Description

When a container is created with live-restore it is supposed to remain active even when the containerd daemon is stopped. But when the containerd is restarted with a container running in background with live-restore, the container is no longer active in the running process.

Steps to reproduce the issue

1. mkdir /tmp/html
2. Ensure /etc/docker/daemon.json has live-restore: true
3. echo "Hello World" > /tmp/html/index.html
4. Run docker run --name nginx-test -v /tmp/html:/var/www/html/:ro -d --network host gcr.io/google-containers/nginx
5. Check for the process with docker ps
6. Restart containerd process with systemctl stop containerd and systemctl start containerd
7. Check for the process with docker ps now. The container would be no longer active.

Describe the results you received and expected

docker ps

Expected:
Where it should show the

CONTAINER ID   IMAGE                            COMMAND   CREATED        STATUS        PORTS     NAMES
61cb4e0e5e9c   gcr.io/google-containers/nginx   "nginx"   18 hours ago   Up 18 hours             fb4edff8

What version of containerd are you using?

1.6.0_rc.3

Any other relevant information

runc version 1.1.0

Linux 5.10.98+ #1 SMP Tue Feb 8 18:42:19 UTC 2022 x86_64 Intel(R) Xeon(R) CPU @ 2.20GHz GenuineIntel GNU/Linux

Show configuration if it is related to CRI plugin.

No response