No cirrus by lsm5 · Pull Request #697 · containers/aardvark-dns

lsm5 · 2026-05-08T15:29:32Z

See individual commits.

packit-as-a-service · 2026-05-08T15:51:49Z

Tests failed. @containers/packit-build please check.

Luap99 · 2026-05-08T16:03:30Z

Thanks. Note to myself once this is ready we need to update merge protection settings with the new job names.

I think we looked at this a while ago but can packit do do task dependencies? I suppose it is not critical but it is nice to first run validate before triggering all the other jobs.
And that would allow us to have final status task, like "Total success". Without this in the github merge protection settings I need to set all exact task names like show in the github interface, testing-farm:fedora-44-x86_64:validate, etc... which is highly annoying because not only do I need to set a ton of names these names will change with distro upgrades so require ongoing churn for updates.

Luap99 · 2026-05-08T16:05:32Z

+/msrv:
+    summary: Minimum Supported Rust Version (MSRV) check
+    test: |
+        curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y --default-toolchain 1.86


just a heads up #695 will bump to rust 1.88

ack, I'll rebase on that PR

The test will fetch rust version directly from Cargo.toml now.

lsm5 · 2026-05-08T16:58:39Z

And that would allow us to have final status task, like "Total success". Without this in the github merge protection settings I need to set all exact task names

There's discussion and WIP happening in packit/packit-service#1720 .

lsm5 · 2026-05-08T17:29:56Z

I'd say this is ready for reviews. Will keep in draft until #695 is done as I rebased it already but this PR doesn't need to depend on it AFAICT.

@containers/netavark-maintainers @timcoding1988 PTAL

Luap99

LGTM overall, just some minor comments.

Luap99 · 2026-05-11T11:56:29Z

+    adjust:
+        enabled: false
+        when: initiator is not defined or initiator != packit


this is done so they do not run as part of Fedora CI/dist git setup I assume?

Unit tests are real test cases and they are faster than integration tests anyway so I think it would also make sense to run them as gating function. Not that I think this would catch anything particular but if we have this we should run this, better safe than sorry here.

Note I do agree we must skip validate, because a new rust will likely cause new validation errors due new lint rules.

Are you ok leaving this for followup? IIUC, this would involve cloning the source tree in some form at downstream CI and we can't just use the -tests subpackage. That would need some more changes and downstream testing.

ack sure absolutely not blocking, lets get it working as main upstream CI first and foremost right now

Not sure if we would actually want to do this but AFAIK some packages run unit tests as part of the rpm spec %check which would have access to the full source tree? Anyhow not important right now.

Not sure if we would actually want to do this but AFAIK some packages run unit tests as part of the rpm spec %check which would have access to the full source tree?

Ack yes, but that runs as part of rpmbuild / koji build which would be not so nice imo. I'm thinking either package the source as a -src subpackage or fetch TMT plan from the upstream tag. Let's see later ...

Luap99 · 2026-05-11T12:00:33Z

+    trigger: pull_request
+    identifier: validate
+    skip_build: true
+    notifications: *test_failure_notification


Do you actually want this? If this is the primary CI than it is the author/maintainers job to look at the tests results. There is no point in spam pinging you for this.

If I or another maintainer sees general packit problems we can ping you then directly so you do not drown in noise.

will remove all notifications in the upcoming update to this PR. Thanks!

Luap99 · 2026-05-11T12:04:41Z

+        curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y --default-toolchain "$MSRV"
+        source $HOME/.cargo/env


generally not a huge fan of this as hoc installs due to flake risk but I guess we will see how it turns out. If it starts flaking to often we can consider another approach.

ok, leaving this as is then.

timcoding1988 · 2026-05-11T13:48:02Z

Tests failed. @containers/packit-build please check.

small. am i missing something here, looks like tests all passed here. also the total success check seems still visible from here. maybe remove it possible ?
other than that, all LGTM. thank you for setting this up !

lsm5 · 2026-05-11T14:52:09Z

small. am i missing something here, looks like tests all passed here.

