Comment 13 for bug 1078926

Revision history for this message
Steve Langasek (vorlon) wrote :

Seems the problem is that / and /run have both been mounted in the first pass, and mountall is stuck waiting (in try_mounts()) for the 'mounted MOUNTPOINT=/' event to return before it considers the 'mounted MOUNTPOINT=/run' event, which has already triggered. Since the former *depends* on the latter as a result of cloud-init and virtual-filesystems, we need to not block while waiting for it.