| === john-cabaj1 is now known as john-cabaj | ||
| JackFrost | aha: FWIW, actually hit that on a Debian trixie system too. | 05:45 |
|---|---|---|
| === JanC is now known as Guest4670 | ||
| === JanC is now known as Guest1329 | ||
| adrien | I can see openmpi at https://ubuntu-archive-team.ubuntu.com/proposed-migration/update_excuses.html#openmpi but I can't find it in https://ubuntu-archive-team.ubuntu.com/proposed-migration/update_output.txt ; I'm not sure that's a big issue but I'm quite surprised and not sure how to completely understand that situation. Is there a known explanation for that? | 11:45 |
| ginggs | adrien: openmpi is no longer a candidate due to new autopkgtest regressions | 11:52 |
| ginggs | you can see it here https://ubuntu-archive-team.ubuntu.com/proposed-migration/update_output_notest.txt | 11:52 |
| seb128 | adrien, update_output.txt is the next step, it only includes things have no blocker on the first report | 11:52 |
| seb128 | +which | 11:52 |
| slyon | @pilot in | 11:54 |
| === ChanServ changed the topic of #ubuntu-devel to: Archive: Open | Devel of Ubuntu (not support) | Build failures: http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of Focal-Oracular | Patch Pilots: slyon | ||
| adrien | Thanks both of you; it makes perfect sense indeed. I had assumed that every package always appears in update_output.txt, not only packages without regressions (which have all been retriggered and are now passing btw) | 12:04 |
| adrien | ginggs: is the _notest variant the same but basically with the assumptions that there are no failing tests? | 12:04 |
| ginggs | adrien: exactly | 12:05 |
| adrien | thanks! :) | 12:06 |
| * sudip sees adrien has all the new openmpi autopkgtest regressions under control | 12:08 | |
| ahasenack | schopin: hi, did you see my updates in the unprivileged-ping bug? | 12:13 |
| ahasenack | schopin: I'm about to upload workarounds for backuppc and fence-agents | 12:13 |
| ahasenack | basically setting the group range to 0-65534, that works from inside the lxd, and should be sufficient | 12:13 |
| ahasenack | dep8 runs in the infra confirmed | 12:14 |
| ahasenack | https://pastebin.ubuntu.com/p/dr6nbTrfPN/ second line | 12:14 |
| adrien | sudip: tsimonq2 did a few retries too; and I'm merely monitoring stuff from a distance at the moment since I'm also busy with openssl 3.4 (and m2crypto, nodejs and ruby3.3 which have failing tests with openssl 3.4 :/ ) | 12:17 |
| adrien | but I see the transition tracker is almost completely happy about openmpi, nice! :) | 12:18 |
| ahasenack | something I can help with? I'm on +1 | 12:20 |
| ahasenack | picking my targets | 12:20 |
| sudip | adrien: yeah, if boost passes then it should be a valid candidate on next run | 12:27 |
| sudip | thanks doko for removing those binaries and the extra ones :) | 12:28 |
| sudip | ahasenack: if you are looking for something then I can suggest 'iptux'. i could not figure out why the build tests are failing but working in Debian | 12:39 |
| ahasenack | ok | 12:39 |
| ypwong | Hi! There are two packages that could use some sponsorship love: backport-iwlwifi-dkms in Oracular for unblocking oracular kernel build in next kernel sru, and usbio-drivers in Oracular which is needed for some mipi camera to work. Thanks! | 12:43 |
| sudip | thanks ahasenack | 12:50 |
| schopin | ahasenack: re iputils, I saw. I'm wondering if we shouldn't set that range in the default file to begin with, though. | 13:30 |
| ahasenack | unsure | 13:30 |
| ahasenack | whatever range we pick, it's bound to leave some people confused when their user is outside that range | 13:31 |
| ahasenack | I'm not so sure I like this change | 13:31 |
| ahasenack | on one hand, we are giving up on NET_RAW capability for ping, which is more than it needs | 13:31 |
| ahasenack | but on the other hand, we have to pick gids | 13:31 |
