirv: hm, so we had an upload of the slideshow before release, but it didn't seem to include new translations [07:16] I guess that's what hid the issue [07:16] sil2100: right, it may have confused someone following the process that "oh, that was uploaded" [07:18] That might have been my responsibility, ouch [07:18] acheronuk, I'm on Ubuntu Budgie - thanks for the confirmation Mirv [07:19] sil2100: well it's not the first time, and can be fixed for .1, but it's good to get the process bullet proof for updating the nonlangpack translations for release [07:19] I dunno if KDE front end even has that option there. certainly never noticed [07:31] Ok, switching to the office for a short while [07:32] brb [07:36] are python3-networkx or python-networkx on any image? [07:40] Thoughts on this - three times this morning I've tried to boot into a live session on UB and I'm getting this https://imgur.com/a/NwCJRt6 [07:42] that's the same error handsome_feng mentioned earlier [07:43] * fossfreedom scrolling back [07:43] -queuebot:#ubuntu-release- Unapproved: python-networkx (bionic-proposed/main) [1.11-1ubuntu1 => 1.11-1ubuntu2] (ubuntu-server) [07:43] fossfreedom, https://paste.ubuntu.com/p/x9cwtjBGGh/ [07:44] yep -same issue [07:45] Laney, ^^ any thoughts? same error on 2 different flavours [07:45] fossfreedom, which build of which flavour ? [07:45] 64bit Ubuntu Budgie [07:45] Laney: s390x autopkg testers unhappy (unknown versions everywhere) [07:47] fossfreedom, what is the build number? [07:47] 201804?? [07:48] jibel, fossfreedom: which bug is this about? [07:51] sil2100, "unable to create file '/run/user/999/dconf/user': Permission denied. dconf will not work properly" when at-spi-dbus-bus is activated [07:51] I don't think any bug has been filed [07:51] I didn't see any in the backscroll [07:54] is it already possible to upload to -proposed? [07:57] Was it ever not? === infinity1 is now known as infinity [07:58] can't remember [07:59] jibel: what are the steps to reproduce that? [07:59] I guess I'm missing some backscroll [07:59] doko, we always upload to -proposed all cycle, via the redirect for 'bionic' to -proposed === LocutusO- is now known as LocutusOfBorg [08:00] apw: well, my question is, if I already can queue up -proposed uploads [08:01] doko, as long as things are prepped as SRUs i believe they are fine to upload [08:02] doko, though i would not think we want anything major in there in case we have to rebuild anything [08:02] ok [08:02] infinity, ^ that make sense ? [08:03] sil2100, I don't know. I didn't reproduce it myself but 2 testers on 2 different flavours reported it?. [08:03] -? [08:05] fossfreedom, can you report a bug please and attach the journal? [08:06] also describe the steps to reproduce the problem if you can do it reliably [08:14] if this the no free space thing again, or something else? [08:18] slangasek: I found this problem also occurs in Ubuntu, see https://paste.ubuntu.com/p/sZykyz5BVg/ [08:19] handsome_feng, how do you reproduce it? [08:19] can you reproduce it in a VM or only hardware? [08:19] jibel: Just burner the iso and then go into live system [08:20] jibel: I found that is only occurred in Dell computer. [08:20] And in the vm, all works fine [08:21] yeah but fossfreedom did it in vbox apparently [08:21] hmmmmm [08:21] jibel: emmm, It's wired.. [08:22] handsome_feng, can you report a bug in LP please? [08:23] it'll be easier to track the status than on IRC [08:23] jibel, OK, I will do it now [08:23] thanks [08:28] In latest ubuntu iso in virtualbox, I click 'try' and punts me to gdm login screen instead of straight into live session? is the intended? [08:29] no it means that the shell failed to start [08:30] you'll find more info in the journal [08:30] jibel: LP: #1767067 [08:30] Launchpad bug 1767067 in Ubuntu "[Live]Wrong Permission in live mode." [Undecided,New] https://launchpad.net/bugs/1767067 [08:30] and if I click live