Tracking down NFS problems can be tricky. There are several failure points in the NFS system, ranging from server misconfiguration to network difficulties.
The table below lists symptoms, causes, and resolutions for the most frequent NFS problems.
|NFS clients cannot mount your NFS server's exported filesystems
||NFS isn't running or misconfiguration in
/etc/rc.d/init.d/nfs status to check that NFS is running. If NFS is running, look through
/etc/exports for problems.
|Cannot modify a file on the NFS server
||Insufficient privileges, user or group mismatch, or stale lock file
||Make sure you have write access to the file and that user and group IDs on both the client and server match. If both of these are OK, you have a stale lock file created by a buggy NFS implementation. Obtain a patch from the vendor.
|Slow performance or NFS clients hang indefinitely
||Check that the connection between the client and server is up with
ping. If the connection is up, check latency with
traceroute. It is possible that an intermediate gateway is dropping your NFS packets.
|Random crashes of client or server
||Client and server incompatibilities
||Check the vendors' known incompatibility list for both the client and server. Upgrade as necessary.
For example, overloaded, mis-configured, or malfunctioning switches, firewalls, or networks may cause NFS requests to get dropped or mangled between the NFS Client and NFS Server.
Some specific instances have been:
- A damaged security appliance mangling packets between the NFS Client and NFS Server:
- The port-channel aka EtherChannel aka bonding configuration on the switch was incorrect:
- A second system on the network had duplicated the IP address of the NFS Server
- The switch was dropping TCP SYN,ACK packets:
- Issue was with a Riverbed WAN optimizer device