On Wednesday 05 January 2005 13:35, Keith Matthews wrote:
>On Sat, 18 Dec 2004 10:01:41 +0000
>
>Keith Matthews <amanda AT frequentous.co DOT uk> wrote:
>> On Sat, 18 Dec 2004 09:17:30 +0000
>>
>> Keith Matthews <amanda AT frequentous.co DOT uk> wrote:
>> > In the light of messages just posted I'll report this in case it
>> > didn't get out. Apologies to those who got it first time, I
>> > don't like those who assume that no answer simply maans people
>> > don't want to answer either.
>> >
>> > I'm having trouble getting a remote backup to work.
>> >
>> > The report states that the disk backup failed due to a timeout.
>> > Examination of /var/log/messages at the client shows that
>> > amandad exited'status 1' but gives no other indication of the
>> > cause of the problem. This is happening with all disks on that
>> > client.
>
>OK, for the sake of posterity I'd better post some more for this.
>
>The problem seems to be related to the number of entries in the
> disklist for the relevant host.
>
>I originally had
>
><hostname> wd0a user-tar -1
><hostname> wd0e comp-user-tar -1
><hostname> wd0g comp-user-tar -1
><hostname> /var/amanda user-tar -1
><hostname> /var/backups user-tar -1
><hostname> /var/clamav user-tar -1
><hostname> /var/cron user-tar -1
><hostname> /var/mysql user-tar -1
><hostname> /var/named user-tar -1
><hostname> /var/spool comp-user-tar -1
><hostname> /var/www user-tar -1
>
>(I've replaced the real, fqdn, hostname for security reasons).
>
>After some considerable amount of cut-and-try testing I discovered
> that the system worked quite happily with just one disklist entry.
> Further testing revealed that it was quite happy with as many as 7
> entries, more would cause the first few to fail, and the whole set
> would cause the lot to fail.
>
>Replacing the above with one entry per filesystem (i.e wd0a, wd0e,
> wd0g) where the whole filesystem was needed, and the top level
> directory (/var) for the other case, with an exclude file to
> eliminate the unwanted had the whole set dumping correctly. I have
> no idea if this is a generic Amanda issue or one specific to the
> OpenBSD port.
>
>Debugging was complicated by the disk entries being tried in reverse
>order, something else that does not seem to be mentioned in the
>documentation.
>
>In case anyone wonders about the effect of 'inparrallel' I left it
> at the default of 4.
Did you perchance increase the etimeout and dtimeout values first?
I have had as high as 53 entries for a single client in my disklist
without any problems. As long as the spindle numbers used are
assigned to the individual drive per number, I've not even had any
disk thrashing problems either.
--
Cheers, Gene
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.31% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com attorneys please note, additions to this message
by Gene Heskett are:
Copyright 2004 by Maurice Eugene Heskett, all rights reserved.
|
