Bacula-users

Re: [Bacula-users] Backup of many small files

2013-11-05 10:28:41
Subject: Re: [Bacula-users] Backup of many small files
From: Josh Fisher <jfisher AT pvct DOT com>
To: bacula-users AT lists.sourceforge DOT net
Date: Tue, 05 Nov 2013 10:24:47 -0500
On 11/4/2013 6:55 PM, Dimitri Maziuk wrote:
> On 11/04/2013 04:17 PM, Phil Stracchino wrote:
> ...  In at least one of the
>> cases I know about, though, the problem was not a failure of DRBD per
>> se, it was that someone accidentally started up mysqld on the second
>> node, which normally would not be allowed to happen because the second
>> instance accessing the same filesystem would not be allowed to lock the
>> InnoDB tablespace.  In the DRBD situation, though, there was nothing
>> whatsoever locking the tablespace on the local machine, so it quite
>> cheerfully allowed both mysqlds to write to their own local instances of
>> the tablespace on the replicated filesystem, whereupon the two mysqlds
>> proceeded to destroy the tablespace by making it internally inconsistent.
> Ah, the "active-active" setup. I agree: avoid like plague.
>
> The simple stupid (and bulletproof) setup is an active/passive cluster
> with drbd, mysql, and floating ip. Unfortunately if it fails over your
> bacula-dir is going to lost its connection to the db and kill the
> running backup -- see my post from 2 days ago.

And I think that will always be the case in a fail-over situation as 
long as Bacula requires the persistent network connection and fails jobs 
without attempting a reconnect. I think that Open vSwitch could allow 
live migration of DIR and SD without the connection loss, but not in a 
fail-over situation where the hardware has dropped the connection. That 
is why I said that Bacula is not entirely cluster friendly. That is not 
a deal breaker, since much of the tape hardware SD can use as backup 
devices require manually plugging the device into another node anyway. I 
have been running DIR and SD in a kvm VM on a 2-node cluster for a 
couple of years now without issue, other than it is not fail-over 
capable. I can run the VM on either node, though it requires manually 
migrating the VM. It would be nice if Bacula were fail-over capable, but 
still it works just as well running in a VM on a cluster as on bare metal.


------------------------------------------------------------------------------
November Webinars for C, C++, Fortran Developers
Accelerate application performance with scalable programming models. Explore
techniques for threading, error checking, porting, and tuning. Get the most 
from the latest Intel processors and coprocessors. See abstracts and register
http://pubads.g.doubleclick.net/gampad/clk?id=60136231&iu=/4140/ostg.clktrk
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users