Apple M1, macOS Monterey - stream copy error: reading form a closed fifo

[BenjaminPoulain]

On macOS Monterey, we randomly run into failures while running small containers on after the other.

When the issue happens, the logs show the following error: dockerd time="..." level=error msg="stream copy error: reading from a closed fifo". Logs attached. vm.zip

The test themselves do:

• Install Docker.
• Wait for the VM to come up.
• Repeatedly do Create containers -> Execute I/O operations -> Delete Containers.

The bug typically happens within 50 repetitions but it's random, it's not always after running N containers.

From a previous run, the console.log contained this which seems relevant:

[45342.682241] oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),cpuset=docker,mems_allowed=0,global_oom,task_memcg=/podruntime/docker,task=dockerd,pid=18194,uid=0
[45342.682728] Out of memory: Killed process 18194 (dockerd) total-vm:5273852kB, anon-rss:58204kB, file-rss:0kB, shmem-rss:0kB, UID:0 pgtables:1992kB oom_score_adj:1
[45343.226076] oom_reaper: reaped process 18194 (dockerd), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB

This makes me speculate it could be a leak on the guest side.

[stephen-turner]

Thanks, Benjamin, we'll have a look. cc @djs55

[docker-robott]

Issues go stale after 90 days of inactivity. Mark the issue as fresh with /remove-lifecycle stale comment. Stale issues will be closed after an additional 30 days of inactivity.

Prevent issues from auto-closing with an /lifecycle frozen comment.

If this issue is safe to close now please do so.

Send feedback to Docker Community Slack channels #docker-for-mac or #docker-for-windows. /lifecycle stale

[djs55]

/lifecycle frozen