I've split the upstream test in 247.3-5ubuntu1 and also forwarded the change:
https://salsa.debian.org/systemd-team/systemd/-/merge_requests/128
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1915126
This bug was fixed in the package systemd - 247.3-1ubuntu4
---
systemd (247.3-1ubuntu4) hirsute; urgency=medium
[ Dimitri John Ledkov ]
*
d/p/debian/UBUNTU-resolved-Mitigate-DVE-2018-0001-by-retrying-NXDOMAIN-with.patch:
Patch updated to reduce log level to debug
Sorry for not replying to comment #11, although we chatted on IRC.
The main difference is the passing of the test name I think, but the
full autopkgtest command that I used is at the top of the log. I hope
using this exact command and adding --shell-fail is enough for anyone to
reproduce.
--
@laney Thank you for looking into the problem.
I will mark systemd-fsckd as flaky and look into the root cause later.
It used to be flaky before as well.
I have checked the last 10 failed runs from
http://autopkgtest.ubuntu.com/packages/systemd/hirsute/amd64 , with the
last one being from
@paelzer I haven't made changes related to the failure rate. I guess the
infra got lighter load for some time.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1915126
Title:
autopkgtest times out (or
Also @rbalint - it seems that the success rate with 247.3-1ubuntu2 got a bit
better.
Were there changes in the upload or was it actually converted to be a big_test
already ?
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
Thanks Laney for chiming in here (as one of the affected/blocked I appreciate
every help).
I was last week already trying that myself on canonistack and for me it didn't
reproduce.
This is the command I used:
autopkgtest --no-built-binaries --apt-upgrade --testname=systemd-fsckd
Ok, here you go. I reproduced the (or a) failure in Canonistack. This
was the first run FWIW, so it wasn't too hard to get it to fail:
autopkgtest [14:05:12]: test systemd-fsckd: [---
test_fsck_failure (__main__.FsckdTest)
Ensure that a failing fsck doesn't prevent fsckd to
I'm running a test on Canonistack. I'm hoping that this will reproduce
the failures we're seeing, and then they can be investigated.
That is closer to what is running in production than the QEMU runner.
I don't think we have evidence that this is an 'infra issue', so I'm
marking the
** Tags added: update-excuse
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1915126
Title:
autopkgtest times out (or fails before that) in hirsute on test infra,
passes locally
To manage
On Thu, Feb 11, 2021 at 01:49:45PM -, Balint Reczey wrote:
> There is not much to do about the failures on systemd's side. The tests
> are reliably passing in qemu on my i5-4670 testbed from 2013, also
> passing on @paelzer's presumably newer machine.
>
> This is an infra issue. The Release
There is not much to do about the failures on systemd's side. The tests
are reliably passing in qemu on my i5-4670 testbed from 2013, also
passing on @paelzer's presumably newer machine.
This is an infra issue. The Release Team can decide if they want to hint
the test and lose gating coverage,
** Also affects: auto-package-testing
Importance: Undecided
Status: New
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1915126
Title:
autopkgtest times out (or fails before that) in
** Merge proposal linked:
https://code.launchpad.net/~rbalint/autopkgtest-cloud/+git/autopkgtest-cloud/+merge/397740
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1915126
Title:
autopkgtest
** Merge proposal unlinked:
https://code.launchpad.net/~rbalint/autopkgtest-cloud/+git/autopkgtest-cloud/+merge/383953
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1915126
Title:
autopkgtest
@paelzer I seems infra issues got worse, but there is a pending review
to bump systemd to big which I linked. Even in the absence of the infra
issues it would have made sense to make system run big test instances,
but I expect the big instances would also improve the current situation.
Instead
** Merge proposal linked:
https://code.launchpad.net/~rbalint/autopkgtest-cloud/+git/autopkgtest-cloud/+merge/383953
** Summary changed:
- autopkgtest times out (or fails before that) in hirsute
+ autopkgtest times out (or fails before that) in hirsute on test infra, passes
locally
--
You
Running the same on a Canonistack instance creates "different" issues
and won't help much debugging it. Locally it works fine even on multiple
re-runs.
@rbalint / others that feel systemd'ish: What shall we do about this
short term to resolve all blocking packages?
--
You received this bug
>From the log:
402209 Checking in progress on 1 disk (92.7% complete)fsckd:1:92.8:Checking in
progress on 1 disk (92.8% complete)
402210 Checking in progress on 1 disk (92.8% complete)fsckd:1:92.9:Checking in
progress on 1 disk (92.9% complete)
402211 Checking in progress on 1 disk (92.9%
On average the upstream tests loose ~55 minutes compared to former runs (~90 ->
~145 minutes).
Also that test times out:
autopkgtest [20:14:12]: test upstream: ---]
upstream FAIL timed out
Also the later test for systemd-fsckd hangs until
I've compared older good and newer runs that timeout.
$ for i in good.hirsute-amd64.*; do printf "\n\n$i\n"; grep
'^autopkgtest .* test .*: - - - - - - - - - - results' $i; done
We see that good cases need ~110-130 minutes to complete "test boot-
smoke" and then another 4-12 minutes to
In our worker config we see https://git.launchpad.net/autopkgtest-cloud:
long_tests timeout => ~11h
big_pkg timeout => ~5.5h
default timeout => ~3h
Currently systemd is only marked as long_test on ppc64el, so ~3h is what
applies to us here.
The systemd runs that failed with timeouts are all in
22 matches
Mail list logo
