[15:59] <slyon> o/
[15:59] <bdmurray> o/
[16:00] <jawn-smith> o/
[16:00] <bdmurray> #startmeeting Weekly Ubuntu Foundations team
[16:00] <meetingology> Meeting started at 16:00:11 UTC.  The chair is bdmurray.  Information about MeetBot at https://wiki.ubuntu.com/meetingology
[16:00] <meetingology> Available commands: action, commands, idea, info, link, nick
[16:00] <bdmurray> #topic Lightning Round
[16:00]  * bdmurray shuffles names like cards
[16:00] <bdmurray> ogayot xypron jawn-smith vorlon schopin waveform alexghiti sil2100 enr0n juliank mwhudson ginggs doko dbungert slyon bdmurray
[16:00] <ogayot> * subiquity:
[16:00] <xypron> Still in call
[16:01] <ogayot>   * consolidated API for UA contract token
[16:01] <ogayot>   * still troubleshooting issues with overlays for third-party drivers
[16:01] <ogayot> * whoopsie / apport:
[16:01] <ogayot>   * manual tests run in no-UI & UI mode
[16:01] <ogayot>   * regression when running whoopsie in path-based activation mode only
[16:01] <ogayot>   * working on another prototype with timer-based + path-based activation
[16:01] <ogayot> * merge of python-testtools
[16:01] <ogayot> * fix of mosquitto autopkgtest
[16:01] <ogayot> (done)
[16:01] <bdmurray> since xypron is still in a call lets move to the next person
[16:02] <jawn-smith> * ubuntu-image:
[16:02] <jawn-smith>     * Created https://github.com/snapcore/snapd/pull/11355
[16:02] <jawn-smith>     * Work done locally with full coverage for the ubuntu-image side of this
[16:02] <jawn-smith>     * Testing these changes on lots of architectures
[16:02] <jawn-smith> * icu transition:
[16:02] <jawn-smith>     * aegisub: has been removed from the archive
[16:02] <jawn-smith>     * gspell: unrelated to icu. filed LP: #
[16:02] <jawn-smith>     * icu-ext: fix-uploaded
[16:02] <jawn-smith>     * libcolumbus: is unrelated to icu 70, but I uploaded a fix anyway. RISC-V still FTBFS
[16:02] <jawn-smith>     * libsimpleini: just needed a rebuild
[16:02] <jawn-smith> * golang 1.18 transition:
[16:02] <jawn-smith>     * golang 1.18 does not play nicely with lto. New version of golang 1.18 uploaded to my ppa
[16:02] <jawn-smith>     * Once it has been published I will start no change rebuilds in my ppa
[16:02] <jawn-smith> * Assisting security team with golang-modules
[16:02] <jawn-smith> * LP: #1960083 include dirname in busybox to allow encrypted rootfs to boot again
[16:02] <jawn-smith> * Proposed migration: ansible vs python3.10
[16:02] <jawn-smith>     * submitted bug against dh-python for not cleaning broken symlinks
[16:02] <jawn-smith>     * Thanks ginggs for helping identify that the behavior should be fixed in ansible instead and moving the bug
[16:02] <jawn-smith> [done]
[16:02] <xypron> Package cd-boot-images-riscv64 in preparation of installer images LP: #1960216
[16:02] <xypron> Analyzed autopkgtest problems of package keras. Opted for deletion LP: #1959918
[16:02] <xypron> Fix autopkgtests of package aiocoap: LP: #1959315
[16:02] <xypron> Start packaging jh17xx-tools for recovering VisionFive board
[16:02] <xypron> <done />
[16:03] <bdmurray> vorlon:
[16:04] <vorlon>  * processing Debian package removals, which led to
[16:04] <vorlon>   * reviewing our mlocate delta and porting it to plocate so that we can do the m/p-locate transition Debian has already done
[16:04] <vorlon>   * needs an MIR for this new codebase, if someone is in need of MIR experience
[16:04] <vorlon>  * spending some time on TB issues
[16:04] <vorlon>  * working on NBS cleanups; removed php8.0 from release finally
[16:04] <vorlon>  * doing interviews this week
[16:04] <vorlon> (done)
[16:04] <bdmurray> schopin:
[16:05] <schopin> * rustc: worked on rustc 1.58.1 for the MIR, with multitude of smaller items in there
[16:05] <schopin> * distro: worked on several packages for the Python 3.10 migration
[16:05] <schopin> (done)
[16:06] <bdmurray> waveform:
[16:06] <waveform> * Investigated pi-bluetooth issue on UC18, UC20, and UC22 on CM4, 4B
[16:06] <waveform> * .. unable to replicate on 4B
[16:06] <waveform> * .. replicated on CM4 under UC20 and UC18; fixed on UC20, left UC18 as CM4 isn't supported there
[16:06] <waveform> * Finished tentative fix for debhelper restart issue (LP: #1959054)
[16:06] <waveform> * Incorporated fix for ^^ into debhelper merge (LP: #1960248)
[16:06] <waveform> * .. fix needs sponsorship, but also serious testing for any affected packages
[16:06] <waveform> * .. would also be good to grep the archive for any packages using --no-restart-{after,on}-upgrade in d/rules (how?)
[16:06] <waveform> * Merged extra interface changes to pi-gadget
[16:06] <waveform> * Cherry-picked Pi Zero 2, and interface changes across UC20 & 22 branches (for both supported archs)
[16:06] <waveform> * Merged a few outstanding PRs on pi-gadget (including better build docs \o/)
[16:06] <waveform> * Worked on updated "classic" and "desktop" branches of pi-gadget (sorry, still not *quite* done!)
[16:06] <waveform> * Verification of linux-firmware-raspi{,2} SRUs
[16:06] <waveform> * Analyzed memory usage of "bare+idle" server system (non-essential RSS hogs in desc. order: snapd=21-32MB, multipathd=16-17MB, unattended-upgrades-shutdown=12-16MB, fwupd=15MB, networkd-dispatcher=12-15MB)
[16:06] <waveform> * Pi meetings
[16:06] <waveform> (done)
[16:06] <alexghiti> PlusOne maintenance
[16:06] <alexghiti> Pytest fixed uploaded by ginggs
[16:06] <alexghiti> Pytest-mpi fixed by tox upload by ginggs
[16:06] <alexghiti> Pytest-skip-markers fixed by pyfakefs update (https://salsa.debian.org/python-team/packages/python-pyfakefs/-/merge_requests/1)
[16:06] <alexghiti> Python-ml-collection fixed uploaded by ginggs
[16:06] <alexghiti> Q2cli: upstream fix not enough  https://github.com/qiime2/q2cli/pull/260
[16:06] <alexghiti> Shasta 0.8.0-1 always failed because of lack on memory in VM
[16:06] <alexghiti> Typer fixed by pytest-sugar https://github.com/Teemu/pytest-sugar/pull/231
[16:06] <alexghiti> Pytest-sugar to upload
[16:06] <alexghiti> Unicycler fixed by spades update
[16:06] <alexghiti> Spades update ongoing
[16:06] <alexghiti> Toil may not support python 3.10 (https://github.com/DataBiosphere/toil/pull/3967)
[16:06] <alexghiti> Python-pip  https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1004149
[16:06] <alexghiti> Libapache2-mod-python ongoing
[16:06] <alexghiti> Python-dugong ongoing
[16:06] <alexghiti> (done)
[16:07] <bdmurray> I'm not sure sil2100 is going to make it
[16:07] <bdmurray> enr0n:
[16:07] <enr0n> * MR merged for LP #1947581.
[16:07] <enr0n> * Working on MR for LP #1959458.
