ADSM-L

Re: Stale NFS Handle on HP10.20 during backup

1998-09-23 14:54:34
Subject: Re: Stale NFS Handle on HP10.20 during backup
From: Brandon Moore <brandon.moore.hoq0 AT STATEFARM DOT COM>
Date: Wed, 23 Sep 1998 13:54:34 -0500
Question 3:
During scheduled backup, files in an NFS mounted filesystem
(/a/D/GROUPS/... on Windows NT) are being backed up in filespace /, even
though the DOMAIN specifies only ALL-LOCAL.

If the filesystems are mounted on the root partition(/), then what you
stated about the NFS mounted filesystem is true.  If you have seperate
partitions for
/,/usr./tmp,  the ADSM client recognizes the ufs or jfs or vxfs partitions
created when you built HP-UX O/S not mount points as filespaces that you
could recieve from q fi nodename from
the administrative client.

If you do not want to backup /a/D/GROUPS, you want to issue  presched
command to unmount the articular file system...That is if this machine is
on a schedule...

When I have seen stale NFS mount points, it has been when information from
the mounted filesystem on a client is requested and the source(server) is
not available.  Usually I have had to stop and restart
nfs(rpc.statd,rpc.lockd.rpc.mountd, etc.-depending on what machine, client
or server).  You may want to check out your nfs timeouts for mount and
extend them a little or least to be equal with your commtimeoute for ADSM.

I have in the past timeouts for NFS to fifteen minutes and ADSM to ten
minutes but hey UNIX sysadmins do different things...

FWIW

Brandon W. Moore
Princeton Information, Ltd
<Prev in Thread] Current Thread [Next in Thread>