Well, taking a look at what you have there, let's start with the /etc/exports
Now, does the UID 900 actually exist and is available on the Server A? If not, then change it to something that does exist on Server A that you consider "safe". That is what any new entries will appear to be from on Server A when accessed/written to across the mounted share, which with all_squash
is what everything will appear to be.
Also, an example of a couple additional entries in the /etc/exports
file that have helped me out in the past:
Now, if you are doing 'ro
', then I imagine you don't *need* sync, but you will likely get a warning message if you don't put something. The 'no_subtree_check
' will, at the very least, make the mounting of the shared directory faster since it won't require NFS to check the entire filesystem before completing the mount.
Now, on the other side, let's look at the /etc/fstab
entry of Server B:
172.25.130.109:/ARCHIVE /ARCHIVE nfs defaults 0 0
Well, I have never been one for blind "default" settings, so let me show you an example of one of the mounts of one of our servers:
192.168.2.242:/mnt/repository /mnt/mirror nfs rw,sync,bg,auto,intr,soft,retry=10
So, I have it that it is readable and writable, that changes must make it to disk before next request is serviced, that the mounting process is left in the background, that the mounting is done automatically when the filesystems in the fstab are processed on boot/startup, that the mounting process is interruptible (should there be an issue with the mounting), and that the mounting client will fail if there isn't success in the 10 minutes of retries.
Also, you don't need the filesystem check settings (e.g. "0 0") at the end of Server B's fstab entry as that should be handled by Server A to begin with.
HTH. Let us know.