Amanda-Users

Re: automount and raw devices

2007-12-20 10:58:06
Subject: Re: automount and raw devices
From: Jeffrey Anderson <JDAnderson AT lbl DOT gov>
To: "Dustin J. Mitchell" <dustin AT zmanda DOT com>
Date: Thu, 20 Dec 2007 07:48:35 -0800
On Thursday 20 December 2007, "Dustin J. Mitchell" <dustin AT zmanda DOT com> 
wrote:
> On Dec 20, 2007 7:39 AM, Tim Bunnell <bunnell AT asel.udel DOT edu> wrote:
> >  > Dec 19 23:45:06 helios automount[4754]: failed to mount /vol/rsma
> >  > Dec 19 23:45:07 helios automount[4755]: failed to mount /vol/rhumgen
> >  > Dec 19 23:45:07 helios automount[4756]: failed to mount /vol/rconfocal
>
> Any idea why the automounts themselves are failing?
>
> To rule out Amanda, take a look at the sendbackup and sendsize debug
> logs for lines containing "Spawning".  These will show you the 'tar'
> invocations which, I expect, will not contain the 'r'.
>
> Dustin

I experienced this same situation a couple of weeks ago.  It suddenly started, 
persisted for a couple weeks, then, just when I was about to ask the list, it 
stopped!

The backups always proceeded as expected and there were no bad sideffects, but 
everytime amcheck or amdump ran there was an automount attempt to access 
these non-existent directories.

For every DLE  named /home/SOMEUSER on the client amcheck (and amdump) 
apparently also triggered an automount attempt for /home/rSOMEUSER.  In my 
case, and I think in Tim's as well, the automounts fail because these 'r' 
directories simply don't exist.

It only happened on one client.  It started happening without amanda being 
updated on the client or the server.  I didn't pursue it further because it 
never caused a problem and it stopped on its own, but it was very strange.


-- 
--------------------------------------------------------------
jeffrey anderson                        | jdanderson AT lbl DOT gov
Lawrence Berkeley National Laboratory   | Mailstop 50a-5101
Phone: 510 486-4208                     | Fax: 510 486-6808

<Prev in Thread] Current Thread [Next in Thread>