NFS-Problem between HP-UX and AIX

[Oliver Stoklossa]

Hello everybody ...

I got a strange problem:

Background: I've got an AIX-workstation, which offers a NFS-mount via automountd and some HP-UX 11.0 workstation to mount this share (as well as some additional IBMs)

Problem: On a specific HP machine I always get an error message when trying to do a "ls" in the NSF-mounted directory (/dataus) (NFS server catia7 not responding still trying). All the other machines work well, but the configuration on all machines is the same.
The other thing I can't explain (and I guess that's my problem) is the following output from nfsstat -m:

IBM-machine:
/dataus from /dataus:catia7
Flags: vers=3,proto=tcp,auth=unix,hard,intr,link,symlink,rsize=32768,wsize=32768,retrans=5
All: srtt=0 (0ms), dev=0 (0ms), cur=0 (0ms)

HP-machine:
/tmp_mnt/dataus from catia7:/dataus (Addr 10.1.55.7)
Flags: vers=2,proto=udp,auth=unix,hard,intr,dynamic,devs,rsize=8192,wsize=8192,retrans=5
Lookups: srtt= 10 ( 25ms), dev= 7 ( 35ms), cur= 4 ( 80ms)
Reads: srtt= 3 ( 7ms), dev= 3 ( 15ms), cur= 1 ( 20ms)
All: srtt= 12 ( 30ms), dev= 7 ( 35ms), cur= 5 (100ms)

So why does the HP use NFS version 2 instead of using V3 (On all other HPs the version is 2 as well, but there it works)

Any clues highly appreciated

Cheers
Oli

[kish_1]

I had simillar issue, then added following patch's
also compare this two file
/usr/newconfig/etc/rc.config.d/nfsconf
/etc/rc.config.d/nfsconf


PHCO_23117 bdf(1M) cumulative patch
PHCO_23651 fsck_vxfs(1M) cumulative patch
PHCO_23876 cumulative SAM/ObAM patch
PHCO_23963 libc cumulative header file patch
PHCO_25707 libc cumulative patch
PHKL_18543 PM/VM/UFS/async/scsi/io/DMAPI/JFS/perf patch
PHKL_20016 2nd CPU not recognized in G70/H70/I70
PHKL_22677 fix of getdirentries, MVFS, rcp, mmap & IDS
PHKL_24027 VxFS 3.1 cumulative patch
PHKL_25613 IDS/9000; file/socket syscalls; eventports
PHKL_26059 syscall, signal, umask cumulative patch
PHKL_26800 Probe,IDDS,PM,VM,PA-8700,asyncio,T600,Hang
PHNE_25116 Cumulative STREAMS Patch
PHNE_25626 ONC/NFS General Release/Performance Patch
PHNE_26250 LAN product cumulative patch
PHNE_26771 cumulative ARPA Transport patch

[Oliver Stoklossa]

Hello Babu ...

I just applied the "missing" patches to my systems and checked the nsfconf - files.
But the problem/output persists.
I can't explain why it works on all machines except one and why the protocol is udp version 2 on the HPs and tcp version 3 on AIX ...

Any other tips ?

[Oliver Stoklossa]

Some additional info:

I just mounted the /dataus - directory manually under /tmp/dataus and got the following:

/tmp/dataus from catia7:/dataus (Addr 10.1.55.7)
Flags: vers=3,proto=udp,auth=unix,hard,intr,link,symlink,devs,rsize=8192,wsize=8192,retrans=5
Lookups: srtt= 13 ( 32ms), dev= 8 ( 40ms), cur= 5 (100ms)
All: srtt= 13 ( 32ms), dev= 8 ( 40ms), cur= 5 (100ms)

The directory still isn't accessible via "ls" but the version changed from V2 to V3, so I guess the protocol udp is the issue...

Is there a problem with the automount, for manually mounting results in the correct NFS version ...

Running out of clue

Oli

[Oliver Stoklossa]

And again something strange:

When manually mounting as described above I get udp as default.
Trying to mount with

mount -o proto=udp catia7:/dataus /tmp/dataus

the NFS mount is built, but inaccessible.

Doing
mount -o proto=tcp catia7:/dataus /tmp/dataus

I get the error

nfs mount: set_args: invalid value for option: "proto=tcp"

[Oliver Stoklossa]

THE PROBLEM IS SOLVED!!!

I finally got it and the sollution is strange ... I found that the entries for nfsd in /etc/services differed from that one machine...

Thanks again babu, 'cause you led me to the right direction (the tip with nfsconf was good, for the entry NFS_TCP=1 led me to the sollution)