I’m trying to troubleshoot an NFSv4 mounting issue on Rocky 9.4.
My mount command exits with status 0 and without printing any error message, but the filesystem is not successfully mounted. In particular, it is not listed by the df command, and the mount point directory is empty (whereas the exported directory is not). On the other hand, the mount appears in /proc/mounts, and if I attempt to repeat the same mount command in verbose mode, the last message it emits is “mount.nfs: mount(2): Device or resource busy”. That appears to be associated with the mount point directory, not the remote filesystem, because I can try again (once) without getting that message if I use a different mount point.
Those “Device or resource busy” conditions persist as long as I’ve been willing to wait, at least tens of minutes, even though I’m performing a foreground mount, and I’ve set options that I would expect to make the mount attempt fail relatively quickly (and the mount command does terminate quickly). For example (as root):
mount -v -o rw,fg,exec,nosuid,nodev,vers=4,retry=1,timeo=50 my.server.org:/remote/filesystem /mnt/tmp
I don’t see anything relevant in the system log.
Also, I can successfully mount a different filesystem exported by the same server. If I’m grasping at straws, it may be relevant that krb5p security is being negotiated (both for the mount that fails and for the one that succeeds), which is as I expected.
What are some plausible explanations for this problem?
What options do I have for further troubleshooting?