Author: Brian Cuttler <brian AT wadsworth DOT org>
Date: Mon, 21 Sep 2009 15:59:32 -0400
I thought we would get by when we saw this error for a single partition, but its now preventing the client from amchecking at all. Server Amanda 2.6.1, Solaris 10/x86 Client amanda 2.6.1 Solaris 10/S
Author: Brian Cuttler <brian AT wadsworth DOT org>
Date: Mon, 21 Sep 2009 16:13:54 -0400
Fearing the number of DLE for the single client, I commented out 6 of them. Amcheck ran fine. I restored 2, leaving 4 commented out, amcheck ran fine. I restored 2 more, leaving 2 commented out, amch
Author: Brian Cuttler <brian AT wadsworth DOT org>
Date: Tue, 22 Sep 2009 09:42:29 -0400
Using amanda 2.6.1 on Solaris 10 (server x86, client Sparc) I currently have entries in the disklist like these, currently there are 55+ entries. finsen /export/home/agaupel zfs-snapshot finsen /expo
You might want to include a catch-all, something like below: finsen /export/home-others /export/home { zfs-snapshot include "./[!a-z]*" } that you expect to be empty (or not :) but which will get the
Author: Brian Cuttler <brian AT wadsworth DOT org>
Date: Tue, 22 Sep 2009 15:23:51 -0400
Jon, So it will work ? I was afraid of what would happen if I tried to snapshot the directories that way, that are all unique mount points, not just subdirectories of /export/home. The grep is meant
Hopefully someone with more experience with zfs snapshots can tell you. It would certainly work without snapshots. I'm concerned that you might still be backing up the mounted, active filesystems rat
We are in the process of making a number of significant changes to our internal network. At the moment, an amcheck on 52 machines, which ultimately reports no problems, is taking over 6 minutes. Hist
Author: Charles Curley <charlescurley AT charlescurley DOT com>
Date: Wed, 23 Sep 2009 08:16:14 -0600
I am running amanda-server-1:2.5.2p1-1--i386 on Ubuntu 8.04, Hardy Heron, and backing up to vtapes. For about a week, I have gotten the following error messages in my reports: dzur.localdomain /home
Author: Brian Cuttler <brian AT wadsworth DOT org>
Date: Wed, 23 Sep 2009 14:48:28 -0400
I have a very old version of amanda 2.4.4, backing up a new version, 2.6.1 across Solaris client and Solaris server. The client has ZFS partitions, /export /export/home which are actually ZFS mount p
Author: "Roger Williams" <R.Williams AT gns.cri DOT nz>
Date: Thu, 24 Sep 2009 11:55:37 +1200
Having just added lots of disk space to a network of Linux workstations and servers, I decided to take the plunge and check out Amanda as a replacement for a set of home-grown dump scripts that I hav
Author: Chris Hoogendyk <hoogendyk AT bio.umass DOT edu>
Date: Wed, 23 Sep 2009 21:18:43 -0400
Roger Williams wrote: Having just added lots of disk space to a network of Linux workstations and servers, I decided to take the plunge and check out Amanda as a replacement for a set of home-grown d
Author: Matt Webster <amanda-forum AT backupcentral DOT com>
Date: Wed, 23 Sep 2009 11:55:01 -0400
It appears the problem is the Superloader/SCSI interface is sending a SG_ERR_CAT_RESET. I have just put a work around into mtx scsi command to wait for 5 mins if receives this status, and then ignore
Author: Jason L Tibbitts III <tibbs AT math.uh DOT edu>
Date: Wed, 23 Sep 2009 21:45:39 -0500
Yes, I've seen the same. I've no idea what's gone wrong, but switching to tar seems to be the best solution. Unfortunately the idea of writing to the filesystem in order to back it up seems... dumb
Author: "Dustin J. Mitchell" <dustin AT zmanda DOT com>
Date: Thu, 24 Sep 2009 08:55:26 -0400
On Wed, Sep 23, 2009 at 2:48 PM, Brian Cuttler <brian AT wadsworth DOT org> wrote: We've seen problems with --one-file-system in certain versions of gnutar on solaris. What are you using? Dustin -- O
Author: "Dustin J. Mitchell" <dustin AT zmanda DOT com>
Date: Thu, 24 Sep 2009 09:29:12 -0400
On Mon, Sep 21, 2009 at 12:50 PM, Brian Cuttler <brian AT wadsworth DOT org> wrote: Are you using BSD or BSDUDP? Those can both run into limits on the size of a UDP packet when you get too many DLEs.
Author: Brian Cuttler <brian AT wadsworth DOT org>
Date: Thu, 24 Sep 2009 09:37:20 -0400
/usr/sfw/bin/bin/gtar is GNU tar 1.17, but it seems to work find for /export/home, just not for export, which times out... Its just a couple of zones in an ldom... 2.4.4 server didn't have estimate s
Author: "Dustin J. Mitchell" <dustin AT zmanda DOT com>
Date: Thu, 24 Sep 2009 09:36:10 -0400
On Mon, Sep 21, 2009 at 3:59 PM, Brian Cuttler <brian AT wadsworth DOT org> wrote: Yep, that's a lot of bytes to fit in a UDP datagram! Dustin -- Open Source Storage Engineer http://www.zmanda.com
Author: Brian Cuttler <brian AT wadsworth DOT org>
Date: Thu, 24 Sep 2009 09:35:14 -0400
I believe it defaults to BSD authentication, if your asking for the protocal for the daemon activation from inetd.conf, I think amandad listens to the udp port. I think I'm going to implement the inc
Author: "Dustin J. Mitchell" <dustin AT zmanda DOT com>
Date: Thu, 24 Sep 2009 09:33:14 -0400
On Mon, Sep 21, 2009 at 12:58 PM, Brian Cuttler <brian AT wadsworth DOT org> wrote: It does give that same error, but it gives it in the amreport rather than failing at the amdump level (thereby gene
Author: "Dustin J. Mitchell" <dustin AT zmanda DOT com>
Date: Thu, 24 Sep 2009 09:27:03 -0400
Tar should be running via the setuid script, runtar, so it should have root privs to back up anything. The client-side amcheck checks do *not* run with these privileges, so the top-level directory fo