-
Notifications
You must be signed in to change notification settings - Fork 2.1k
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
cgroupv2 checkpoint/restore is not working #2328
Comments
I already have a patch to fix this, just need more time to create a PR |
Currently, CRIU 3.14 (not yet released) is required to checkpoint fstype=cgroup2 mount[1]. Also, some runc changes are needed (since cgroup2 mount is the real mount, not an external bind mount[2]). Therefore, let's disable checkpoint tests for cgroup v2 (the ones that are running on a Fedora 31 Vagrant box) for now. [1] checkpoint-restore/criu@3a15076405c [2] opencontainers#2328 Signed-off-by: Kir Kolyshkin <[email protected]>
When I implemented the freezer CRIU cgroup2 changes for runc and CRIU it was still working. That was a few weeks ago. The integration tests are actually running checkpoint/restore using vagrant-fedora-31 with cgroup2 in the container and outside of the container: https://github.com/opencontainers/runc/blob/master/libcontainer/integration/checkpoint_test.go#L86 I am certain I do not understand the whole cgroup details as good as you, just wanted to mention that some checkpoint/restore tests using cgroup2 are working. |
Currently, CRIU 3.14 (not yet released) is required to checkpoint fstype=cgroup2 mount[1]. Also, some runc changes are needed (since cgroup2 mount is the real mount, not an external bind mount[2]). Therefore, let's disable checkpoint tests for cgroup v2 (the ones that are running on a Fedora 31 Vagrant box) for now. [1] checkpoint-restore/criu@3a15076405c [2] opencontainers#2328 Signed-off-by: Kir Kolyshkin <[email protected]>
@adrianreber I'll take a closer look at it, but I think I found (and fixed) a genuine bug, please see #2335. The reason it works in your case is that you explicitly set Maybe I need to redo my fix differently, changing "cgroup" to "cgroupv2" in case we're using v2 unified hierarchy, but let's discuss it in #2335 |
One reason it is not working is cgroupv2 mount (and BTW cgroupv1 mounts in case
--cgroupns
is used) should not be treated like (presented to criu as) external bind mount(s).fstype
ofcgroup
)fstype
iscgroup2
)There might be other bugs, too.
The text was updated successfully, but these errors were encountered: