NFS Read/write issues...

[jmckinzie]

I am once again having an issue reading and writing a specific nfs mount poiny.

here are some outputs:
netfmt -f /var/adm/nettl.LOG001


***********************************STREAMS/UX*******************************@#%
Timestamp : Mon Aug 07 EDT 2006 01:17:30.740668
Process ID : [ICS] Subsystem : STREAMS
User ID ( UID ) : -1 Log Class : ERROR
Device ID : 0 Path ID : 0
Connection ID : 0 Log Instance : 0
Location : 00123
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
10393 01:17:30 159518395 1 T.. 0 0 miip_reassemble_setup: Frag memory exceeded, 1 unknown disposition packet(s) dropped on interface: lan0

***********************************STREAMS/UX*******************************@#%
Timestamp : Mon Aug 07 EDT 2006 01:17:51.740198
Process ID : [ICS] Subsystem : STREAMS
User ID ( UID ) : -1 Log Class : ERROR
Device ID : 0 Path ID : 0
Connection ID : 0 Log Instance : 0
Location : 00123
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
10394 01:17:51 159520495 1 T.. 0 0 miip_reassemble_setup: Frag memory exceeded, 1 unknown disposition packet(s) dropped on interface: lan0


# lanadmin -x 0
Current Config = 100 Full-Duplex MANUAL

Any ideas on what the error listed above means would be great.

Thanks,

Jody

[Steven E. Protter]

Shalom Jody,

Have you looked aat syslog /var/adm/syslog/syslog.log

Is there anything else to help.

As far as I can tell without background, there is nothing in the error output to lead to the read write issue, which is not clearly defined or referred to in your thread.

Most NFS issues either deal with Network connectivity, as in the switch the wire, the configuraiton, the NIC card, or permissions.

If permissions is not an issue, run cstm or mstm or xstm diagnostics on the NIC card and see if there are a lot of problems on the switch concerning this host.

Also, make sure always you have the latest, greatest and best NFS patches on your system along with dependencies.

SEP

[Mel Burslan]

Jody,

which protocol are you using to access the NFS mountpoint ? UDP or TCP. If it is UDP, and if you can try TCP as it provides a little slower but more robust access.

Also, check to see if your network interfaces' firmware revisions are upto date.

[jmckinzie]

Syslog:

Aug 7 15:37:14 cfcap01d vmunix: NFS readdir+ failed for server nasfsa01.fsa.citicorp.com: RPC: Timed out
Aug 7 15:51:40 cfcap01d vmunix: NFS write failed for server nasfsa01-gss1: RPC: Timed out
Aug 7 15:51:55 cfcap01d vmunix: NFS readdir+ failed for server nasfsa01-gss1: RPC: Timed out
Aug 7 15:45:20 cfcap01d vmunix: NFS readdir+ failed for server nasfsa01.fsa.citicorp.com: RPC: Timed out

I am running on UDP....

Permissions do not appear to be the problem.

[Mel Burslan]

RPC timeout clauses on the syslog error lines making believe that you are having some sort of network disconnects during your NFS access attempts. In general, such disconnects can be attributed to one of the two things: either your network segment is heavily congested at times or a network interface at either end of the connection is dropping packets.

I have checked your previous posting abut the issue and you said you do not have control over the netapp, so I will not re-iterate my try TCP connectivity instead suggestion, but if you have any way to talk to the admins of the netapp box, try convincing them to turn on TCP connectivity, just for the testing if not for permanent, so that you can diagnose the problem. If that fails, ask your network administrator folks to place a sniffer to this network segment and try to find out what exactly is happening at this RPC timeout moments.

Other than that, I don't know what else to suggest.

[Dave Olker]

Hi Jody,

I just saw this post (I only check ITRC once ever couple of weeks when I get really busy at work). Are you still having this problem or have you figured out the root cause? If you're still having the problem let me know and I can try to help you troubleshoot it further.

Regards,

Dave

[jmckinzie]

Dave,

This problem has popped up once again.

Any ideas on how i can fix it?

Last time it was due to the NFS server having two NICs and i was trying to mount against the wrong one but, thats not the case now.

Any ideas would be greatly appreciated.

[Dave Olker]

Hi Jody,

Well, isn't this a blast from the past? Nearly a year between responses. :)

Let's refresh a bit here. What OS is the NFS client running? What OS is the NFS server running? What NFS patch levels are both of them running? On the NFS client issue the command "nfsstat -m" and post the output so I can see exactly how the filesystem is mounted.

Is this a problem that happens once a year or is it reproducible? If so, what steps are you using to reproduce the problem? If you can reproduce it at will I'd suggest starting a network trace and capturing one of these events in the trace. Hopefully the trace will provide some hints as to the reason for the error.

Regards,

Dave

[Reshma Malusare]

Hi Jody,

I think its a network connectivity issue...

First check whether servers are reachable..

then also check nfs status on both servers.

[Safarali]

check with rpcinfo if you can not mount possible reason is that mount daemon is not running on the server
you can use the option
-p -u (for udp)

Reggards
Safar

[jmckinzie]

I am successfully mounting the same export on other hp-ux hosts.

IE HP server located in the smae building is successfully mounting the export from the server.

However copying the exact mount references from the /etc/fstab still gets the errors.

I have checked networking to verify that the Lan card is mounting at 100FD.

Any other ideas?

-TIA

[Dave Olker · Accepted solution]

Hi Jody,

You haven't answered any of the questions in my previous post. Please answer those and also cut/paste the exact command you're using to reproduce the problem and the exact error message you're getting.

Dave

[jmckinzie]

I was monuting the filesystems via the wrong NIC.

It was mountinig using the controller NIC (on the filer) vice the lan NIC.