[16:07] <enr0n> * nanolyse vs python3-defaults (with bdmurray)
[16:07] <enr0n>   - Was failing due to minimap2 issue (LP #1960028)
[16:07] <enr0n>   - minimap2 in -proposed removed
[16:07] <enr0n>   - nanolyse autopkgtest succeeds with python3.10 and previous minimap2 version
[16:07] <enr0n> * Administrative stuff
[16:07] <enr0n> (done)
[16:08] <bdmurray> juliank:
[16:08] <juliank> * grub2 upstream work
[16:08] <juliank> * reviewed grub2 riscv patchset
[16:08] <juliank> * proposed-migration djangorestframework
[16:08] <juliank> * reviewed shim submissions from:
[16:08] <juliank>   - Isoo (https://github.com/rhboot/shim-review/issues/202)
[16:08] <juliank>   - OS ALT (accepted, https://github.com/rhboot/shim-review/issues/205)
[16:08] <juliank>   - Microsoft CBL-Mariner (accepted with caveats, https://github.com/rhboot/shim-review/issues/223)
[16:08] <juliank> * apt merges review
[16:08] <juliank> * lots of silly autopkgtest-cloud work to make amd64/i386 become more stable
[16:08] <juliank>   - Retry stderr printing if we received EAGAIN :(
[16:08] <juliank>   - Move /tmp/autopkgtest-run-wrapper to /var/tmp from /tmp, as the latter is deleted by systemd-tmpfiles which might run after we create the file
[16:08] <juliank>   - use nova reboot to reboot in addition to reboot in the VM, to workaround broken network after reboot
[16:08] <juliank>   - poked IS to restart OVS in lgw01
[16:08] <juliank>   - set ConnectionAttempts=20 in auxverb to work around flaky networking
[16:08] <juliank>   - limit our lcy01 workers to 20 (from 28) as IS instituted a limit there
[16:08] <juliank>   - limit our code to only boot one new VM every 30 seconds (per cloud)
[16:08] <juliank> (done)
[16:09] <bdmurray> ginggs:
[16:09] <ginggs> * Working on FTBFS and autopkgtest regressions in Python 3.10-as-default transition
[16:09] <ginggs> * FTBFS reports
[16:09] <ginggs> * Sponsored some uploads; tox, pytest and python-ml-collections
[16:09] <ginggs> (done)
[16:09] <doko> - started the icu transistion
[16:09] <doko> - going on with the perl transition
[16:09] <doko> - currently handicapped by failing builds without logs. LP team is aware of this
[16:09] <doko> - package removals related to icu and perl
[16:09] <doko> (done)
[16:09] <dbungert> * +1: bobcat vs icmake - send partial fix upstream (debbug 1004986), upstream
[16:09] <dbungert> can reproduce core problem (endianness issue)
[16:09] <dbungert> * python3.10 vs oz: collections -> collections.abc, LP: #1960273
[16:09] <dbungert> * subiquity - update storage/v2 to require a gap object for partition creation,
[16:09] <dbungert> will be used later to control where on disk partitions are created
[16:09] <dbungert> * subiquity - pick up partial branches from mwhudson on partition editing
[16:09] <dbungert> * subiquity - fix issue where sensitive information could be logged to log
[16:09] <dbungert> files, and tighten permissions on files that subiquity writes
[16:09] <dbungert> (done)
[16:09] <slyon> - take-home test review
[16:09] <slyon> - MIR: improve protobuf-c packaging, to make it ready for main, LP: #1956617
[16:09] <slyon> - merged sudo (still having issues on autopkgtest.u.c, working locally)
[16:09] <slyon> - systemd SRU for OEM team (Impish + prep-work for Focal), LP: #1955997
[16:09] <slyon> [16:09] <slyon> - more release planning & initial testing
[16:09] <slyon> - merged SmartNIC work https://github.com/canonical/netplan/pull/253 \o/
[16:09] <slyon> - reviewed & merged https://github.com/canonical/netplan/pull/259
[16:09] <slyon> - some discussions around OVS+GRE tunnel offloading
[16:09] <slyon> - experiemnts around netplan vs NetworkManager vs udev vs networkd allow-
[16:09] <slyon>   and denylist integration, LP: #1951653
[16:09] <slyon> (done)
[16:10] <bdmurray> review of and merging of the fix for LP: #1947581
[16:10] <bdmurray> searched for duplicates of the above bug too
[16:10] <bdmurray> reported bug regarding /boot partition sizing LP: #1959971
[16:10] <bdmurray> uploaded partman-auto for Focal with changes for the above
[16:10] <bdmurray> modified ubuntu-release-upgrader fallback initrd size estimate
[16:10] <bdmurray> merged, uploaded software-properties MP regarding ppa.launchpadcontent.net
[16:10] <bdmurray> pinged IS regarding firewall rule change for the above
[16:10] <bdmurray> updated mirrors, demotions etc for ubuntu-release-upgrader for Jammy
[16:10] <bdmurray> investigation into germinate with -updates (LP: #1921862)
[16:10] <bdmurray> submitted removal request for lupin (LP: #1960457)
[16:10] <bdmurray> no change rebuild of minimap2 for python3.10
[16:10] <bdmurray> reran nanolyse tests for the above
[16:10] <bdmurray> uploaded python-apt, ubuntu-release-upgrader for 20.04.4
[16:10] <bdmurray> met with ogayot regarding whoopsie and apport-noui
[16:10] <bdmurray> sru-reviews and releases
[16:10] <bdmurray> preparing for and conducting interviews
[16:10] <bdmurray> mentoring
[16:10] <bdmurray> ✔ done
[16:10] <bdmurray> any questions on status?
