The two comments are back-to-back in both man mount and man nfs.
I still say that from what she is saying, she wants Soft mounts,
since they don't require user intervention; a hard mount, can't
be killed without killing the process trying to use it; even with
an intr, you still break the process.  I've had a single machine
error cascade to killing a network of systems, due to a hard
mount.  The same error with a soft, and I MIGHT have a minor problem
on some of the machines (depending on what they were doing), but
those machines are still up and ready for work.

Bill Ward

-----Original Message-----
From: Dan Stromberg [mailto:[EMAIL PROTECTED]]
Sent: Thursday, April 26, 2001 7:51 PM
To: [EMAIL PROTECTED]
Subject: Re: NFS questions


Some say you should never use soft mounts if you value your data.

mounting intr would proably suffice.

You could also try umount -f.

On Wed, Apr 25, 2001 at 02:30:44PM -0400, Ward William E DLDN wrote:
> Softmounts
> 
> man nfs
> 
> -----Original Message-----
> From: Ashley M. Kirchner [mailto:[EMAIL PROTECTED]]
> Sent: Wednesday, April 25, 2001 12:45 PM
> To: Red Hat Mailing List
> Subject: NFS questions
> 
> 
> 
>     Past experience has taught me that if an NFS server happens to go
> down for whatever reason (and then restarted), the machines connected to
> it will be in a state of 'being stuck' and the only way to fix that is
> to either try to unmount those NFS volumes, and remount them, or to
> restart the client (which solves the problem all together).
> 
>     Is there any better mechanism to check for things like this
> automatically?  If the NFS server gets rebooted, or the connection died
> long enough for the clients to lose it.  What can be done?  Is there any
> specific way I should set the clients up to reconnect or something?
> 
>     For that matter, if and when the NFS server goes down, is there any
> way of telling the clients to stop trying to use that particular mounted
> volume, and switch to a local (mirrored) partition till the NFS server
> comes back up?
> 
>     AMK4



_______________________________________________
Redhat-list mailing list
[EMAIL PROTECTED]
https://listman.redhat.com/mailman/listinfo/redhat-list

Reply via email to