| ahasenack | oh, hmm, it's group ids | 13:31 |
| ahasenack | ah, but we have private groups | 13:31 |
| ahasenack | otherwise we could just pick the group "users" | 13:31 |
| schopin | Well, I can't see a context for which you have a gid > 65k (from your PoV) | 13:32 |
| schopin | AIUI above values are for userns mapping aren't they? | 13:32 |
| ahasenack | active directory users, for example | 13:32 |
| ahasenack | very high ids by default | 13:34 |
| schopin | yes but from the user perspective you get the low one don't you? | 13:35 |
| ahasenack | not sure what you mean | 13:35 |
| ahasenack | you get stuff like | 13:35 |
| ahasenack | $ id EXAMPLE\\jammy | 13:35 |
| ahasenack | uid=2001103(EXAMPLE\jammy) gid=2000513(EXAMPLE\domain users) groups=2000513(EXAMPLE\domain users),2001103(EXAMPLE\jammy) | 13:35 |
| schopin | Oh, nevermind, I misread the forum thread I was looking at. | 13:35 |
| schopin | Damned. | 13:35 |
| ahasenack | that range could allow something like 0 -1, to "wrap around" :) | 13:36 |
| schopin | Wouldn't that be nice. | 13:36 |
| ahasenack | or "all" | 13:37 |
| schopin | https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1089626 I was wondering yesterday why nobody complained in Debian... | 13:38 |
| -ubottu:#ubuntu-devel- Debian bug 1089626 in iputils-ping "iputils-ping: ping command does not work any more as non-root user" [Important, Open] | 13:38 | |
| ahasenack | I also don't know yet how this works (or doesn't work) by default inside docker containers | 13:41 |
| ahasenack | so in debian that iputils migrated? | 13:41 |
| schopin | Yes, presumably because they use full virtualization in autopkgtests | 13:41 |
| ahasenack | and they have the systemd change | 13:42 |
| ahasenack | also migrated, perhaps | 13:42 |
| ahasenack | but we have bugs on backuppc on the debian side, dep8 failing | 13:42 |
| ahasenack | and I think debian uses lxc, not vms. VMs is something more recent, not everywhere, I heard | 13:42 |
| schopin | backuppc was removed from testing for a short while. | 13:43 |
| schopin | well, short... 5 months | 13:43 |
| schopin | Ugh. I need some coffee, I can't see straight. It was removed from testing between Nov 17 and Nov 23 | 13:44 |
| schopin | I'm betting iputils migrated during that window. | 13:44 |
| schopin | https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1085289 even worse, it was removed *because of this issue* | 13:44 |
| -ubottu:#ubuntu-devel- Debian bug 1085289 in src:backuppc "Please set net.ipv4.ping_group_range sysctl in autopkgtests" [Serious, Open] | 13:44 | |
| ahasenack | so they had a warning, but removed the package that was warning them :) | 13:48 |
| ahasenack | was iputils code algo changed with the capability drop, i.e., it stopped doing some safety checks? | 13:49 |
| ahasenack | just wondering if we could/should revert that and again install it with the capability set | 13:49 |
| ahasenack | did security have a say in this change, like "oh yeah, drop the capability, much safer"? | 13:49 |
| schopin | Yeah that's my current line of thought as well. | 13:49 |
| schopin | No, we inherited this from Debian and only noticed because autopkgtests started failing. | 13:50 |
| schopin | I never contacted security on this. | 13:51 |
| ahasenack | let me ping | 13:51 |
| slyon | ypwong: looking. for backport-iwlwifi-dkms did you already get a written SRU exception? Considering an older full version backport was rejected? https://bugs.launchpad.net/ubuntu/+source/backport-iwlwifi-dkms/+bug/2046871/comments/14 | 13:51 |
| -ubottu:#ubuntu-devel- Launchpad bug 2046871 in backport-iwlwifi-dkms (Ubuntu Noble) "migrate to upstream release/coreNN branch" [High, Incomplete] | 13:51 | |
| slyon | ypwong: sorry, not sponsoring backport-iwlwifi-dmks yet: https://bugs.launchpad.net/ubuntu/+source/backport-iwlwifi-dkms/+bug/2085453/comments/8 | 14:21 |
