Author: "Dustin J. Mitchell" <dustin AT zmanda DOT com>
Date: Mon, 19 Oct 2009 10:57:16 -0400
On Mon, Oct 19, 2009 at 10:34 AM, Gene Heskett <gene.heskett AT verizon DOT The whitespace is significant in the command I showed: if [ `/bin/ls /dumps` != "" ] btw, the shell strips leading and trai
Author: Gene Heskett <gene.heskett AT verizon DOT net>
Date: Mon, 19 Oct 2009 11:14:03 -0400
Bingo! Its running normally with the original syntax. And I would be willing to bet that many of the conditionals in that script which have the same basic syntax had to be surrounded by [[ and ]] to
Author: "Dustin J. Mitchell" <dustin AT zmanda DOT com>
Date: Mon, 19 Oct 2009 10:08:23 -0400
On Mon, Oct 19, 2009 at 8:05 AM, Gene Heskett <gene.heskett AT verizon DOT net> wrote: This can be a bit tricky in portable shell, but maybe this is closer to what you want: if [ `/bin/ls /dumps` !=
Author: Gene Heskett <gene.heskett AT verizon DOT net>
Date: Mon, 19 Oct 2009 11:20:40 -0400
wrote: Except that this time you use dbl-quotes. I used singles as you did in the first reply, and it worked, with the added white space around the [ and ] as Paul pointed out. It may be that due to
Author: Nathan Stratton Treadway <nathanst AT ontko DOT com>
Date: Mon, 19 Oct 2009 11:40:39 -0400
However, if /dumps is empty, the above will return a parse error (at least in bash) since `/bin/ls /dumps` will be blank and so no string will be found on the left side of the != after command subsit
Author: Brian Cuttler <brian AT wadsworth DOT org>
Date: Tue, 20 Oct 2009 11:00:39 -0400
Jean-Louis, I reconfigured one of my globs to use dumptype zfs-snapshot rather than user-tar. Backups ran well but I believe that we did a single snapshot of /export/home rather than a snapshot of /e
Author: Jean-Louis Martineau <martineau AT zmanda DOT com>
Date: Tue, 20 Oct 2009 11:21:55 -0400
Brian, What is the path of the snapshot, post the Amzfs_snapshot.20091019*.debug files. Jean-Louis I reconfigured one of my globs to use dumptype zfs-snapshot rather than user-tar. Backups ran well b
Author: Jean-Louis Martineau <martineau AT zmanda DOT com>
Date: Tue, 20 Oct 2009 11:50:34 -0400
It create one snapshot for each DLE. Can you post the disklist file so I can interpret the debug file correctly. Jean-Louis From directory (on the client) /tmp/amanda/client [finsen/tmp/amanda/client
Author: Brian Cuttler <brian AT wadsworth DOT org>
Date: Tue, 20 Oct 2009 11:58:41 -0400
Jean-Louis, Yes, however I'm afraid that if it creates a single disklist for a glob, for instance /export/home/a*, creates snapshot /export/home which I think would be acceptable. But 'attempts' to c
Author: Alan Griffiths <ap_griffiths AT hotmail DOT com>
Date: Tue, 20 Oct 2009 16:58:03 +0100
Thanks for the pointer. I've re-compiled with define STREAM_BUFSIZE (NETWORK_BLOCK_BYTES * 16) And I can see that has taken effect on the server dumper: try_socksize: send buffer size is 524288 But b
Author: Jean-Louis Martineau <martineau AT zmanda DOT com>
Date: Tue, 20 Oct 2009 12:05:58 -0400
That's what I wanted to see, For the disklist: finsen /export/home-A /export/home it create a 'amanda-_export_home-current' snapshot, this is bogus, the snapshot should be 'amanda-_export_home-A-curr
hello everyone, I've prepared a few examples of amreport xml output, filtered through xmllint for readability's sake. These were prepared using the test cases found in installcheck/Amanda_Report.pl
Author: Alan Griffiths <ap_griffiths AT hotmail DOT com>
Date: Tue, 20 Oct 2009 18:42:55 +0100
In the amandad log from 2.5.0p2 I see the following amandad: try_socksize: send buffer size is 65536 amandad: try_socksize: receive buffer size is 65536 amandad: time 0.006: bind_portrange2: trying p
Author: "Dustin J. Mitchell" <dustin AT zmanda DOT com>
Date: Tue, 20 Oct 2009 13:48:47 -0400
Ah, my "old" logs don't stretch back to 2.5.0 :) Did a tcpdump demonstrate that the buffer sizes were actually in effect? Dustin -- Open Source Storage Engineer http://www.zmanda.com
Author: "Dustin J. Mitchell" <dustin AT zmanda DOT com>
Date: Tue, 20 Oct 2009 12:43:24 -0400
Hmm, I can't find that indication in any of my older logs. From what I can tell, because amandad just uses the handles (x)inetd gives it, it never sents the buffer size. From my read, there is a bit
Author: "Stefan G. Weichinger" <sgw AT amanda DOT org>
Date: Tue, 20 Oct 2009 23:50:57 +0200
Paul Mantz schrieb: Sorry if I missed something ... Could you please explain the context of this? I appreciate the direction but I would like to get the picture ... Thanks, Stefan
On Tue, Oct 20, 2009 at 4:50 PM, Stefan G. Weichinger <sgw AT amanda DOT org> wrote: Hi Stefan, I'd be glad to explain. I'm working on a general update and rewrite of amreport in Perl. One of the com
Author: "Dustin J. Mitchell" <dustin AT zmanda DOT com>
Date: Tue, 20 Oct 2009 22:45:29 -0400
I'm curious whether anyone is post-processing either trace logs or amreport output to generate statistics, and if so, if this XML report gives you the information you need. Dustin -- Open Source Stor
Author: Jean-Louis Martineau <martineau AT zmanda DOT com>
Date: Wed, 21 Oct 2009 06:59:01 -0400
Use the amgtar application and set the TAR-BLOCKSIZE to a bigger value (half the STREAM_BUFSIZE. Why do you believe it is bottleneck? Jean-Louis I've re-compiled with define STREAM_BUFSIZE (NETWORK_B
Author: Brian Cuttler <brian AT wadsworth DOT org>
Date: Wed, 21 Oct 2009 15:29:15 -0400
We hit the problem of the TCP packet being too large with amanda 2.6.1p1, though this threat indicates a fix at 2.5.2. http://forums.zmanda.com/showthread.php?t=1005 Is the problem corrected or does