Real container state lost across reboot

[cjeanner]

Is this a BUG REPORT or FEATURE REQUEST?:

[//]: # Uncomment only one, leave it on its own line:

kind bug

[//]: # kind feature

Description

Steps to reproduce the issue:

1. Run a container that fails starting properly. It should be listed with an "Exited (non-0)" in podman ps --all

2. Reboot the computer/server

Describe the results you received:
After the reboot, podman ps --all shows the container as "Created"

Describe the results you expected:
It should keep the "Exited (error-code)" instead, in order to stay consistent.

Additional information you deem important (e.g. issue happens only occasionally):

Output of podman version:

podman version 0.9.4-dev

Output of podman info:

host:
  BuildahVersion: 1.4-dev
  Conmon:
    package: podman-0.9.3.1-3.gita3c4ce6.el7.x86_64
    path: /usr/libexec/podman/conmon
    version: 'conmon version 1.11.4, commit: 88d5276957861dea88b85f1d3f786612042cb4dd-dirty'
  Distribution:
    distribution: '"centos"'
    version: "7"
  MemFree: 8404660224
  MemTotal: 15871639552
  OCIRuntime:
    package: runc-1.0.0-54.dev.git2abd837.el7.x86_64
    path: /usr/bin/runc
    version: 'runc version spec: 1.0.0'
  SwapFree: 0
  SwapTotal: 0
  arch: amd64
  cpus: 6
  hostname: undercloud.lain.internux.ch
  kernel: 3.10.0-862.14.4.el7.x86_64
  os: linux
  uptime: 10m 3.02s
insecure registries:
  registries:
  - 192.168.24.1:8787
  - 192.168.24.3:8787
registries:
  registries:
  - docker.io
  - registry.fedoraproject.org
  - quay.io
  - registry.access.redhat.com
  - registry.centos.org
store:
  ContainerStore:
    number: 90
  GraphDriverName: overlay
  GraphOptions:
  - overlay.override_kernel_check=true
  GraphRoot: /var/lib/containers/storage
  GraphStatus:
    Backing Filesystem: xfs
    Native Overlay Diff: "true"
    Supports d_type: "true"
  ImageStore:
    number: 37
  RunRoot: /var/run/containers/storage

Additional environment details (AWS, VirtualBox, physical, etc.):
Running on a libvirt domain - although this shouldn't affect the state reported by podman.

[TomSweeneyRedHat]

@cjeanner thanks for the problem report. Do you by chance have a Dockerfile or a recommended way to create a container that fails to start?

[cjeanner]

Hello @TomasTomecek
I guess a command as "exit 1" in the Dockerfile should do the trick. I can confirm that in about an hour if you want - I have some running tests I can't really stop right now, and a bit out of resources :/

Cheers,
C.

[mheon]

This should be pretty easy now that we have a dedicated Exited state. Instead of unconditionally setting the Created state, we set it only if we are not in Exited state. Stopped also probably transitions to Exited on refresh.

…

On Tue, Oct 23, 2018, 07:55 Cédric Jeanneret ***@***.***> wrote: Hello @TomasTomecek <https://github.com/TomasTomecek> I guess a command as "exit 1" in the Dockerfile should do the trick. I can confirm that in about an hour if you want - I have some running tests I can't really stop right now, and a bit out of resources :/ Cheers, C. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#1703 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AHYHCK2PAiolQZXEKdDqrWC_jdJn3Mc-ks5unwO0gaJpZM4X1GkX> .

[rhatdan]

@mheon Has this been fixed?

[mheon]

Nope. I'll take it, should be a quick fix.

[cjeanner]

@mheon I guess we can close this issue now since you merged #2273 ?

[mheon]

Yes. Good call.