| -ubottu:#ubuntu-devel- Launchpad bug 2085453 in backport-iwlwifi-dkms (Ubuntu Oracular) "New upstream core91 release" [High, In Progress] | 14:21 | |
| === JanC is now known as Guest3444 | ||
| slyon | cc vicamo ^ | 14:27 |
| === JanC is now known as Guest9590 | ||
| schopin | ahasenack: is it possible to have AD users in a container? If so, how does the mapping work? | 14:29 |
| ahasenack | schopin: it's possible, but usually you would have to increase the mapping range | 14:31 |
| ahasenack | container == lxd, I don't know about docker et al | 14:31 |
| ahasenack | for my dep8 tests around AD integration, I restrict those to VMs, for that reason | 14:32 |
| mfo | ubuntu-archive: hi o/ update_excuses for jammy apparently stopped updating at December 4th; others look good (10th/today), looking at https://ubuntu-archive-team.ubuntu.com/proposed-migration/log/<series>. Can someone take a look please? (I'd ask for pointers to try to help, but it seems this requires access to a specific system). Thanks in advance! | 14:33 |
| ahasenack | sudip: re iptux, it seems like an assertion error inside gtk, I'm really not familiar with that area and I think my time would be better spent in other issues where I can help more quickly | 14:34 |
| seb128 | mfo, hey, it's probably more of an #ubuntu-release topic | 14:34 |
| mfo | seb128, ack, thanks! | 14:34 |
| seb128 | I also personally don't know where that's set up | 14:34 |
| ahasenack | things to try would be drop lto, or frame pointers (less likely), and perhaps -Wl,-Bsymbolic-links | 14:34 |
| ahasenack | oh, and -O3, move back to -O2 | 14:34 |
| ahasenack | there have been issues recently with the combination of -O3 and LTO | 14:34 |
| ahasenack | actually, I would switch to -O2 first | 14:35 |
| adrien | I've investigated openmpi migrating based on update_output_notest.txt and I think there will be an issue around some packages (scotch|parmetis|metis, deal.ii, opm-*) but I haven't been able to investigate fully (plus, what is going on with scotch/parmetis/metis/whatever?!) | 14:49 |
| slyon | ypwong: sponsored usbio-drivers | 14:50 |
| sudip | ahasenack: ack, thanks. will try those later. | 14:52 |
| ahasenack | sudip: fwiw, building with -O0 also fails the build-time test | 14:54 |
| ahasenack | also SIGTRAP | 14:55 |
| === JanC is now known as Guest9046 | ||
| bluca | so currently in noble-updates linux is at version 6.8.0-49.49 but linux-meta is already at version 6.8.0-50.51 | 15:01 |
| bluca | linux 6.8.0-50.51 is in noble-proposed, not in noble-updates | 15:01 |
| bluca | so we get: linux-headers-generic : Depends: linux-headers-6.8.0-50-generic but it is not installable | 15:01 |
| bluca | is this a known issue? | 15:01 |
| bluca | ie, the linux-meta binary packages in noble-updates depend on the linux binary packages in noble-proposed | 15:02 |
| ahasenack | apw: ^? | 15:05 |
| === JanC is now known as Guest3297 | ||
| === dviererbe1 is now known as dviererbe | ||
| slyon | @pilot out | 15:57 |
| === ChanServ changed the topic of #ubuntu-devel to: Archive: Open | Devel of Ubuntu (not support) | Build failures: http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of Focal-Oracular | Patch Pilots: N/A | ||
| lvoytek | @pilot in | 16:00 |
| === ChanServ changed the topic of #ubuntu-devel to: Archive: Open | Devel of Ubuntu (not support) | Build failures: http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of Focal-Oracular | Patch Pilots: lvoytek | ||
| ypwong | slyon: thanks for sponsoring usbio-drivers, and will look at the sru exception tomorrow | 16:37 |
| sudip | thanks for checking ahasenack_ | 19:51 |
| lvoytek | @pilot out | 20:01 |
| === ChanServ changed the topic of #ubuntu-devel to: Archive: Open | Devel of Ubuntu (not support) | Build failures: http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of Focal-Oracular | Patch Pilots: N/A | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!