All meeting minutes can be found here:
Meeting is hosted here
DocB, ddemaio, lkocman, DimStar, Dirk, maxlin, Maurizio (m4u)
Heads up !!! Update on SLE Next on project@
TODAY (April 20th) at 15:00 is the code submission deadline for Leap
15.4 RC. Only bugfixes afterwards.
What about the backlog queue?
We have to process all that arrived until ~15:00, we can't easily have
different deadline than SLES 15 SP4 Public RC (also today).
Ask from QA: https://progress.opensuse.org/issues/109223
Update description of job groups with insufficient description
Leap 15.5 - we have to aim for very early planning announced
Focusing on Leap Micro to SLE Micro migration in spare time
* Having a discussion about existing migrated systems from CentOS 7 to
Leap 15.X and how does Leap 15.5 being the last release and transition
to ALP fits in. Users seem to be worried.
We have several possible ways: provide SP6/SP7 updates to Leap, some
nice offer to SLES migration (probably not an option for most), easy
migration to ALP ... let's see what's possible and what not.
* Traditional Leap user, and their journey after 2023.
These are two scenarios that I'd like to have active discussion on ^
## openSUSE Tumbleweed
WSL/Tumbleweeed - openQA somehow stopped working 9 months ago
https://openqa.opensuse.org/group_overview/68 do we know why? This has
been asked by Jose (Cloud/WSL QA)
openSUSE:Factory build fail stats: 177 failed, 11 unresolvable (last
* Easter-weekend: Tumbleweed was rolling, but with very few inputs (the
number of SRs was lower and could thus still be handled)
* Kernel 5.17.2
* Moved from LLVM13 to LLVM14
## Richard (Kubic/TW MicroOS)
Podman 4.0.x is in Factory
Pods in Podman 4.0.x are currently broken, due to catatonit 0.1.7 not
being in Factory. It's on the way, meanwhile rolling back is a valid
openQA tests for Pods in Podman is written, reviewed, just being held
back until catatonit 0.1.7 lands in Factory
Experimenting with MicroOS Desktop (GNOME) SelfInstalling images. If
experiments go well, next step will be to use TIU for OS updates.
Ultimately may obsolete the current YaST-based installation method for
lkocman: should I file inclusion requests for Leap/SLE Micro?
Richard: no plan to have it there as of now.
Working on various interesting MicroOS research tracks:
- kiwi-less/container first VM image building Status: POC
- nerdctl support in toolbox Status: Broken prototype
- buildkit and its limit as a container builder
- TW-to-MicroOS conversion
- curl-minimal and making it default (like Fedora) Status:
- Re-arranging Factory so products like MicroOS can build with
different PrjConf than Tumbleweed
- MicroOS Desktop using TIU instead of transactional-update (ie.
remove all individual rpm package management) (research not sure if
it's worth updating docs yet)
Please ping firstname.lastname@example.org if you are interested/have opinions on any
of the above, feedback/thoughts/existing knowledge is welcome.
* Deleted duplicated SLE package rebuilt in Backports project, there is
exceptions of multibuild package, because some specific build flavor
has enabled on openSUSE only
* Look into rpmlint issue, while debugging I found that it seem to we
did introduce entry to DBUSServices.WhiteList in the wrong format
therefore whitelisting doesn't work for kpmcore and others, bsc#1198521
* Tweaked openqa settings for disabling RNG device on upgrade tests for
* Build fail stats in Leap 15.4: 42 build fails, 0 unresolvables
lkocman: I'll open individual bugs for remaining build failures.
Max: looking into pending submissions on Staging
lkocman: I'll re-check status of the SLE submissions prior the RC
Santiago mentioned that he'll have more free cycles for Leap/openQA
this week. I'll reach out to him regarding focus on the migration
## Guillaume - Arm
* Factory:ARM recovered from GCC12 build failure. A new snapshot
should go to openQA later today or tomorrow.
* Upcoming GCC 12 tested for aarch64:
* Go compiler has been updated in SLE.
* Enlightenment package still fails in rpmlint check (missing
lkocman: will reach out to simotek and Wolfgang
* scc product class looks wrong (Marina/Marcus) (for all non-intel
arches). x86_64 has OPENSUSE-BETA while for the others the class is
SLES-ARM64-BETA, SLES-PPC-BETA, SLES-Z-BETA.
Ticket with SCC team, team lead is aware of the request). Request is
now resolved. (lkcoman: Correction trello shows me that 2/4 items are
lkocman to check if we could remove the firefox tests from the
42.X/15.0 upgrade test suite. It doesn't seem to happen outside of that
(same for TW).
lkocman will try to look into this
## Sarah - s390x
* Is rolling
openSUSE Leap Beta:
* no issues at the moment
* fakeroot updated, that it is buildable for s390x again
* arpack-ng requires 12GB of memory for successful builds (constraints
lkocman: slightly related topic, should we perhaps re-consider
recommended memory amount in Release Notes? I think we still recommend
something like 2G minimal / 4G or 8G optimal.
* KDE for s390x adopted
* Temporary suspension of TSP until we can find a payment solution
* Had meeting with accounts payable (need other experts on call)
* Another meeting schedule for beginning of next week
* 108 registrations, 63 submissions, 39 hours
* Approving talks
* Have mock schedule and feedback from people on talk times
* Catering offer recieved (hope to approve it this week)
* Meeting today at Z-Bau
lkocman: I'd really like to see a panel discussion with our ALP
steering comittee on the conference (similar to Fedora in the past).
* OSCAL oSC22
* CfP goes until April 26
* LAS talk next week about Leap Micro
* GSoC is May 12 deadline
Git Hub maintainer day/week? mentioned on the #oscc
* switched obs-arm-11 worker to aio=io_uring away from aio=threads
which appears to reduce I/O cost (bsc#1197746)
* fixed file(1) speedup regression
* Still to work rust 1.59 / ghc issues on armv6 and armv7: all have the
same root cause, use of SWP/SWPB instructions which is imperfectly
emulated by the aarch64 host kernel. works fine with armv6/v7 kernel.
In backlog: https://code.opensuse.org/leap/features/issue/17
## Wolfgang (Package Hub), Scott Bahling
## Maintenance team (Marina or Marcus, Maurizio (m4u))
- Marcus: 15.4 is set up channel wise, updates will go in testing
occasionaly, if you need more ping me, no openqa setup yet
lkocman: could you please help me with cleanup of maintenance tasks in
- Marcus: 15.3 openQA is ok again.
lkocman: (still pending) will open a release notes entry for a change
Security Advisory identification
TODO Marcus: could you give me example of old and new values?
- Marina and Maurizio: SLE-PackageHub overlaps
in progress following the initial list of overlapping packages
(https://progress.opensuse.org/issues/106610#note-3). More incidents
are under QA. Work is going on and under control (and tracked
internally as jsc#MSC-303).
Already released: (AH PERFECT!)
SUSE:Maintenance:22929:265929 (babl libbabl-0_1-0 typelib-1_0-Babl-0_1)
SUSE:Maintenance:23000:266006 (MozillaThunderbird enigmail)
SUSE:Maintenance:23009:266004 (freerdp freerdp-devel freerdp-proxy
libfreerdp2 libgsm) -> libgsm1 was missing in PH making freerdp not
SUSE:Maintenance:22928:266007 (argyllcms csync libcsync0 libcsync-
Max alrady removed packages from Backports
Maurizio: Most packages from the list for 15-SP3 on the ticket are now
staged, waiting to be sent for QA testing. Update on status to come as
soon as package updates are released but it will likely take time.
Leap Micro 5.2 - Maintenance setup should be rather small, as we do
override only branding.
Lubos to open a progres-o-o ticket for the maintenance update. Marcus
would also prefer as small overhead as possible. Adrian: we could have
a policy that if there is an updated for Micro it should go only to the
official code-stream even if the package didn't exist there yet.
Raised by Marcus: removal of SUSEConnect in favor of suseconnect-ng
lkocman: the obsoletes are set correctly, none of packages is installed
by default. And installation of suseconnect-ng currently removes
SUSEConnect. Documentation was updated accordingly. In my opinion this
is already good enough.
## Adrian - OBS
feedback loop on mirrored submissions doesn't workstation. Lubos wills
end an email with an example to Marco and Adrian. (Sent last week)