[16:12] <bdmurray> #topic Team proposed-migration report
[16:12] <bdmurray> #link http://reqorts.qa.ubuntu.com/reports/rls-mgr/rls-jj-incoming-bug-tasks.html#foundations-bugs
[16:12] <vorlon> wrong #topic?
[16:13] <bdmurray> yes
[16:13] <bdmurray> #topic Release Incoming Bugs
[16:13] <bdmurray> bug 1960263
[16:14] <bdmurray> Is anybody looking at this with the the python3.10 transition?
[16:14] <bdmurray> It doesn't sound like it so lets just card it
[16:15] <ginggs> I was not aware of it, it did build recently on  2022-01-10
[16:15] <bdmurray> bug 1958720 is tagged server-todo so let's see if they get it done
[16:16] <bdmurray> bug 1960089
[16:16] <doko> there is no python-is-python2 in jammy anymore
[16:16] <bdmurray> doko: Can you comment on the bug?
[16:17] <bdmurray> As I understand it some users are installing multiple flavors of kernels (e.g. oem and stock) and /boot partition size calculation only provides room for one flavor of kernel.
[16:17] <bdmurray> And when someone's /boot partition becomes full it is a rather difficult situation to recover from.
[16:18] <bdmurray> vorlon: Do you have an opinion here?
[16:18] <vorlon> I could also write chunks of /dev/random to /boot and run it out of space but I don't think that's something we should be taking into consideration when sizing /boot at install time
[16:19] <vorlon> I do not think a user having two kernel metapackages installed in parallel is a supported use case
[16:19] <bdmurray> How will users know that though? It seems like an easy mistake to make.
[16:19] <schopin> Perhaps we could warn the user when installing a new kernel metapackage?
[16:20] <vorlon> how are these users coming up with the idea to manually install a different kernel metapackage than the one selected for them by the installer?
[16:20] <vorlon> you don't get there from the gui
[16:20] <vorlon> is someone giving bad advice on the Internet telling them to run apt?
[16:21] <dbungert> the user on that bug is making the case that they have users intentionally using 2 or 3 kernels
[16:21] <vorlon> schopin: the UX for being able to pop such a warning makes that difficult
[16:21] <juliank> you get two kernel flavors if you upgrade from oem, fwiw, but I'm not sure how this ends up in terms of kernel count
[16:21] <vorlon> dbungert: well, I think it's clearly intentional, I just don't think it's sensible
[16:21] <bdmurray> Ubuntu Studio users seem to switch between the lowlatency and generic kernel
[16:22] <juliank> Their calculation is off in any case
[16:22] <vorlon> bdmurray: anyway my view is that there needs to be a broader design discussion here, perhaps on ubuntu-devel, not buried in a bug report
[16:22] <juliank> as we keep 3 kernels, and our calculation is to allocate space for 2*3 kernels (and initrds)
[16:22] <dbungert> vorlon: I'm inclined to agree but with my installer hat on it's hard to know how to respond to this
[16:22] <bdmurray> Yeah, I'd use our calculation but just adjust it for 2 kernel flavors if appropriate
[16:22] <juliank> historically
[16:23] <vorlon> juliank: AIUI you end up with two kernel metapackages but one is supposed to be a dummy pointing to the other, and any resulting increase in installed kernels is a bug
[16:23] <juliank> it should be
[16:23] <schopin> can flavours hook into the installer at this stage?
[16:23] <bdmurray> Okay, I'll suggest they bring it up on the ubuntu-devel mailing list.