I think Packit does a few retries so that's why you saw the failure notification without failures. But I'm removing the notifications in the upcoming force push, so this should be a non-issue.

also the total success check seems still visible from here. maybe remove it possible ?

I think that needs to be removed from branch protection in repo settings which @Luap99 will take care of.

other than that, all LGTM. thank you for setting this up !

Thanks

Luap99 · 2026-05-11T15:01:23Z

I think that needs to be removed from branch protection in repo settings which @Luap99 will take care of.

correct as long as a task is required in merge rule settings GH will display it

lsm5 · 2026-05-11T17:14:15Z

/packit retest-failed

lsm5 · 2026-05-11T19:09:10Z

three subnets, one container on two of the subnets, network connect flaked once. Is that known?

I've rebased only on main, so this is good to go without waiting for #695 .

Luap99 · 2026-05-11T19:23:46Z

three subnets, one container on two of the subnets, network connect flaked once. Is that known?

I have no observed any failures with cirrus in the past, but overall a lot of these test things are timing related so some flakes can be possible I think.

Is there a way like in cirrus where I can see all runs from the repo so I can see the history which failed run you meant, i.e.
https://cirrus-ci.com/github/containers/aardvark-dns/main

Also the cirrus task always displayed at the top if they were a rerun and the previous log, I don't packit marks rerun in any way do they?

lsm5 · 2026-05-11T19:25:07Z

Is there a way like in cirrus where I can see all runs from the repo so I can see the history which failed run you meant, i.e. https://cirrus-ci.com/github/containers/aardvark-dns/main

Also the cirrus task always displayed at the top if they were a rerun and the previous log, I don't packit marks rerun in any way do they?

Hmm, don't know offhand. I'll ask around and file a request if they don't have it already.

Luap99 · 2026-05-11T19:30:32Z

https://dashboard.packit.dev/projects/github.com/containers/aardvark-dns?forge=github.com&namespace=containers&repo-name=aardvark-dns

seems to show a list of PRs and its copr builds links but not the testing farm links AFAICT

lsm5 · 2026-05-11T19:35:20Z

I think https://dashboard.packit.dev/jobs/testing-farm will show all jobs, but that's currently a nightmare to navigate, no filtering from what I can tell.

cc @LecrisUT see #697 (comment)

LecrisUT · 2026-05-11T19:49:51Z

Unfortunately I do not have any better solution and web-dev is not in my skillset to navigate https://github.com/packit/dashboard/. If there are ideas of FOSS dashboards that we could host, we can explore that as well

Luap99 · 2026-05-11T19:52:33Z

I think https://dashboard.packit.dev/jobs/testing-farm will show all jobs, but that's currently a nightmare to navigate, no filtering from what I can tell.

right. One of the nice thing of the cirrus UI is I can filter this by branch. So I can only look at results on main because it should not contain broken code so any failure there is most likely a flake and could be worth reporting/tracking/fixing...
And of course all of this is available via API as well so I have some basic scripts to check flake rates via that.

On PRs contributors push broken code all the time so a failing status on a PR does not mean anything really.

So overall I think it would be a nice feature request if we could have a webpage+API to get a list of all created copr and testing-farm jobs filtered by repo and branch if possible.

lsm5 · 2026-05-12T17:45:49Z

@containers/netavark-maintainers I think we're good to go here unless anything unaddressed. PTAL.

Luap99 · 2026-05-12T18:07:40Z

yes logic wise this seems good to me, I like to have #695 merged first before I update the merge protection settings here

Signed-off-by: Lokesh Mandvekar <lsm5@linux.com>

- Remove all COPR build failure notifications - Remove all test failure notifications from integration jobs - Add identifiers to integration jobs (integration-fedora, integration-centos) - Add tmt_plan to existing integration jobs (/plans/integration) - Add new validate test job (skip_build: true, fedora-latest-x86_64) - Add new unit test job (skip_build: true, fedora-latest-x86_64) Signed-off-by: Lokesh Mandvekar <lsm5@linux.com>

