machine-mac: unable to clean up gvproxy: checking if process running: input/output error

[edsantiago]

I don't think this is the same one being discussed in Slack right now, but please close if it is.

/Users/MacM1-1-worker/ci/task-4908461394231296/bin/darwin/podman machine reset --force
Error: 1 error occurred:
	* unable to clean up gvproxy: checking if process running: input/output error

Four instances, all of them in the last 24 hours.

• darwin : machine-mac podman darwin rootless host sqlite
  • PR CI: e2e: use distinct ports, not just 5000 #21834
    • 02-27 07:23 in machine-mac podman darwin rootless host sqlite: unknown failure in test
  • PR some podman compose fixes #21825
    • 02-27 08:39 in machine-mac podman darwin rootless host sqlite: unknown failure in test
  • PR Enable lint for Darwin and fix identified issues #21753
    • 02-26 11:56 in machine-mac podman darwin rootless host sqlite: unknown failure in test
  • PR Change QEMU netdev to Unix domain socket #21594
    • 02-26 13:57 in machine-mac podman darwin rootless host sqlite: unknown failure in test

[baude]

is this a flake that fails a test ?

[baude]

it looks like remove does it too; and yes, it is a flake; i will try to look next time im on my mac

[edsantiago]

I would assume that it causes a CI failure, since by definition that's what my flake-logger logs. Confirmation. (Wow, it is so hard to work with non-logformatted logs....)

[baude]

@edsantiago are you still seeing this error?

[edsantiago]

The last one in my logs was on 02-27, shortly after I filed this issue.

[baude]

ok, i was asking because #21778 had a potential fix for this.

[edsantiago]

CI load has been super low the last few days, but I've seen no darwin flakes. If you're confident that #21778 fixed the gvproxy issue, feel free to close this: one less dangling sword to track. Ah what the heck, I'll close it myself.