The problem really has nothing to do with NFS.  If a read times out the stale
NFS handle message is used.  I am assuming that the developer assumed that a
local filesystem would never time out (hmmm... does the rule about assuming
apply when I assume re: someone else's assumption?)  In my shop I have had to
set the nfstimeout value to 120, but now I don't get aborted backups that
complain about NFS.




 (Embedded
 image moved   Jim Hicks <jdhicks AT IBM DOT NET>
 to file:      05/27/99 03:21 PM
 pic21793.pcx)




Please respond to "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>

To:   ADSM-L AT VM.MARIST DOT EDU
cc:    (bcc: Chuck Mattern/IS/SSC/THD)
Subject:  Re: Stale NFS handle




Jeff,
Did you get anywhere with the 'stale nfs mount' message?  I also am getting
this error on "/" filesystem.  The backup will abort and skip this
filesystem but continues on with other filesystems etc. to completion.
ADSM client  3.1.20.6   AIX 4.3.1.0 on SP.

I tried Eric's suggestion of NFSTIMEOUT 10  to no avail.  There is no NFS
mounts on the node I'm using....

Jeffrey Fulmer wrote:

> Thanks,
>
> I really should have posted this somewhere else.  This is an AIX problem
> and not related to ADSM.  I ran "fuser -k" on the filesystem and ADSM
> backed it up that evening.  But I can't figure out why I am getting a
> stale handle on a system that is NOT even running NFS.
>
> Cheers,
> Jeff
>
> Eric Tang wrote:
>
> > You can try adding undocumented parm in dsm.sys    NFSTIMEOUT     10
> >
> > Regards,
> > Eric Tang