2

I am experimenting with mount namespaces. I have a python script, which creates a bind mount, after that a mount namespace with the unshare system call. I print the PID, and pause the script. I can list the mounts with:

nsenter --mount=/proc/PID/ns/mnt
mount

Let the python process be interrupted before calling umount (or let us say it terminates due to some error). Now /proc/PID disappears. Will the system be in a consistenst state? I mean will umount be called automatically, or will the mounts "cleaned up" some other way?

Improve this question

1 Answer 1

Reset to default
5

A mount namespace, like any other namespace, is removed once all its member processes exit (unless it is kept alive via a bind mount: see man 7 namespaces, search for “Bind mounting”), and when a mount point is removed, all its mounts are implicitly unmounted, see man 7 mount_namespaces:

A mount ceases to be a member of a peer group when … the mount is implicitly unmounted because a mount namespace is removed (because it has no more member processes).

Improve this answer

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.