NAS/NFS issue

[Prashant Zanwar_4]

I have enterprise Linux setup with NetApps NAS storage. Also it has service guard setup. NAS is NFS mounted on both cluser nodes.
Thousand activities are going around, lots of NFS issues and Linux hangs keeping day always busy.
I am facing a issue with jar file extract.
It is on NAS storage, Linux is same on both nodes and also the jar executable. I copy file on NAS storage just at one server. Try to extract, that goes ok. When I try to extract on other node having same storage mounted I run through error.
Has any one faced it any time. Error would be the file is corrupt..like that..

Appreciate everybody's response..

Thanks in advance
Prashant

[Dave Falloon]

What nfs options are you using ie.

softmount
intr
rsize
wsize

Do you see any errors in the logs?

Check out the linux NFS faq as well:

http://nfs.sourceforge.net/

--Dave

[Steven E. Protter]

Prashant,

Nice to see you on the Linux side. Please post your Linux mount options from /etc/fstab.

I think a soft mount might be in order.

I need however to know what the options are on the shares from Linux as well /etc/exports

The parameters that work on HP-UX are not enough for smooth Linux operation.

Here is the /etc/exports on my succesful, home grown NAS server.

/share/web 192.168.0.40(rw,no_root_squash,async) 192.168.0.35(rw,no_root_squash,
async) 192.168.0.70(rw,no_root_squash,async) 192.168.0.41(rw,no_root_squash,asyn
c) 66.92.143.194(rw,no_root_squash,async) 66.82.153.195(rw,no_root_squash,async)
66.82.153.196(rw,no_root_squash,async)


Note: I am limiting access to certain nodes.

Thanks for posting your question, I need to authorize a new node in my setup.

SEP

[Prashant Zanwar_4]

I dont have access to servers, but FSTAB entries are as below..Am I missing anything here...Thx
Prashant

proto=tcp,ro,hard,intr,rsize=32768,wsize=32768,vers=3 0 0

[Prashant Zanwar_4]

I meant the filers..I dont have access to..I have access for Linux boxes..sorry for that..
Prashant

[rmueller58]

We have a Dell NAS, it occasionally looses itself.. I have a job croned to check the mount point, if the mount point isn't attached it attempts to mount.

nasmount=`mount |grep nas |awk '{print $3}'`
if [ "$nasmount" != "nas:/NAS" ]
then
echo "NAS is NOT MOUNTED ATTEMPTING MOUNT" > /nas.log
date +%Y%m%d:%H%M >>/nas.log
mount /spi/nas
else
echo "NAS is MOUNTED" >/nas.log
fi

Problem is the NAS is not under my direct control and our MS-Tech doesn't get in to the office till 8, when I check it, and it's down I notify him asap so it can be rebooted..
(typical MS or Apple) reboot the thing.

[Serviceguard for Linux]

I don't believe Serviceguard for Linux supports NFS as shared storage. At the very least do all packages have different NFS mounts?

[Prashant Zanwar_4]

Well, Now I get little more confined situation. This problem is with particulat file..And it happens on various server which are NAS storage connected..
Now it has got error like below, where moreof I am feeling it's some bug with jar/gcc..what could be solution...
But it is still strange as both systems have same NAS mounted and one server only gives problem/...please reply..
Thanks and its gr8 help of u guys
Prashant


[root@11i-tst-01 pz]# jar tvf XXCAC_FC.jar
Error in JAR file format. zip-style comment?

[Prashant Zanwar_4]

Well..Story becomes more interesting as I see inode of file is same, but cksum differs on this mount with this file...
Strange....
Did any one faced it with filers...will appreciate any type of help..

[Stuart Browne]

NFS is known for going bonkers periodically,I've witnessed this more times than I care to admit.

You've tried unmounting then remounting on the affected nodes and then re-trying the file?

I understand this might be somewhat difficult if it's in a busy cluseter environment (fail the node then do it).

The other possibility is that the NAS it's self is the one having the idfficulty. For some reason the inode opened, locked the file, and never released it. At a 'quiet' time, try checking the NAS to see what files are opened or locked, and see if this is one of them.

Another thing to look at is the 'nfslock' service on your server boxes. Whilst NFS supports various file locking types, not all that OS's can do are supported at the base level. You might want to try starting this service as well to see if reliability increases.

Just some thoughts, hopefully some will help.