session user, the 1st time seems to crash and push me back to gdm. 2nd try logs me in to live user [08:31] And feel free to edit it [08:32] jibel: https://i.imgur.com/Hx7YJGI.png [08:33] acheronuk, yeah same bug than handsome_feng [08:35] sil2100: sorry - dropped out of IRC - on mobile at the moment - this bug is about trying to launch into a live session. When I tried earlier it was failing 50% of the time. Basically left at the login screen. Cannot launch into the session [08:37] It is intermittent. sometimes the live session launches - sometimes it doesn't. Nothing obvious as to why [08:37] okay, I reproduced it [08:37] Never seen this before until last nights respin [08:40] I saw it during the day yesterday [08:40] (but thought it was just me) [08:46] handsome_feng, when did you first see this problem? with which build? [08:46] willcooke, did you ever see it before? [08:47] jibel, Not sure if we're talking about the same thing. I saw the live image booting to a login screen instead of a logged in session during the day yesterday, and I /think/ Tuesday evening [08:47] jibel: We test 20180421, and it works fine [08:49] handsome_feng, did you get any imagine in between those days? like on monday 23 or 24? === hc_ is now known as Guest6903 [08:49] willcooke, that would be the symptom but without the log it cannot be confirmed. [08:49] * willcooke checks his trash [08:50] seb128: no...I just got the 20180425 [08:50] I found if I manually select "try but not install" , all goes fine! [08:51] handsome_feng, jibel, in what mode is the issue happening exactly? [08:51] jibel: were you able to reproduce it finally? Is this random? [08:51] live mode [08:52] handsome_feng, but you just said that if you select "try" it goes fine? [08:52] I can reproduce, If I manually select "try but not install" , all goes fine, and if I wait until the windows shows, the problem occurs [08:52] isn't try = live session? [08:53] seb128, just boot, then in ubiquity-dm select 'try ubuntu' [08:53] jibel, it happens only through that codepath? picking the live session from syslinux works? [08:53] seb128, yes [08:54] jibel, what flavor(s) are we talking about? [08:54] seb128, ubuntu, budgie and kylin [08:54] oh, ubuntu as well? [08:54] yes [08:54] wait until the windows show and then chose try [08:54] *fun* [08:54] jibel, do you know on what iso it started? [08:54] we had ubuntu isos in between monday and today [08:54] no, I've 24 here, I'll try it [08:54] thx [08:55] that sucks, looks like a release blocker issue? [08:56] question: can I sru virtualbox or sync it? I guess the former, right? [08:56] FFS I just tried again to video capture what it does, and it didn't happen [08:56] Laney, are you guys in London poking at it? [08:56] seb128, and you see the bug only on BIOS machine because on UEFI you don't start ubiqutiy-dm [08:58] 1. when start, hit the arrow keys, choose "try Ubuntu but not install" -> works fine. 2. Wait until the windows show and then choose the "try ubuntu" -> failed. [08:58] jibel, handsome_feng: I was testing kylin i386 on my VM and it worked correctly [08:58] jibel, handsome_feng: just now re-tested with clicking 'try ubuntu' and everything works [08:58] sil2100: Yeah, I also test it in vm, and works fine [09:00] I have the bug in a VM and fossfreedom too [09:00] urgh. randomly hard to reproduce! fun [09:00] jibel, I'm testing [09:00] jibel: are you able to reproduce it on a VM? Or you're on the real hardware? [09:00] sounds racy [09:00] sil2100, yes in qemu [09:00] other in vbox [09:00] Or maybe it's because I'm testing the i386 image [09:01] the platform is unrelated [09:01] the flavour too [09:02] I just tried on an amd64 iso, didn't get the issue [09:03] I wonder when this started [09:04] if I had to guess I would say this week when we started fixing ubiquity use of gsettings, that uncovered issue with code that never worked properly [09:04] seb128: a bit, but