Tag Archives: nfs
I stumbled on a strange problem regarding NFS, which I hadn’t used for a while on my server. I had some old exports which worked last time I used them, but not any more. I was doing an rsync and suddenly saw messages about chgrp or chmod failing. These messages are expected when uid/gid is not matching between server and client, so what’s up?
Running dmesg showed me messages like these:
Looking in /var/log/syslog I saw messages like these:
I finally found a working solution here.
THIS IS NOT A BEST PRACTICE ON HOW TO INSTALL A LINUX BOX!
This document serves as a reminder to myself, on what needs to be done when reinstalling my workstation, and I have put it here to have it conveniently at hand. Because of this, the document may not be of any help to anybody but me. The last month I’ve had 3 failing harddrives, one new and 2 used ones, so this makes me want to document the process, at least in an overall fashion…
- I have a script listing all my script dependencies and all software I always want installed, so this part is easy:
I found a scribbled note regarding the matter and just had to post this as a reminder:
Mounting an NFS export from a Linux server in AIX will surely fail, unless you run this command before mounting:
After this you can proceed with regular mounting.
The explanations I’ve seen to why AIX doesn’t use reserved ports to begin with, vary from AIX follows NFS standard to It’s more secure this way. Whatever the reason behind this, it’s annoying when mounting fails without an abvious and reasonable explanation.
Solving disk problems is a little more complicated when devices are encrypted. You need to keep in mind that a filesystem on an encrypted device can’t be seen unless the device is decrypted. Here is a few problems I’ve come in touch with, together with solutions that have worked for me.
Error unmounting a volume
Error message which may occur when trying to unmount truecrypt volume:
Trying to remove the volume with cryptsetup gives a similar message:
Device /dev/mapper/truecrypt1 is busy.
As the error messages implies, something is keeping the volume open. If lsof doesn’t show anything, it might be an active NFS export keeping it busy.
I recently switched mainboard in one of my computers, and suddenly eth0 becomes eth1. This makes my /etc/network/interfaces point to the wrong interface, thus not giving me the correct IP address, which in return makes DNS fail, which in return makes NFS fail… Let’s just say that a plethora of problem arise!
Showmount gives me this error message when I try to list exports on the NFS server: