fix: improved "containers/{name}/wait" endpoint

[matejvasek]

Previous implementation wasn't working with short living containers.
Polling missed the very brief moment the container was running.

Now Event API is used instead of polling.

resolves #10256

[matejvasek]

/cc @mheon

[rhatdan]

One Nit

[openshift-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: matejvasek, rhatdan

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

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [rhatdan]

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

[mheon]

Looks like this could leak goroutines - think you need to close the channel once you have the event so event reading terminates

…

On Sat, May 8, 2021 at 06:19 Daniel J Walsh ***@***.***> wrote: ***@***.**** commented on this pull request. ------------------------------ In pkg/api/handlers/utils/containers.go <#10271 (comment)>: > @@ -7,6 +7,8 @@ import ( "strconv" "time" + "github.com/containers/podman/v3/libpod/events" + LGTM — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#10271 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB3AOCBTP4PTDYMRN4XOT2DTMUF2ZANCNFSM44LACLWA> .

[matejvasek]

Looks like this could leak goroutines - think you need to close the channel once you have the event so event reading terminates

@mheon Good catch. Closing the eventChannel would cause panic as the gorutine would try to write to closed channel. Correct way to terminate the event watcher is to cancel the passed context. The context here is done once the HTTP request is ended. So in the end there is no leak. However I will add explicit cancellation for clarity.

[matejvasek]

@mheon I tried to compile podman without systemd support and this is not working. Any idea why?

Edit: my bad it works

[vrothberg]

LGTM, thank you!

[rhatdan]

/approve
/lgtm
/hold

[openshift-ci]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: matejvasek, rhatdan

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

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [rhatdan]

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

[TomSweeneyRedHat]

fedora-33 root test is failing and I don't think it's a flake.

[+0650s] not ok 153 podman pod create - hashtag AllTheOptions
[+0650s] # (from function `die' in file test/system/helpers.bash, line 412,
[+0650s] #  from function `run_podman' in file test/system/helpers.bash, line 220,
[+0650s] #  in test file test/system/200-pod.bats, line 208)
[+0650s] #   `run_podman build -t $infra_image - << EOF' failed with status 125
[+0650s] # # /var/tmp/go/src/github.com/containers/podman/bin/podman rm --all --force
[+0650s] # # /var/tmp/go/src/github.com/containers/podman/bin/podman ps --all --external --format {{.ID}} {{.Names}}
[+0650s] # # /var/tmp/go/src/github.com/containers/podman/bin/podman images --all --format {{.Repository}}:{{.Tag}} {{.ID}}
[+0650s] # quay.io/libpod/testimage:20210427 aadc32e2a626
[+0650s] # # /var/tmp/go/src/github.com/containers/podman/bin/podman build -t infra_ho2waiwlxz -
[+0650s] # STEP 1: FROM quay.io/libpod/testimage:20210427
[+0650s] # STEP 2: RUN ln /home/podman/pause /pause_2Izc4dQiEZ
[+0650s] # open state file /proc/122860/stat: No such process
[+0650s] # open pidfd: No such process
[+0650s] # error running container: error reading container state from /usr/bin/crun (got output: ""): exit status 1
[+0650s] # Error: error building at STEP "RUN ln /home/podman/pause /pause_2Izc4dQiEZ": error while running runtime: exit status 1
[+0650s] # [ rc=125 (** EXPECTED 0 **) ]
[+0650s] # #/vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
[+0650s] # #| FAIL: exit code is 125; expected 0
[+0650s] # #\^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[+0650s] # # /var/tmp/go/src/github.com/containers/podman/bin/podman pod rm -f -a
[+0650s] # # /var/tmp/go/src/github.com/containers/podman/bin/podman rm -f -a
[+0650s] # # /var/tmp/go/src/github.com/containers/podman/bin/podman image list --format {{.ID}} {{.Repository}}
[+0650s] # aadc32e2a626 quay.io/libpod/testimage
[+0650s] # # [teardown]
[+0650s] # # /var/tmp/go/src/github.com/containers/podman/bin/podman pod rm --all --force
[+0650s] # # /var/tmp/go/src/github.com/containers/podman/bin/podman rm --all --force

[edsantiago]

@TomSweeneyRedHat yes, "open state file" is a flake, we hope it's fixed in containers/crun#661 but I assume it will take a VM-rebuild dance to get it fixed in CI.

[rhatdan]

/hold cancel