Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

CI: add actuated-arm64 #4142

Merged
merged 5 commits into from
Apr 24, 2024
Merged

CI: add actuated-arm64 #4142

merged 5 commits into from
Apr 24, 2024

Conversation

AkihiroSuda
Copy link
Member

@AkihiroSuda AkihiroSuda commented Dec 12, 2023

See https://actuated.dev/blog/arm-ci-cncf-ampere.
Thanks to Alex Ellis, Ampere Computing, and Equinix.


Host information:

  • CPU: aarch64 (ARMv8)
  • Kernel: 5.10.201
    • Lacks CONFIG_CHECKPOINT_RESTORE, CONFIG_BLK_CGROUP_IOCOST, etc.
  • Cgroup: v2
  • OS: Ubuntu 22.04
    • Lacks newuidmap, newgidmap, etc. (still apt-gettable)
    • sshd is not running

@AkihiroSuda AkihiroSuda marked this pull request as draft December 12, 2023 11:14
@AkihiroSuda
Copy link
Member Author

cc @alexellis @vielmetti

@alexellis
Copy link

Thanks for the ping. The build should run once the GitHub app is installed.

@alexellis
Copy link

Thanks to @amye the app is now installed.

@AkihiroSuda AkihiroSuda force-pushed the actuated branch 23 times, most recently from 715693c to c844665 Compare December 15, 2023 08:33
@alexellis
Copy link

The Kernel may have been "born" some time ago, but isn't old, it's a relatively new patch revision.

That said, if 5.10.x and Ubuntu 22.04 are not going to work for CRIU, can you merge this as a step forward for Arm support, and keep an action item to enable CRIU e2e tests for when we have something like a 6.x Kernel available?

Would having CI for 99% of runc running on Arm CI be better, than blocking everything due to not having CRIU coverage?

@adrianreber
Copy link
Contributor

Maybe it is enough to retry with the latest kernel. The latest error I see here is not related to restartable sequences and the files which are running in the test containers are using binaries from glibc 2.31 (busybox) and glibc 2.28 (debian).

At this point, with the latest kernel, it might just work.

@AkihiroSuda
Copy link
Member Author

Rebased.
I guess the criu stuff can be revisited in a separate PR later.

@alexellis
Copy link

We have a 6.x Kernel on the roadmap, but it's a significant amount of labour and iteration to take a Firecracker guest Kernel and only enable the specific features for containers and ebpf, so if that resolves the issues with CRIU we will reach out when it's ready.

I'd suggest you merge this PR and skip those tests for the time being on arm64. The project's test coverage on native hardware will be in a much better position already. Would you agree?

@AkihiroSuda
Copy link
Member Author

Ping @opencontainers/runc-maintainers
Let's merge this

Copy link
Contributor

@kolyshkin kolyshkin left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM overall, just one minor nit

Signed-off-by: Akihiro Suda <akihiro.suda.cz@hco.ntt.co.jp>
kcmp is often missing: https://man7.org/linux/man-pages/man2/kcmp.2.html

> Before Linux 5.12, this system call is available only if the
> kernel is configured with CONFIG_CHECKPOINT_RESTORE, since the
> original purpose of the system call was for the
> checkpoint/restore in user space (CRIU) feature.  (The
> alternative to this system call would have been to expose
> suitable process information via the proc(5) filesystem; this was
> deemed to be unsuitable for security reasons.)  Since Linux 5.12,
> this system call is also available if the kernel is configured
> with CONFIG_KCMP.

Signed-off-by: Akihiro Suda <akihiro.suda.cz@hco.ntt.co.jp>
Signed-off-by: Akihiro Suda <akihiro.suda.cz@hco.ntt.co.jp>
On cgroup v2, cpu burst needs kernel >= 5.14
torvalds/linux@f418371

Signed-off-by: Akihiro Suda <akihiro.suda.cz@hco.ntt.co.jp>
README.md Show resolved Hide resolved
See <https://actuated.dev/blog/arm-ci-cncf-ampere>.
Thanks to Alex Ellis, Ampere Computing, and Equinix.

Host information:
* CPU: aarch64 (ARMv8)
* Kernel: 5.10.201
  * Lacks ~CONFIG_CHECKPOINT_RESTORE~, CONFIG_BLK_CGROUP_IOCOST, etc.
* Cgroup: v2
* OS: Ubuntu 22.04
  * Lacks newuidmap, newgidmap, etc. (still apt-gettable)
  * sshd is not running

vmmeter is added from:
https://gist.github.com/alexellis/1f33e581c75e11e161fe613c46180771#file-metering-gha-md

Signed-off-by: Akihiro Suda <akihiro.suda.cz@hco.ntt.co.jp>
@lifubang lifubang merged commit 52bd9fd into opencontainers:main Apr 24, 2024
40 checks passed
@lifubang
Copy link
Member

Thanks @alexellis

@alexellis
Copy link

You're welcome.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

7 participants