-
Hi. I have an active single-node swarm with 10 services on my Debian system. The docker engine runs containerd standalone and is started via systemd service ( |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 2 replies
-
Existing containers will continue to run, which is handled by the containerd shim (which acts as parent process for the container), but creating and deleting containers won't work while containerd is down; you can try this on a test machine; # create a container that prints the date every 2 seconds
docker run -d --rm --name mycontainer alpine sh -c 'while true; do date; sleep 2; done'
# restart containerd, and check the container logs
systemctl restart containerd
docker logs mycontainer
Mon Sep 11 11:28:10 UTC 2023
Mon Sep 11 11:28:12 UTC 2023
Mon Sep 11 11:28:14 UTC 2023
# stop containerd and check that it's still running
systemctl stop containerd
docker logs mycontainer
Mon Sep 11 11:28:10 UTC 2023
Mon Sep 11 11:28:12 UTC 2023
Mon Sep 11 11:28:14 UTC 2023
Mon Sep 11 11:28:16 UTC 2023
Mon Sep 11 11:28:18 UTC 2023
Mon Sep 11 11:28:20 UTC 2023
# but creating or starting a new container will fail
docker run -it --rm alpine
docker: Error response from daemon: connection error: desc = "transport: Error while dialing dial unix:///run/containerd/containerd.sock: timeout": unavailable. I should mention that the "remove containers" while containerd is not running situation could use some improvement (in general you'd not want containerd to be stopped for longer periods of time). I'll open a ticket for that to see if we can make that slightly better. |
Beta Was this translation helpful? Give feedback.
Existing containers will continue to run, which is handled by the containerd shim (which acts as parent process for the container), but creating and deleting containers won't work while containerd is down; you can try this on a test machine;