Signed-off-by: Lokesh Mandvekar <lsm5@linux.com>

Fixes: RUN-4616 Signed-off-by: Lokesh Mandvekar <lsm5@linux.com>

ashley-cui · 2026-05-13T13:12:42Z

LGTM

lsm5 · 2026-05-13T13:32:59Z

Rebased on the latest main. PTAL @containers/netavark-maintainers

Luap99

LGTM,

branch protection is set to require this now

testing-farm:fedora-44-x86_64:msrv
testing-farm:fedora-44-x86_64:unit
testing-farm:fedora-44-x86_64:validate
testing-farm:fedora-44-x86_64:integration-fedora
testing-farm:fedora-44-aarch64:integration-fedora

When fedora upgrades to 45 we need to update this but one update every 6 months is not the end of the world.

lsm5 force-pushed the no-cirrus branch from 218e34b to 2f76095 Compare May 8, 2026 15:37

Luap99 reviewed May 8, 2026

View reviewed changes

lsm5 force-pushed the no-cirrus branch 2 times, most recently from 28ac9a5 to 4852111 Compare May 8, 2026 17:12

Luap99 reviewed May 11, 2026

View reviewed changes

lsm5 force-pushed the no-cirrus branch 3 times, most recently from c167ed1 to 731d0b4 Compare May 11, 2026 15:40

lsm5 marked this pull request as ready for review May 11, 2026 19:07

lsm5 mentioned this pull request May 12, 2026

(flake) integration: three subnets, one container on two of the subnets, network connect #699

Open

lsm5 added 3 commits May 13, 2026 09:05

TMT: remove tags and filters, use individual plans

adbcf51

Signed-off-by: Lokesh Mandvekar <lsm5@linux.com>

TMT: improve prepare steps

9de40d6

Signed-off-by: Lokesh Mandvekar <lsm5@linux.com>

lsm5 added 2 commits May 13, 2026 09:05

Packit/TMT: add MSRV job

f2e6e71

Signed-off-by: Lokesh Mandvekar <lsm5@linux.com>

Remove Cirrus CI configuration

d13167e

Fixes: RUN-4616 Signed-off-by: Lokesh Mandvekar <lsm5@linux.com>

lsm5 force-pushed the no-cirrus branch from 731d0b4 to d13167e Compare May 13, 2026 13:06

Luap99 approved these changes May 13, 2026

View reviewed changes

Luap99 merged commit b25cf6a into containers:main May 13, 2026
30 checks passed

lsm5 deleted the no-cirrus branch May 13, 2026 14:40

		curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs \| sh -s -- -y --default-toolchain "$MSRV"
		source $HOME/.cargo/env

Conversation

lsm5 commented May 8, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

packit-as-a-service Bot commented May 8, 2026

Uh oh!

Luap99 commented May 8, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

lsm5 commented May 8, 2026

Uh oh!

lsm5 commented May 8, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Luap99 left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

timcoding1988 commented May 11, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

lsm5 commented May 11, 2026

Uh oh!

Luap99 commented May 11, 2026

Uh oh!

lsm5 commented May 11, 2026

Uh oh!

lsm5 commented May 11, 2026

Uh oh!

Luap99 commented May 11, 2026

Uh oh!

lsm5 commented May 11, 2026

Uh oh!

Luap99 commented May 11, 2026

Uh oh!

lsm5 commented May 11, 2026

Uh oh!

LecrisUT commented May 11, 2026

Uh oh!

Luap99 commented May 11, 2026

Uh oh!

lsm5 commented May 12, 2026

Uh oh!

Luap99 commented May 12, 2026

Uh oh!

ashley-cui commented May 13, 2026

Uh oh!

lsm5 commented May 13, 2026

Uh oh!

Luap99 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

lsm5 commented May 8, 2026 •

edited

Loading

lsm5 commented May 8, 2026 •

edited

Loading

timcoding1988 commented May 11, 2026 •

edited

Loading