[16:24] <juliank> Please point to the old bug with the calculation we did
[16:24] <juliank> :)
[16:24] <juliank> Well cyphermox did I think
[16:24] <vorlon> the most recent calculations in u-r-u are mine working with bdmurray, AFAIK
[16:24] <bdmurray> I did in our irc chat about it
[16:24] <juliank> also we do have a card to reevaluate size of /boot already, no?
[16:25] <bdmurray> juliank: yes, its assigned to me and I upload partman-auto for focal. I'm waiting on impression for compression decsisions
[16:25] <juliank> ok
[16:25] <bdmurray> s/impression/jammy/
[16:25] <bdmurray> okay so bug 1959362
[16:26] <bdmurray> Let's ask seb if they are working on it
[16:26] <bdmurray> bug 1960255
[16:26] <vorlon> - Setup a password for Secure Boot MOK enrollment
[16:26] <vorlon> sorry, that's on the ubiquity bug
[16:26] <juliank> u-r-u that's a no-brainer
[16:26] <vorlon> does that mean ubiquity failed to get the signed nvidia drivers for them?
[16:27] <bdmurray> Yes, there has been similar work done with u-r-u so let's target and card that.
[16:27] <vorlon> (but yes, desktop can have it for now)
[16:27] <juliank> the nvidia issue happens occassionally I think
[16:27] <vorlon> :|
[16:27] <juliank> I've seen other bugs about that in the past weeks, I think they're working on those
[16:28] <bdmurray> bug 1959759
[16:29] <bdmurray> that has ubuntu-archive subscribed to it, is that enough?
[16:29] <vorlon> that's in the ubuntu-archive team queue, I don't think it belongs in the rls-jj-incoming escalation path
[16:29] <bdmurray> okay
[16:31] <bdmurray> #link https://reqorts.qa.ubuntu.com/reports/rls-mgr/rls-ii-incoming-bug-tasks.html
[16:31] <bdmurray> bug 1959085
[16:33] <bdmurray> let's add a zfsutils task and let the kernel team look at it
[16:34] <vorlon> added
[16:34] <bdmurray> the rls-ff-incoming list just has the /boot size bug
[16:34] <bdmurray> so that's it from bugs
[16:34] <bdmurray> #topic Team proposed-migration report
[16:34] <vorlon> #link https://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses_by_team.html#foundations-bugs
[16:34] <vorlon> 126 packages needing attention, 14 packages not yet considered late
[16:34] <vorlon> looks quite dire
[16:35] <vorlon> the first couple are ruby-defaults, --> server team
[16:35] <vorlon> python-click is the mailman3 issue, which is with mwhudson and unblocked now (we will remove mailman3 as Debian has)
[16:35] <bdmurray> Is there / does there need to be a bug opened for that?
[16:35] <vorlon> python3-defaults is looking much better, with only a short list of blockers; are these all assigned out or otherwise in progress?
[16:35] <vorlon> bdmurray: we will need a removal bug with ubuntu-archive subscribed
[16:36] <bdmurray> No, we did not assign everything out
[16:36] <vorlon> so I have barbican assigned to me and I think I would like to pass it on to someone else
[16:36] <doko> it's now clarified that we want to remove mailman3
[16:36] <vorlon> LP: #1959626
[16:36] <slyon> I can take that
[16:36] <alexghiti> Regarding python3-default: I am taking a look at all the packages after pytest
[16:36] <bdmurray> the mosquitto ones will just need retrying with python3-defaults
[16:36] <vorlon> slyon: thanks
[16:37] <vorlon> when apache2 is restarted with libpython3.10 loaded, the wsgi requests hang with no useful logs
[16:37] <vorlon> so someone more conversant with wsgi should look at it
[16:37] <vorlon> I'll have a look at faiss since it's ppc64el
[16:38] <vorlon> alexghiti: are you still working on libapache2-mod-python?  looks like you were assigned to it last week
[16:38] <alexghiti> vorlon: Yes, apache2 segfault and I struggle to understand why
[16:38] <alexghiti> I want to update the package to see if it helps
[16:38] <vorlon> ok
[16:39] <vorlon> then it sounds like the remaining blockers are handled
[16:39] <vorlon> then we have a series of packages waiting for python3 and/or ruby...
[16:39] <ginggs> LP: #1960088 is tagged server-todo
[16:40] <alexghiti> Ahah thanks, same segfault as mine
[16:41] <vorlon> media-types has a python3.9 autopkgtest regression on ppc64el, I'll look at this it's probably a false positive (and ultimately ignorable because python3.9)
[16:41] <vorlon> opensbi, I heard some back chatter about this - alexghiti are you working on this?
[16:41] <alexghiti> Yes, I'll do that next week
[16:41] <vorlon> thanks
[16:42] <vorlon> linux-firmware-raspi, is this blocked by bluez-firmware or something else?
[16:44] <vorlon> "unsatisfiable dependency" so I assume it's related and already handled, moving on
[16:44] <vorlon> and then we start getting into autopkgtest waits (libwww-perl)
[16:44] <vorlon> initramfs-tools vs klibc warrants someone looking at it
[16:45] <vorlon> ogayot: could you look at initramfs-tools/klibc?
[16:45] <ogayot> sure
[16:45] <vorlon> xypron: could you look at pillow?
[16:46] <xypron> ok
[16:46] <vorlon> glibc, I'm not sure the status of these, I'll skip over them for now
[16:47] <vorlon> oh, python3.9/ppc64el also blocking glibc, that's fun; I'll certainly look at that one
[16:47] <vorlon> systemd autopkgtest blocking linux-meta-aws, jawn-smith can you look?
[16:47] <jawn-smith> sure
[16:48] <vorlon> cross-toolchain-base: schopin?
[16:48] <schopin> ack
[16:48] <vorlon> cdebconf: waveform?
[16:48] <waveform> vorlon, ack
[16:48] <vorlon> busybox vs. casper: sil2100 ?
[16:49] <dbungert> casper issue is a known problem, I suggest retest
[16:49] <vorlon> gdbm: enr0n?
[16:49] <sil2100> vorlon: ACK o/
[16:49] <enr0n> can do
[16:49] <sil2100> ...or just a re-test then
[16:49] <vorlon> and then there are piles of perl stuff I'm not going to dig into
[16:49] <vorlon> sorry I didn't find enough to assign one to everyone
[16:50] <vorlon> but if you didn't get one feel free to dive in and grab something else to look at :)
[16:50] <vorlon> bdmurray:
[16:50] <bdmurray> #topic AOB
[16:50] <bdmurray> I believe Monday is a US holiday
[16:50] <slyon> I'll be off tomorrow
[16:50] <mclemenceau> isn't it next Monday?
[16:51] <doko> fix  zeek, zzuf ftbfs please
[16:51] <jawn-smith> It's the day after the super bowl, which should be a holiday :)
[16:51] <bdmurray> Oh, I'm off by one
[16:51] <bdmurray> Okay, if there is nothing else.
[16:51] <bdmurray> Thanks everybody.
[16:52] <bdmurray> #endmeeting
[16:52] <meetingology> Meeting ended at 16:52:00 UTC.  Minutes at https://ubottu.com/meetingology/logs/ubuntu-meeting/2022/ubuntu-meeting.2022-02-10-16.00.moin.txt
[16:52] <schopin> o/
[16:52] <slyon> o/
[17:31] <Eickmeyer> bdmurray: For clarification, it's not Ubuntu Studio. It's Kubuntu Focus (OEM laptop) users.
[17:32] <Eickmeyer> Hence the OEM kernel being a factor.
[17:34] <Eickmeyer> bdmurray: Additionally, it's mostly due to a bug in PackageKit (which KDE Discover uses as a backend) which causes packages that were originally marked as auto to be marked as manual when upgrading.
[17:34] <Eickmeyer> Doesn't have much to do with having more than one kernel flavor installed, though that can be a factor.
[17:35] <Eickmeyer> The problem is there's no safety net when the /boot directory overfills.
[17:36] <Eickmeyer> *sigh* this should be in #ubuntu-devel.
[17:53] <bdmurray> Eickmeyer: Having the discussion on the ubuntu-devel mailing list would be best as this seems like an involved conversation
[18:26] <vorlon> making /boot bigger does not solve the problem of there not being a safety net when the /boot directory overfills.
[18:26] <vorlon> and yes, ubuntu-devel please
[20:00] <mwhudson> bdmurray, vorlon: tada https://bugs.launchpad.net/ubuntu/+source/mailman-hyperkitty/+bug/1960547
[20:00] <bdmurray> mwhudson: thanks