Author: Chris Hoogendyk <hoogendyk AT bio.umass DOT edu>
Date: Mon, 28 Sep 2009 07:58:05 -0400
I've been happily using Amanda for some time now, but I seem to have run into a recent snag. The backup server has multiple addresses on a single interface and, for some reason, it has taken to using
Author: "Dustin J. Mitchell" <dustin AT zmanda DOT com>
Date: Mon, 28 Sep 2009 10:58:52 -0400
On Sun, Sep 27, 2009 at 10:36 PM, Jason Frisvold <friz AT godshell DOT com> wrote: No, at the moment there is no configuration for this -- Amanda does not bind its sockets to any particular IP, so th
I've been happily using Amanda for some time now, but I seem to have run into a recent snag. The backup server has multiple addresses on a single interface and, for some reason, it has taken to using
Author: "Dustin J. Mitchell" <dustin AT zmanda DOT com>
Date: Mon, 28 Sep 2009 10:52:02 -0400
Actually, I've been away on vacation, but thanks :) I think we're going to need to move in that direction. I want to have a look at the pthread implementation to see if I can understand in more detai
Author: Nathan Stratton Treadway <nathanst AT ontko DOT com>
Date: Mon, 28 Sep 2009 12:25:13 -0400
I haven't tried running Amanda on a host with two IP numbers on the same subnet, but another possibility is that the routing configuration on the backup server host have been changed somewhat during
Author: "Dustin J. Mitchell" <dustin AT zmanda DOT com>
Date: Mon, 28 Sep 2009 14:11:51 -0400
On Mon, Sep 28, 2009 at 1:48 PM, Dustin J. Mitchell <dustin AT zmanda DOT com> wrote: Apologies for replying to myself, but while putting together the patch (attached), I noticed that this will affec
Author: "Dustin J. Mitchell" <dustin AT zmanda DOT com>
Date: Mon, 28 Sep 2009 13:48:37 -0400
OK, I have a more in-depth summary of exactly what's going on here, and why the fcntl() calls fix it. The good news: we've stumbled on a pretty stable "fix" for this problem. As background, the Amand
Hi Dustin, Great analysis; thanks for sharing the details of the problem. On Mon, Sep 28, 2009 at 12:11 PM, Dustin J. Mitchell <dustin AT zmanda DOT com> wrote: I patched the 0928 snapshot, but it di
Author: Nathan Stratton Treadway <nathanst AT ontko DOT com>
Date: Mon, 28 Sep 2009 16:55:32 -0400
Well, as I understand it, the kernel uses the source IP that matches the outgoing route entry that it has selected for sending the packet in question. When the routes are choosing between interfaces
Author: Nathan Stratton Treadway <nathanst AT ontko DOT com>
Date: Mon, 28 Sep 2009 18:29:59 -0400
Congrats on tracking down all that! :) One followup, question, though: Early on in this thread (Aug 30 and 31), Michael reported that one of the early patches that Jean-Louis sent to him (or to Stan?
Author: Nathan Stratton Treadway <nathanst AT ontko DOT com>
Date: Mon, 28 Sep 2009 21:42:39 -0400
[...] [...] As you say, having this extra call shouldn't really hurt anything, but I wondering if it would make sense to tweak the #ifdef so that is possible to compile Amanda on OpenBSD without havi
Author: "Dustin J. Mitchell" <dustin AT zmanda DOT com>
Date: Tue, 29 Sep 2009 08:29:51 -0400
On Fri, Sep 25, 2009 at 4:41 PM, Gene Heskett <gene.heskett AT verizon DOT net> wrote: OK, revision 2148 should have fixed this, and should be in last night's snapshot. I'm sorry it took so long -- i
On Sep 28, 2009, at 12:25 PM, Nathan Stratton Treadway wrote: I haven't tried running Amanda on a host with two IP numbers on the same subnet, but another possibility is that the routing configuratio
On Sep 28, 2009, at 4:55 PM, Nathan Stratton Treadway wrote: Well, as I understand it, the kernel uses the source IP that matches the outgoing route entry that it has selected for sending the packet
Author: "Dustin J. Mitchell" <dustin AT zmanda DOT com>
Date: Tue, 29 Sep 2009 09:31:17 -0400
I wouldn't mind doing this, but it's only six fcntl() calls on startup, so the cost really is negligible. I don't know a good way to do this -- the best I can think of is to check GLIB_LDFLAGS for /p
Author: Gene Heskett <gene.heskett AT verizon DOT net>
Date: Tue, 29 Sep 2009 10:20:33 -0400
wrote: I'll take a look at it sometime later today, I just had a major disaster here, trying to setup a separate config to test backup up another machine, I found some of my startup helper scripts we
Author: Nathan Stratton Treadway <nathanst AT ontko DOT com>
Date: Tue, 29 Sep 2009 11:23:13 -0400
Well, I brought up this issue on the general principle that this sort of workaround hack might have its own unexpected side effects, and thus there all else being equal it seems like a good idea be a
I added the following lines to util.c as a temporary fix: I ran three test backups, and each succeeded: - forced full, without compression, no index - forced full, with compression and index - unforc
If this is a dual ether box, I've seen this issue with Red Hat / CentOS. You need to have only ONE default gateway on one of the interfaces. If you have two, it will alternate back and forth. On Sep