Hi!

I have a Raspbian Raspberry Pi 4 nfs server running with the following exports:

showmount -e 192.168.1.YY lists
I have a 4 disk harddisk case with currently 3 disks inside since one broke and i am waiting for the replacement. On my 192.168.1.XX Linux desktop i mount the disks in fstab via

The idea is to have hd and D01 as read-only disks for my media library available on the whole network and to have all exports available rw on my desktop 192.168.1.XX. It works except one strange phenomenon: On my desktop hd,D01,D02 are mounted correctly, but D03 is simply a remount of D01. How could this have happened? As i mentioned D04 is currently out of my 4-disk-case-JBOD. On my 192.168.1.YY nfs server the drives are mounted completely correctly. On my desktop both the terminal and the GNOME file explorer show the wrong assigment, so i guess something went wrong while mounting the nfs share or while advertising and exporting it in the network.

I do not understand fully, but I think you have just mistyped something, somewhere...

I don't think so. I have cross-checked. The drives are actually mounted in a wrong way over the network.

How can it happen? You may need to mount them one by one (and by hand) to see what's happening. And I'm sure you will find something suspicious (sooner or later).

I checked /etc/mtab on my NFS client desktop and indeed for D03 it lists two mounts in a row, the D01 remount being the more bottom one and the correct D03 mount being before it. So it seems the correct /etc/fstab mount is somehow overridden. But why i do not understand. The only place that i ever mounted the disks were /etc/fstab, though once one disk broke and i placed another one into the slot instead. On the machine where the disks are attached the mounts are completely correct. So i do not understand why the client mounts are mingled up now.

for me 2 lines in mtab means 2 mounts. You might want to comment out it in /etc/fstab and check what's happening at the next boot. It looks like automounter is running (or ??)

Hi! I commented out in /etc/fstab the entry for D03 and after a reboot /etc/mtab had neither the correct mount line nor the other one with D01 mounted to /mnt/D03. So commenting out the mount line in fstab of D03 removed both entries from mtab after a reboot. Then i uncommented the line again and after a new reboot, again both mount lines for D03 were there again.

If i do a manual umount of D03 both lines vanish. If i do a manual mount of D03 ("mount /mnt/D03" or "mount 192.168.1.YY:/mnt/D03 /mnt/D03") only the wrong mount appears (i.e. D01 gets mounted to /mnt/D03). I guess that the data structures of the mount procedure got messed up in a completely intransparent way or maybe something is wrong on the server side?

I think i have an idea about the problem. Once i switched disks after one broke in my disc tower. Maybe the network mount client has somehow stored the old partition id somewhere and that is why it is remounting D01 and also doing an overridden mount of D03. That must be then an error on the client side.

I found some post from 2012 in another forum where someone had the same problem https://forums.opensuse.org/showthre...-folder/page2?. Reading that i think that there the problem did not get solved either. At the end someone reported that some deletion of a file on the server-side or some export option "nohide" solved the problem. I tried the nohide option and it did not change anything. I also tried to apt purge both the client side nfs-common package and the server-side nfs-kernel-server. I even purged remaining files in /var/lib/nfs and looked in other files of the installed package for information. None of that could solve the wrong filesystem mount. Since the problem was already there in 2012, maybe it resurfaced again? The person in the post had a similar problem also related to multiple NFS volumes mounted.