I've got an NFS server which is happily serving out files to 10 NFS clients (and has been doing so for 18 months+). I recently noticed that one of the clients has stopped working correctly, NFS mounts perfectly fine, but doing any operation on the mount itself results in an endless hang, and frozen terminal.

I've searched around for several hours online and have checked pretty much everything I can, but some info:

  • All clients (and server) are the same Dell 2950 machines running Centos 5.6/5.7
  • All clients are connected to the same local network
  • Network connection on both problem client and server is the same (1000mb/s Full Duplex), with no connection/latency/packet loss issues between them.
  • Server responds to rpcinfo, showmount -e and nfsstat straight away without issue
  • nfs, portmap are installed on the problem client (like all the other clients) and started
  • Mount works perfectly, returning instantaneously
  • I've tried mounting using a variety of mount options (intr, soft, nolock)
  • Reversing the situation (setting up server on the client, and mounting from the server), this works fine!

Can anyone help me out as to something else I could try to resolve this?

  • What errors show up if you look at the client-inteface with ethtool -S eth0 (replace eth0 with your interface)? You should also start concurrent tcpdump-sessions on a working and on the hanging client. Do you see the same output regarding brodcast and multicast traffic? – Nils Sep 21 '11 at 19:58

can you post a snippit from something like...

$ cd /broken/nfs/mount

$ strace touch foo

Just curious what syscall it's hanging at.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.