please do too [09:05] Laney, do you have any clue so far? I didn't manage to reproduce yet :/ [09:05] no didn't make it happen [09:05] I guess something in ubiquity-dm triggers a gsettings write with the worng user/env [09:05] but I've not idea what [09:11] -queuebot:#ubuntu-release- Unapproved: python-redis (bionic-proposed/main) [2.10.6-2 => 2.10.6-2ubuntu1] (ubuntu-server) [09:19] Just to update the channel.... I consider this a release blocker. Release team in London are trying to reproduce it right now [09:20] I've let JamieBennett know that this is a problem [09:20] willcooke, thx for the update, difficult to follow for those who are not in London [09:20] I'll relay messages [09:20] thx [09:21] * didrocks doesn't succeed in reproducing :/ [09:21] and looking at the code, I don't see anymore obvious guilty path [09:21] * seb128 neither [09:22] seb128, no failure with 24 and 50% of the boot are failing with 25.1 [09:23] yeah, that was the day we fixed the a11y… correct? [09:23] jibel, can you boot to -maybe, go to a tty and look at the permission of the /run ... file which is listed in the error? [09:23] willcooke, to reproduce just boot 25.1 in qemu and select "try ubuntu", retry until it fails [09:23] didrocks, that's the day we fixed the session reboot [09:23] jibel, are we talking about the "login fails, you get gdm" or about the dconf errors? [09:24] willcooke, I kept a VM running with the problem if they need more data in london [09:24] just reconfigured my test laptop in to legacy mode, L_aney is writing a USB stick [09:24] #livetweet [09:24] willcooke, you can relay the message ;) [09:24] seb128: hum, are you sure? we fixed the a11y issue on the 24, so 25.1 contains the a11y fix, not the one from 24 [09:25] jibel, shall we get you on a HO? [09:25] willcooke, great that you are here, at least we have some update of what is going on in London [09:25] otherwise it feels like complet disconnect [09:25] infinity, Laney: are you aware of https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-bionic/bionic/s390x/b/backblaze-b2/20180426_070912_e76c3@/log.gz ? [09:25] willcooke: a HO seems sensible [09:25] sure, give me a minute [09:25] which is a bit sad [09:25] doko: yes [09:25] seb128: :/ [09:25] we're trying to reproduce it [09:26] but please try as well [09:26] Laney, sorry, feels like we have pretty much being blocked or informed with hours delays of anything happening since tuesday, it's a bit the suck :/ [09:26] yeah, i'm trying [09:26] but I'm a bit confused now [09:26] what delays? [09:26] Laney: looks like a quoting error? [09:26] doko: no, the apt update failed [09:26] are we investigating the dconf permission errors? [09:26] ahh [09:26] yes [09:26] or the "get login screen" [09:26] what has the release sprint delayed in that though? [09:27] Laney, if you guys were not at the same place there would be IRC discussions about the issues being considered/worked on and others would be in the loop of what's going on/what are the issues/what is the status [09:28] this is a bug that was raised on IRC? [09:28] yes [09:28] 1h30 ago [09:28] that's the first we knew [09:28] seb128, I think the "get loing screen" is the same problem [09:28] anyway that's not an important discussion to have now [09:28] jibel, I don't understand, does it happen in live or not? [09:29] seb128, yes in a live session [09:29] 4213 [ 306.097343] ubuntu gnome-session-binary[4872]: dconf-CRITICAL: unable to create file '/run/user/999/dconf/user': Permission denied. dconf will not work properly. [09:29] jibel, do you have the buggy session atm? [09:29] seb128, you boot to a live session, and get a login screen [09:29] tes [09:29] yes [09:30] do you need to go through ubiquity-maybe? [09:30] or does it happen also in direct live from syslinux? [09:30] HO: https://hangouts.google.com/hangouts/_/canonical.com/will?authuser=0 [09:30] seb128, yes, I didn't reproduce it when booting