Description

After checkpointing a container and attempting to restore it ctr throws an error:

Error: [ctr error](ctr: OCI runtime restore failed: cannot allocate tty if runc will detach without setting console socket: unknown)

Although the restore fails, the task is actually created and can be found using ctr tasks ls. The task is in a created state with a PID of 0. Attempting to delete the task also fails ctr task rm restoredContainer=> err: ERRO[0000] unable to delete restoredContainer error="task must be stopped before deletion: created: failed precondition" ctr: task must be stopped before deletion: created: failed precondition. It appears that this issue is separate from the restore failure but seems to be caused by it. #7357 also describes this bug (task created but cannot be deleted)

Steps to reproduce the issue

1. Start a container with a tty - ctr run -t --rm docker.io/library/busybox:latest testContainer
2. In another shell checkpoint the container - ctr c checkpoint --rw --task --image testContainer testCheckpoint
3. Now attempt to restore the container - ctr c restore --rw --live restoredContainer testCheckpoint
4. err => [ctr error](ctr: OCI runtime restore failed: cannot allocate tty if runc will detach without setting console socket: unknown)

Describe the results you received and expected

Restore should not fail. Either a terminal should be allocated to the new container or it should be run in a detached mode. Similar to how ctr run handles IO.

What version of containerd are you using?

containerd github.com/containerd/containerd v1.6.10 770bd01

Any other relevant information

$ runc --version

runc version 1.0.0
commit: 84113eef6fc27af1b01b3181f31bbaf708715301
spec: 1.0.2-dev
go: go1.15.14
libseccomp: 2.4.1

Show configuration if it is related to CRI plugin.

No response