On Mon, 2023-10-23 at 10:30 -0400, Greg Wooledge wrote: > On Mon, Oct 23, 2023 at 04:17:11PM +0200, hw wrote: > > I have an entry in the fstab to mount an NFS share via IPv6. For > > unknown reasons, the entry is being ignored on boot, so after booting, > > I have to log in as root and do a 'mount -a' which mounts the share > > without problems. > > Do your IPv4 NFS mounts have the same behavior?
No, I can mount the same share with the host name or the IPv4 address of the same server, and it works as it should. The IPv6 connection is set to be required at boot. > > [fd53::11]:/srv/example /home/example/foo nfs > > _netdev 0 0 > > On my systems, the options field contains "defaults,_netdev" and possibly > a few other things. I don't know whether that matters. I'll try --- I think it uses defaults anywhen when no defaults are specified. What else would it use other than defaults ... > If you add an entry to /etc/hosts for this NFS server, and mount the > share by hostname, does that change anything? I have a DNS server (on the same machine that serves the share) which works fine. Making an entry in /etc/hosts is something I want to avoid. If the addresses of the server ever change, I wont remember that entry and will have to spend ever so much time to figure out what's wrong ... > > I have another case in which machines need to be connected to a > > particular VLAN to mount home directories. In case they are not > > connected to that VLAN, I don't want the boot process to proceed at > > all because the home directories won't be available. > > Hmm... can you add the "fail" option after _netdev? I've never tried. Hm, in man mount, there's only 'nofail' for when the mount is allowed to fail. That also means that mounts must not fail and the failure must not be ignored. Same goes man fstab. I doubt there's a 'fail' option. Who would want a mount to fail? > > So how do I force it that the entries in fstab are not being silently > > ignored? I want these shares either mounted, like through like 3 > > retries, or booting to stop when they can't be mounted. > > I do not know how to allow a specific number of retries. Not without > writing your own hacked-up shell scripts, at least. The retries aren't exactly necessary; I merely thought it can't hurt to retry, and systemd should do that anyway when there's a failure the first time.
