I am running a mix of RHEL 6.9 and 7.7 hosts. These hosts are all NFS clients.
I am having a shutdown problem related to unmounting of NFS shares that were set up in /etc/fstab (i.e. not automounted). Specifically, if an NFS share is mounted and the NFS server subsequently becomes unavailable, shutdown of the NFS client machine hangs for about 30 minutes while trying to unmount the NFS share.
Google has much to say about this, but nothing that has really fixed my problem.
A manual fix is to manually perform a lazy unmount (umount -l) of each NFS share before issuing the shutdown command. However, my users' privileges do not allow them to do this. (We do allow them to shut the system down, though). Furthermore, I would not expect my users to do this. They should just be able to shut their machine down.
For various reasons I need not elaborate on, automounting the NFS share in question is not an option. Suffice it to say, this option has been considered. Setting the mount up in /etc/fstab is necessary in our use case.
I do understand the potential for data corruption. In our use case, we really want our users to be able to shut down.
I've considered creating a script for my users to use to perform a shutdown. That script would first do a lazy unmount of all NFS shares and then issue the actual shutdown command.
Before going this route, I'd like to check with community experts who know more than I do...
Is the proposed script the best route to solve this, or does a better solution exist?