I have seen similer behavior between an old E30 and an H50 with massive
storage. No one, including IBM, has given me any explanation.
My workaround is to make it part of every planned outage to unmount all NFS
mounts on the remote machine before the other goes down. After the outage I
again mount the NFS mount, and the problem is avoided. Unless someone here
has a better answer, this is what I recommend.
I tried tuning parameters for NFS and the TCPIP stack, without maing the
problem any better or worse. I have not tested since AIX was upgraded to
5.1 ML 3 from 4.3.3 ML 10 recently, but I do not expect any change in
behavior.
LUCK!
Quote:> Hi all,
> I'm running a 4.3.3 F50 server that exports some file systems to a
> 4.3.3 S80 machine. Those machines are connected by a 2 way gigabit
> ethernet network. Yesterday I had to get the server down for
> maintenance and surprisingly the client machine stopped answering to
> its clients, i.e., the applications that runs on that machine hung.
> There is a known problem related to that situation ? It might be a
> coincidence and the problem with the applications was not really
> related with the shutdown of the nfs server ? I looked for related
> messages on both logs but I didn't find anything.
> Thanks in advance.
> Xyko
---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.445 / Virus Database: 250 - Release Date: 2003-01-21