straight to the live sesion from syslinux [09:31] k, good [09:31] jibel, can you "ps axu | grep dconf" on the buggy system? [09:31] jibel, link ^ [09:31] for HO [09:31] cc seb128, didrocks [09:31] yup 1s, the sound indicator disappeared [09:32] joining once my other vm trial booted [09:32] (don't want to artifically slowing it down and see what happens) [09:32] distro_info.DistroDataOutdated: Distribution data outdated. Please check for an update for distro-info-data. See /usr/share/doc/distro-info-data/README.Debian for details. [09:32] sigh [09:33] backprotpackage broken [09:34] willcooke, same as didrocks, my laptop is already struggling with vms boots etc atm, might join in a bit [09:35] Canonical CPC team checking in. I am CPC vanguard so if you know of any packages server/cloud related that will possibly cause a respin please ping. We're currently getting dailies uploaded to public clouds but will hold off on promotion until we get the OK from you guys. [09:36] jibel, did you see my question about the ps? [09:37] * didrocks does another testboot closing thunderbird this time… [09:38] I'm not getting the issue in virtualbox [09:38] I'm using kvm/qemu through gnome-boxes [09:38] as it's clearly a race depending on who writes first, hard to reproduce :/ [09:39] Laney, https://bugs.launchpad.net/ubuntukylin/+bug/1767067 [09:39] Ubuntu bug 1767067 in ubiquity (Ubuntu) "Booting to live session fails with: at-spi-bus-launcher: unable to create file '/run/user/999/dconf/user': Permission denied." [Critical,New] [09:39] sounds like we lost jibel, probably under pings from r-t and others [09:40] willcooke, jibel, I can't reproduce and can't help if IRC get ignored so I'm stepping out of that for a bit and going to look at launchpad recent reports & co, let me know if I can help/if someone wants to look at what I asked a bit earlier [09:43] :/ [09:43] we're just trying hard to debug it, it's not a deliberate ignoring of you [09:44] I was trying to help [09:44] seb128, sorry, I missed your messsage [09:45] seb128, join the HO [09:45] Laney, it's fine, it's just that it seems impossible from people not with you to get enough interaction bandwith to be useful, which is fair enough [09:46] seb128, gdm 5318 0.0 0.2 280736 5736 tty1 Sl 08:37 0:00 /usr/lib/ibus/ibus-dconf [09:46] ubuntu 5685 0.0 0.0 21536 1068 tty3 S+ 09:45 0:00 grep --color=auto dconf [09:46] Ubuntu bug 5685 in searchandrescue (Ubuntu) "searchandrescue: merge new debian version" [Medium,Fix released] https://launchpad.net/bugs/5685 [09:46] willcooke, I can't, my machine is already swapping from the vm and GNOME session eating 1.5G ram [09:46] jibel, you don't have dconf-service process(es)? [09:46] weird :/ [09:46] no [09:46] same here, it's either vm or HO :/ [09:46] seb128: how about joining the HO via phone? [09:47] update for IRC: xnox thinks he can see the race [09:47] JamieBennett, I guess I can try, I just don't have my canonical account on my phone [09:47] willcooke, ah, nice [09:47] atspi starting before dconf [09:47] and sometimes it doesnt [09:47] which basically matches the first report on atspi [09:48] I guess needing its env vars to see what's up [09:48] who is owner the dconf/user file? [09:49] it doesn't make sense to me [09:49] at-spi-bus-laun[4081]: unable to create file '/run/user/999/dconf/user': Permission denied. dconf will not work properly. [09:49] suggests that's it's not at-spi starting before dconf [09:49] seb128, it doesn't exist [09:49] we think it;s that the file doesn't exist, rather than perms [09:49] while permission denied then? [09:49] we had the same issue in ubiquity [09:49] no [09:50] it was printing "Permission denied" [09:50] ubiquity had the file owned by root [09:50] remember the second one [09:50] because a dconf-server was spawned by root [09:50] and 999 is owned by ubutnu [09:50] the one while you left for lunch :) [09:50] yeah, that was the first issue [09:50] the