This avoids endlessly requeuing the test if the test produces
an older result.
This will make tests "disappear" if the infrastructure returns
old results for newer triggers but avoids the problem right
now where we end up queuing the same tests every run.
For rolling out britney on a new machine, we want to generate update_excuses
and update_output to confirm it's working correctly all the way through, so
we don't want to use the global --dry-run option; but we *do* want to
disable queuing tests and instead let the production instance of britney
queue the tests while we simply query the results. Add support for
ADT_ENABLE=dry-run in britney.conf, parallelling the behavior of other
policies.
britney currently spends a majority of its runtime querying for baseline
test results that it won't find, and that it doesn't need. Refactor to
eliminate many of these excess queries.
@canonical.com is now DKIM signed and SPF published which means emails
from proposed-migration running on snakefruit sending direct would
likely be caught out. Since we're here, the project is Ubuntu related
so switch to using an @ubuntu.com address instead.
When querying swift there is no way to take results only newer than a
specified point, you can only query newer than or equal to. But for sqlite
we can absolutely use > instead of >= and avoid re-processing results we've
already seen.
Logging all force-reset-test hints for every package causes
about 850 MB of logs in the last run of 880 MB of logs in total,
let's only log ones matching the package instead, as we do for
force-badtest.
In Ubuntu, we only fetch results on demand, so we might not
have seen the results yet.
Debian always fetches results at the beginning so has all the
data ready.
Due to the number of hints in standing use in Ubuntu, hints.search() is an
expensive operation, and we call it once for *every single test* referenced
from -proposed. Since force-reset-test are a small proportion of the hints
in use, searching once for all the hints of this type and only searching
this subset for each autopkgtest improves performance (with 23000
autopkgtests referenced in -proposed, this saves roughly 1 minute of
runtime, or 11% on a 9-minute britney run; the number of packages in
-proposed is typically much higher at other points in the release cycle,
therefore the absolute improvement in performance is expected to be
greater.)
The force-reset-test hints are an Ubuntu delta so this is not expected to be
upstreamed; and it could eventually be dropped if and when baseline
retesting is implemented in Ubuntu and the number of hints required drops.
This could be implemented with a more generic, elegant solution in
HintsCollection, but again, the scalability problem of hints is hopefully
short-lived so I didn't consider it worth the investment here.
urlopen() supports non-http URLs, but when called on them, http-related
features are absent - such as getcode(). Make the code work with file:///
URLs.
Since we've never used remove hints in Ubuntu and never will, we shouldn't
pay the cost for them in britney runs. Remove references to them entirely,
to speed up britney runs.
This is an Ubuntu delta not expected to be upstreamed, and can be dropped at
some future point when we don't have 11,000 active hints. (This count of
hints is expected to drop dramatically once we have autopkgtest baseline
retesting.)
The existing code attempts to "short circuit" processing of binaries in
unstable by first checking if the package is subject to a removal hint.
This is invalid in Ubuntu for two reasons:
- we do not use removal hints for removing packages from the release pocket
- there are 11,000 hints in Ubuntu (due to force-reset-test) - searching all
hints is time-consuming and not a short-circuit at all.
Reorder the code so we only scan the hints if there's otherwise an
indication of something to do.
We currently skip the ALWAYSFAIL/REGRESSION handling for kernels. This
can lead to us missing genuine regressions in kernel uploads. The
idea is that results from one kernel flavour shouldn't influence
another.
We can keep this idea but do better and actually check for regressions:
when looking at results, if we're considering a kernel, only look at
results which were triggered by this kernel.
We currently have excuses.html saying:
x264 (2:0.155.2917+git0a84d98-2 to 2:0.159.2999+git296494a-2)
Migration status for x264 (2:0.155.2917+git0a84d98-2 to 2:0.159.2999+git296494a-2): BLOCKED: Cannot migrate due to another item, which is blocked (please check which dependencies are stuck)
Issues preventing migration:
Implicit dependency: x264 nageru (not considered)
Invalidated by implicit-dependency
Implicit dependency: x264 ffmpeg
Implicit dependency: x264 gst-plugins-ugly1.0
Implicit dependency: x264 handbrake
Implicit dependency: x264 libquicktime
Implicit dependency: x264 mplayer
Implicit dependency: x264 mythtv
Implicit dependency: x264 obs-studio
Implicit dependency: x264 vlc
Implicit dependency: x264 xpra
Additional info:
8 days old
Apart from `nageru` it's not clear here that some further packages need
working on to move this migration forward. In fact, the package
`gst-plugins-ugly1.0` is FTBFS
Migration status for gst-plugins-ugly1.0 (1.16.2-2build2 to
1.16.2-2build3): BLOCKED: Maybe temporary, maybe blocked but Britney
is missing information (check below)
as is xpra. The remaining packages are simply those which are part of
this SONAME transition, as such they have
Migration status for mplayer (2:1.3.0-8build6 to 2:1.3.0-8build7):
BLOCKED: Cannot migrate due to another item, which is blocked (please
check which dependencies are stuck)
All of these excuses are (or should be) "invalid", but only the ones
which are not BLOCKED_BY_ANOTHER_ITEM require direct action. We should
show this, so that maintainers can see what they need to concentrate on.
When outputting depdencency problems, consider the policy verdict of the
excuse. If it's one that indicates further work is required for that
excuse, indicate this by outputting "(not considered)".
We often introduce new kernels post-release, and we still want to not
trigger tests for the kernel image packages. Check for -meta in the
*source* suite as well as the target.
Fix the -meta name calculation for linux-signed-foo.
We currently skip running autopkgtests where there is an installability
problem, but in a few cases the depends policy notes these only, but
otherwise doesn't block migration on them.
In these cases, let's try to run the autopkgtests anyway. There will be
a few instances of uninstallability here, but since we migrate the
packages we should give them a chance to be tested.
We currently concatenate all triggers together into a string, but the
AMQP consumer expects this to be a list.
When using AMQP, keep the triggers as a list. Ensure that the "real"
trigger (the package being tested) is kept first, as before.
If we have this situation
- Binary in target taken over by new source
- Not cleaned up yet
- New source updated in source suite
We can have *three* copies of a source package visible to britney at
once. Handle this case by recording the source package a pkgid was seen
in before, and trying to remove from that and the current source.
On some distros (Ubuntu), arch:all packages are built along with one of
the architectures. We shouldn't be listing 'all' as its own arch in this
case. Instead we filter out the binaries except for on the
'all_buildarch'.
Julian Andres Klode
Michael Hudson-Doyle
Brian Murray
Łukasz 'sil2100' Zemczak
Steve Langasek
Haw Loeung
Iain Lane
Iain Lane
Colin Watson