top: fallback to execing ps(1)

[vrothberg]

Fallback to executing ps(1) in case we hit an unknown psgo descriptor.
This ensures backwards compatibility with docker-top, which was purely
ps(1) driven.

Also support comma-separated descriptors as input.

Signed-off-by: Valentin Rothberg rothberg@redhat.com

[openshift-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: vrothberg

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Details

Needs approval from an approver in each of these files:

• OWNERS [vrothberg]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[TomSweeneyRedHat]

ping back to @vrotherber's radar

[vrothberg]

Thanks a lot, @TomSweeneyRedHat! I didn't notice it was tagged for 1.3.0 (didn't receive an email for that). I will pick this up tomorrow morning :)

[rh-atomic-bot]

☔ The latest upstream changes (presumably #2985) made this pull request unmergeable. Please resolve the merge conflicts.

[vrothberg]

Updated ✔️ @TomSweeneyRedHat @mheon @rhatdan PTAL

FYI: the logic is now a bit simpler. If we hit psgo.ErrUnknownDescriptor we just go ahead and execute ps(1). If that fails, we return the previous psgo error. HOWEVER: there is some crazy thing going on in the wild as the ps(1) in Busybox and Alpine is totally happy with ps invalid-arg. I suggest that's something we can accept. Otherwise, we'd run into all kinds of weird corner cases.

[rhatdan]

I thought I read where @mheon suggested that we not document this behaviour? Or was that someone else? Or was I dreaming?

[vrothberg]

@rhatdan, good point! I missed mentioning that I would not do that due to the inconsistent behavior of ps across linux distros, see the following example:

$ ./bin/podman run -d --rm debian sleep 200
57c1c39755641263df3a0aa951891846ac31765a346b3c8c2378a46c3604be95
$ ./bin/podman top -l invalid
Error: 'invalid': unknown descriptor

$ ./bin/podman run -d --rm alpine sleep 200
aeba614cc3802d8c39bcebd3d4025efaac2fbfc8dffe285acf70f52e9bb6003b
$ ./bin/podman top -l invalid
PID   USER     TIME  COMMAND
    1 root      0:00 sleep 200

I removed it from the help message but I vote for keeping it in the man pages. Otherwise, users have to figure out on their own whats going on which might not be the best experience.

[rhatdan]

OK.
/lgtm