Bacula-users

Re: [Bacula-users] Bad disk noises - replace the volume without restarting the backup?

2010-11-04 04:55:37
Subject: Re: [Bacula-users] Bad disk noises - replace the volume without restarting the backup?
From: Mikael Fridh <frimik AT gmail DOT com>
To: bacula-users AT lists.sourceforge DOT net
Date: Thu, 4 Nov 2010 09:53:25 +0100
On Wed, Nov 3, 2010 at 6:34 PM, Mark Luntzel <mark AT luntzel DOT com> wrote:
> The answer is probably no but...
>
> I can hear the disk currently being written to making bad noises, and
> the speed is extremely slow. Bad disk for sure, about to fail. This is
> at the end of a multi-Terabyte backup, just about 500 gig left and I
> would REALLY hate to think there is no way out but to start over
> completely.
>
> So is there a way for me to replace that disk without invalidating the
> entire backup? Running on Linux with an external SATA / FW enclosure,
> bacula version 3.0.2

If there is a way to temporarily freeze a job (possible using signals)
to prevent it from writing anything while using LVM to attempt to
vgextend; pvmove; vgreduce; vgremove that physical disk could be a
viable solution.

Of course you would have to:
1. use LVM
2. Have the same amount of physical extents available somewhere else
in the system as the failing disk.

However, odds are some extents are unreadable and I'm not sure what
exactly happens when a pvmove fails to read an extent. Anyone
experienced with this particular case?

------------------------------------------------------------------------------
The Next 800 Companies to Lead America's Growth: New Video Whitepaper
David G. Thomson, author of the best-selling book "Blueprint to a 
Billion" shares his insights and actions to help propel your 
business during the next growth cycle. Listen Now!
http://p.sf.net/sfu/SAP-dev2dev
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users

<Prev in Thread] Current Thread [Next in Thread>