Containers (spec) not getting cleaned up after stopping the jobs.

[shishir-a412ed]

Seeing this intermittently (not easy to reproduce at this point), when I launch multiple nomad jobs

e.g.

$ nomad job run redis.nomad
$ nomad job run capabilities.nomad
$ nomad job run privileged.nomad
$ nomad job run stress.nomad

After the containers (tasks) are running. Stop the containers (tasks).

$ nomad job stop redis
$ nomad job stop capabilities
$ nomad job stop privileged
$ nomad job stop stress

This stops all the tasks. However, containers (spec) are left behind and not cleaned up properly. This was working before, and there could likely be a regression.

root@linux:/opt/gopath/src/github.com/nomad-driver-containerd/example# ctr task ls
TASK    PID    STATUS
root@linux:/opt/gopath/src/github.com/nomad-driver-containerd/example# ctr container ls
CONTAINER                     IMAGE                             RUNTIME
32522383_lucid_noyce5         docker.io/library/ubuntu:16.04    io.containerd.runc.v2
63c86855_agitated_cray6       docker.io/library/ubuntu:16.04    io.containerd.runc.v2
ba8f121f_ecstatic_volhard3    docker.io/library/ubuntu:16.04    io.containerd.runc.v2
bcddf7ad_jolly_bartik7        docker.io/library/redis:alpine    io.containerd.runc.v2
fb07ad0a_focused_thompson9    docker.io/library/redis:alpine    io.containerd.runc.v2

[shishir-a412ed]

Might be related to this