second was missing XDG_RUNTIME_DIR [09:50] with XDG env was uncorrect [09:50] right, be the result was a file root owned there [09:50] which was printing '/run/user/999/dconf/user': Permission denied [09:50] there was no file [09:52] HO closed for now [09:53] xnox has a plan [09:53] good [09:54] just in case the issue is the same the second fix we got the other day: check XDG_RUNTIME_DIR in that process and if setting it works [09:54] ftr [09:56] xnox, ^ [09:58] would be useful to have the env of at-spi if the process doesn't exit due to the error [09:58] but it's weird, the a11y stack has its own bus iirc, it's not using the standard session one [09:59] -queuebot:#ubuntu-release- Unapproved: val-and-rick (bionic-proposed/universe) [0.1a.dfsg1-5build1 => 0.1a.dfsg1-6~build1] (no packageset) [10:00] -queuebot:#ubuntu-release- Unapproved: accepted val-and-rick [source] (bionic-proposed) [0.1a.dfsg1-6~build1] [10:07] I made it happen [10:07] this is weird [10:08] anything interesting in env/file permission/confirming file doesn't exist? [10:10] https://paste.ubuntu.com/p/VYh4qqJFpf/ that's the install debug log [10:10] first bad line is (ubiquity:1430): IBUS-WARNING **: 09:47:25.837: The owner of /home/ubuntu/.config/ibus/bus is not root! [10:12] I was suggesting here trying to get an strace -f of both ubiquity-dm and the running dbus-daemon (the latter by pid) [10:12] and then assuming you hit the bug, search down for that directory name and then work backwards [10:13] big hammer but it might be quickest, assuming it doesn't perturb the bug out of existence === hc_ is now known as Guest16076 [10:20] Laney, did you find a way to trigger it or just go it after n tries? [10:21] -queuebot:#ubuntu-release- New binary: python-requests-ntlm [amd64] (bionic-proposed/universe) [1.1.0-1] (no packageset) [10:21] seb128: The latter. [10:21] k :/ [10:23] "intermittent race-condition heisenbug" =) [10:25] bug #1767048 :/ (not rc but annoying if that's the case) [10:25] bug 1767048 in ubiquity-slideshow-ubuntu (Ubuntu) "Slideshow translations were not updated for 18.04 LTS" [Undecided,New] https://launchpad.net/bugs/1767048 [10:25] -queuebot:#ubuntu-release- New: accepted python-requests-ntlm [amd64] (bionic-proposed) [1.1.0-1] [10:26] seb128: Yeah, if we find ourselves having to delay, that's a valid candidate. [10:35] seb128, are this strings in a langpack or with the package? when I look earlier today, the untranslated string in French is translated in LP since 2016 [10:36] in the package [10:36] that's true of the installer in general - it can't rely on langpacks [10:36] jibel: part of the nonlangpack translations: https://wiki.ubuntu.com/NonLanguagePackTranslationDeadline [10:37] right, that's what I supposed but everything is translated in French. [10:37] I think the template in LP is correct, but the translations just weren't exported and included in an upload [10:37] last translation update in the package seems to have been 2017-10-17 [10:38] seb128, jibel - everyone here as now reproduced the issue. Adding more cpus to a VM seems to "help". Trying to figure out what's going on now in parallel [10:38] jibel: the link changed a bit: https://bazaar.launchpad.net/~ubiquity-slideshow/ubiquity-slideshow-ubuntu/html/revision/810#slideshows/ubuntu/slides/gethelp.html [10:38] so that's why the two strings on the last page got affected [10:38] I'll see if I can get a translation update going [10:40] seb128, jibel - "help" = help reproduce [10:43] we found another gsettings call too https://bazaar.launchpad.net/~ubuntu-installer/ubiquity/trunk/view/head:/bin/ubiquity-dm#L526 [10:44] without all the sudo shizzle [10:47] -queuebot:#ubuntu-release- Unapproved: python-josepy (bionic-proposed/universe) [1.0.1-1 => 1.1.0-1] (no packageset) (sync) [10:48] cjwatson, thx [10:48] -queuebot:#ubuntu-release- Unapproved: accepted python-josepy [sync] (bionic-proposed) [1.1.0-1] [10:48]