vmunix: NFS fsstat failed for server : RPC: Timed out

[Tyler Booker]

We have been getting sparatic NFS errors showing up in the syslog on our Unix db servers "vmunix: NFS fsstat failed for server : RPC: Timed out". We also see NFS lookup and readdir errors. We are running HPUX 11.23 with CIFS client vs A02.02.02. The mounts are connectiong to a Windows 2003 file cluster and we see zero errors on the Windows servers. The network activity seems to be rather normal when this happens so we don't believe we are over loading the network. Has anyone run into this issue with the versions I listed above?

[Dennis Handly]

>we see zero errors on the Windows servers.

Obviously the windows servers are never going to blame themselves for RPC timeouts. :-) Same thing for NFS servers.

[sujit kumar singh]

that is OK Dennis, Can we ask Tyler to have the latest patches for NFS and CIFSclient to be installed on the HP-UX system?

Tyler can u install the same and check for that.????

[Tyler Booker]

So are you saying there are documented issues with the A02.02.02 CIFS version?

[TTr]

> NFS errors showing up in the syslog
>"vmunix: NFS fsstat failed for server : RPC: Timed out
> We are running HPUX 11.23 with CIFS client vs A02.02.02

So what do the NFS errors have to do with the CIFS client setup? If you are having NFS errors you have to describe your NFS setup. Who is the NFS client (aparently your HP-UX 11.23 server) and who is the NFS server? Is it possible you have set up both services (NFS, CIFS) against the same windows share?

[Tyler Booker]

Well I think that is part of the question. We don't know if the NFS errors have anything to do with CIFS. The filehare mounts are CIFS mounts, but we get the timeouts on the database server syslogs. We show noting in the network captures that we have on the file servers. You would think you would see the timeouts on the response if it was the NFS server with the problem. The db servers are in a cluster with service guard which is dumping tons of broadcasts onto the network as well. Nothing that I have read so far points to any of this but CIFS or NFS timeout tuning.

[Dennis Handly]

>but we get the timeouts on the database server syslogs.

Nothing that points to a server, PID or a filesystem that is having the problem? Any timestamps, particular time of day?
Does this server use NFS for anything?

[Tyler Booker]

We were hoping to see a pattern in the NFS errors which would let us know that there is a high load on the network during certain times of the day but the errors are random. Nfs is used and we have pulled stats from the db server to show that NFS errors are happening but also shows UDP errors happing as well: