Foundations Team Updates - Thursday 19 January 2023

Previous status can be found here: Foundations Team Updates - Thursday 12 January 2023

1 Like

openjdk update:

jtreg7 update:

  • need to update testng and jcommander packages. There is an issue currently being investigated: jcommander tests unable to find testng after update.
  • junit5: an issue with the jar classpath (#1028991), now fixed in Debian.

python-testtools migration:

  • synced package with debian
  • arm64 auto package tests fail in Ubuntu (pass in Debian). Investigating.
1 Like


  • fix problems in error handling
  • merge request created to handle umount errors


  • merge request created to handle UEFI booting systems without variable support


  • Created wiki pages describing Ubuntu installation on PolarFire Icicle Kit and Nezha D1

Lenovo X13s

  • worked on make-shift installer based on preinstalled server image
1 Like

I did more than just working on Apport, but I haven’t recorded that.


1 Like

Drown in triaging FTBFS with Go1.20. Following uploads are directly to Debian.

Go compiler:

  • golang-1.19 1.19.5 point release (after approval from RT)
  • golang-1.20 1.20~rc3-2. Find a regression when triaging FTBFS in golang-github-minio-minio-go package.

Fix Go packages with Go1.20:

  • golang-github-lucas-clemente-quic-go 0.29.2-1 (This is sponsored by @jawn-smith to sync from Debian experimental)
  • golang-github-hashicorp-go-sockaddr 1.0.2-2
  • golang-github-magiconair-properties 1.8.7-1
  • golang-github-joyent-gocommon 0.0~git20161202.b787089-4
  • golang-github-vmware-govmomi 0.24.2-2
  • golang-github-traefik-yaegi 0.14.3-1

Open issues to upstream for failure with Go1.20

  • crewjam/saml#485
  • gatherstars-com/jwz#6
  • google/go-github#2626
  • henvic/httpretty#16
  • jhillyerd/enmime#269

Triage Go packages proposed migration

  • Mostly just asking @ginggs to retry tests. Thanks!
  • Fix golang-github-klauspost-compress autopkgtest in 1.15.12+ds1-3
1 Like

Short week due to holiday


  • Reviewed Docs update PR from OEM team
  • Merged main into feature branch
  • Cleaned up test cases to remove artifacts from the git tree
  • Updated snap to build on core22
  • Tested snap on new architectures
  • Began creating image_definition.yaml files for rootfs base tarballs


  • Take home test reviews
  • Interview prep
  • Investigating pcre2 proposed migration - still one failure
  • Investigating coreutils proposed migration - @adrien to fix armhf datefudge failure
  • Investigating gnupg2 proposed migration - has migrated
1 Like




  • pulse planning & story point estimations
  • candidate interviews & take-home tests
1 Like

Been busy with merges and migrations


  • Worked on gnutls, p11-kit and libgcrypt merges (code ready, soon to be pushed for review)
  • Made pyyaml migrate
  • Merged botch for the above
  • Spent more time on coreutils’ migration:
    • uqm and dgit only required re-triggers (uqm tests are flaky, dgit had
      connection issues IIRC)
    • coreutils itself switched to 64-bit time_t and this makes datefudge and
      faketime without effect; this is likely to have consequences in other
      packages (especially for reproducible builds) since it applies to the
      ‘date’ binary for instance; summary and proposals to follow soon
  • Investigated autopkgtest failures triggered by my openssl merge:
    • ecflow: only glanced at it
    • mysql: re-triggered because a test timed out
    • petsc: only glanced at it
    • seqkit: no clue, fails on my machine but not on the same test as in the
      test runners
1 Like


  • Removed pyside2 from the never_run list given improvements in the time to launch instances / run tests.
  • Submitted an MP modifying how we wait when creating instances in a cloud region.
  • Submitted an MP documenting how to remove EoL releases from the autopkgtest database.
  • Updated the autopkgtest documentation at so the above change would be available.
  • Submitted an RT regarding issues with creating amd64 instances in lgw01.
  • Created a new grafana graph which only shows autopkgtest runs that end with exit code 16 - a testbed failure.


  • Updated with 22.04.2 and Lunar.
  • Updated release team tasks to include updating more often.
  • Created a Release Tracking document for 23.04 with details about daily image build failures.
  • Added @xypron to notify-addresses for ubuntu-server builds so they’d get riscv64 build failures.
  • Testing of the Lunar desktop installer w/ nvidia and w/ nouveau.
  • Reported a bug regarding “Try Ubuntu” being available after already trying it.
  • Reported a bug regarding adding ubuntu-unity support to ubuntu-release-upgrader. Notified @rs2009 about the bug and tagged it rls-kk-incoming for the Foundations team.
  • Reviewed and merged an updated ubuntu-release-upgrader MP adding a FIPS upgrade quirk.
  • Built, tested, and uploaded the above change to the Focal SRU queue - again.
  • Ubuntu Bug Control application reviews.


  • Administrative work.
  • Reviews of candidate essays.
  • Short week due to illness.

