AIX

We recently upgraded 2 of our AIX 6.1 servers from TL4 to TL5.
Both servers are on the same p7 780 frame, installed at the same time from the same image. Both servers are mounting the same nfs share after reboot what worked perfectly fine until the upgrade.
Since the patching, one of the two boxes became unable to see files in the share bigger 2GB - the other system doesnt have this problem.
Workaround we currently use is to drop nfs and have it recreated. After the recreation and remount of the filesystems all files are visible / accessible - until we boot, than the bigger 2 gig files become unavailable again.

Has anyone seen a similar behavior and know how to fix it permanently?

Regards
zxmaus

What do you mean by "unable to see files large than 2GB"?

the nfs mount contains dump files from another server. Some of these dumpfiles are 10 GB, others are 800 mb. While I have no issues to use the small ones on this box, I get some weird 'cannot access file xy - file or directory does not exist' while I do a simple ls. Doing the same on the other box shows all files. I guess this has something to do with the very very old nfs 2 limitation of 2 gig but as far as I am aware of, AIX 6.1 doesnt even support nfs2 any more - and I have never heard about a limitation like this in nfs3 ?

Regards
zxmaus

Can you verify there are no JFS errors logged on the source server? Also if possible take the file system offline and run a fsck if you have not done already. Do you see these errors when running the same commands on the local file system of the NFS server?

I dont see the errors local - and I dont see them either on the other server mounting the same share - that makes it so weird

Seems after dropping / recreating nfs it works fine until files in the shared filesystem are updated / new dumps written - than the problem starts again - but only on the one server, not on the other one - so I'd exclude the nfs server

Regards
zxmaus

and the mount options are exactly the same?

maybe you have different mount options on one of the servers because of non printable characters or spaces in one of the /etc/filesystems files?

check the options directly using the mount command

Code:

mount | grep filesystem

nfs version 2 has a 2 gb limit afaik


also I would check

Code:

lslpp -l bos.net.nfs.client

check the version, and also the cksum of each file, and compare the two systems
maybe one file has not been replaces with the update


also check the nfs services:

Code:

 lssrc -g nfs
Subsystem         Group            PID          Status 
 biod             nfs              3342544      active
 nfsd             nfs              4194350      active
 rpc.mountd       nfs              8912918      active
 rpc.statd        nfs              9961520      active
 rpc.lockd        nfs              10027058     active
 nfsrgyd          nfs                           inoperative
 gssd             nfs                           inoperative