Amanda-Users

Re: unable to create

2007-05-31 17:51:36
Subject: Re: unable to create
From: Peter Kunst <peter.kunst AT swissrisk DOT com>
To: Brian Cuttler <brian AT wadsworth DOT org>
Date: Thu, 31 May 2007 23:48:35 +0200
Confirmed here, too. Silly me, it was a DLE here, that was previously configured to do an ufsdump in amanda's disklist, but it was a subdir of a partition actually.

Now we use gnutar, and this DLE is not going to do level 0's every day.

Thanks to all who put their hints in there.

Any further hints on howto check (and catch) this kind of problem at amanda runtime ? Maybe another check if a DLE to be (ufs)dumped is the root of a partition or not. If it's a DLE which is _not_ the root of a partion _and_ to be backed up with tar, simply go and do it ?

 Peter

On 05/30/2007 07:33 PM, Brian Cuttler wrote:
Unable to create files in /tmp or /var/tm

You guys where so right! Thank you, confirmed via testing
that this is certainly client side and in all cases because
my disklist file on the server was incorrect. Either pointing
to links that did resolve to partitions or pointing to directories
that where not mount points.

I was able to correct problems on 3 of my servers.

There is a 4th server with the same error, which I suspect is
due to the fact that we are backing up the device /dev/dsk/c0t0d0s0
rather than the mount point.

This was done because the FW admin wanted to backup the ABE, alternate
boot env as well as the active partitions and didn't want to have to
update the disklist file when alternating between boot environments.

We should be moving to a new FW platform "soon" and in the mean time
I will recommend to him that we sync the current boot env to the
alternate and stop backing it up. More than a few days should have
elapsed and we are apparently not returning the the alternate boot
and if we do we will want the current config, not an old one. Then we
can backup the valid mount points again.

The other outstanding issue, also on the same system, is one of the
spindles is too active (and slow?) and we fail backups. We need to
move to another model (backup offline mirror or snapshots) and/or
use faster CPU/HW, which will happen "soon".

The FW by the way is a Ultra 30 running Solaris 8.

Thank you very much, are backups are not only running more cleanly
(error-wise) and efficiently (not backing up the same partition a
second time) and effectively, since we are backing up other partitions
that we where missing before.

---
   Brian R Cuttler                 brian.cuttler AT wadsworth DOT org
   Computer Systems Support        (v) 518 486-1697
   Wadsworth Center                (f) 518 473-6384
   NYS Department of Health        Help Desk 518 473-0773



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