Server Team

  • Jenkins Ansible playbook:
    • Install pipx on the nodes
    • Replace the azure-cli virtualenv with a pipx environment. Update the azure-ha Jenkins jobs accordingly.
    • Install oci-cli via pipx. Will be used for cloud-init integration testing on Oracle.
  • Cloud resource cleanup:
    • Add cleanup job for ec2 instances (internal repo)
    • Add cleanup job for Azure instances and resource groups (internal repo)
    • Provided feedback on handling on resource allocation to IS
  • cloud-init:
    • Integration testing: add cleanup step for ec2 instances
  • pycloudlib:
  • simplestreams:
    • simplestreams-tests: fixup retry logic for container cleanup in Jenkins
  • Weekly bug triage rotation


  • Continued monitoring of the state of s390x on bos02, reporting issues to IS.
  • Repeated issues with the arm64 VMs that host the armhf autopkgtests. Required manual intervention restarting the services.
    • This could be related to networking issues in bos02 scalingstack (same issue affecting s390x). The problem is not occurring anymore.
  • Investigating a case reported by @smb where a kernel autopkgtest run fails to install test dependencied.
    • I failed to reproduce locally. Trying to reproduce on the actual infrastructure.

ISO testing

  • Discussed the state of ISO testing and UTAH with @hellsworth
  • General review/scrub of the Jenkins ISO testing jobs

Release Team

  • Verified the status of the Ubuntu Server ISO images in preparation for the Jammy point release
  • Reported status at the point release coordination meeting

Ubuntu Pro


  • Had a call with IS discussing the status of the Jenkins charm and the requirements of the Ubuntu QA team and Server Team. This is in view of a general overhaul of the charm.
  • Reviewed candidate essays
1 Like
1 Like

Remainder of +1 rotation

On the last day of my +1 shift, I touched quite a few packages. See my full report for details.


I’ve mostly been working on getting Rust 1.65 ready for the upcoming Firefox version. I’m currently working on rustc 1.64 (not a typo, we need to do a step-by-step upgrade), and merged cargo 0.66 to lunar. This has taken most of my time, mostly because of a riscv64 FTBFS.


  • netplan: reviewed PR 299 in depth
  • apport: reviewed a new batch of PRs
1 Like



1 Like
  • proposed-migration sphinx-rtd-theme - retests, import pytorch ftbfs LP: #2002685
  • probert PR: 129 - do a refactor from a earlier fix to device filtering
  • subiquity PR: 1535 - fix a long-standing confusion item for users where the inability to reach geoip looks like an install failure
  • mini-iso-tools
    • PR: 1 - open new repo to handle a new archive package intended to handle the ‘boot-chooser’ mentioned in previous reports. Will be used to create a new ISO that allows choosing from a menu of available Ubuntu installer ISO versions.
    • PR: 2 - draft PR with CI for this repo, adding the CLA check and a simple package build
  • hiring stuff
  • specifications
1 Like

I didn’t have a chance to keep much notes or be aware of what I am doing, but um

  • python-apt 2.5.1 brings basic deb822 support. We cannot match on multiple suites or URIs or types yet, that comes later. Potentially dragging that card to next pulse, let’s see how we get on with the grub work.
  • apt 2.5.5 is out
  • grub 2.06 is now available in proposed for focal and bionic as well, signed with 2022v1 key
  • preparing fwupd-efi resigning with 2022v1 keys
  • worked the day coming up with solutions to avoid updating shim if there are no trusted kernels to boot.
  • investigated python-apt test suite failures causing build failure, found issue, didn’t have chance to commit fix yet, switched back to working on boot stuff.
1 Like


  • Investigating libcap2 autopkgtest failures


  • Minor fixes PR#316, PR#315, PR#314
  • Added warning and fixed a crash when invalid interface names are used PR#313
  • Working on packaging and autopkgtests for the NetworkManager+Netplan integration
  • A bunch of PRs reviewed and merged by @slyon (thanks!)


1 Like
  • lsb merge discarded in favour of sync with upstream (LP: #2002318)
  • lsb-release-minimal force sync’d with upstream
  • Investigated changes required in ubiquity to support FDE on Pi desktop images; too complex to persure pursue at this time when the work would have to be restarted under the forthcoming subiquity installer anyway
  • Worked on another sbuild merge (LP: #2003201) which turned into an extended wrestling match with autopkgtest (still on-going; autopkgtest appears to have the upper hand at this point)
  • Investigating differences between experimental pi image builds and the current nightly (some odd differences; unsure of their source currently)
  • Working on spec for an Ubuntu flavour of raspi-config
  • Pi meetings
1 Like
  • help the desktop team to package the adsys project after adding the Rust components (user-mount-handler branch, adsys#583)
  • help the desktop team to review the Rust component code (reviewed adsys#580)
  • currently working on backporting cargo to Kinetic (22.10), Jammy (22.04), Focal (20.04) and Bionic (18.04)
1 Like