Hello Ben,
Great! Thanks for the feedback.
Good luck,
Kern
On 01/17/2014 06:34 PM, Roberts, Ben wrote:
>
> Hi Kern,
>
>
>
> I verified that the failures were happening on a 5.0.x FD as well as a
> 5.2.x FD. At the time, I hadn't realised this was unsupported or that
> it was even happening.
Hi Kern,
I verified that the failures were happening on a 5.0.x FD as well as a 5.2.x
FD. At the time, I hadn't realised this was unsupported or that it was even
happening. Following Martin's observation earlier that the corruption was
happening conveniently closely to the 2^32 overflow boundar
Hmm. Very clever of you Martin to notice that. Sounds like
a strong possibility ...
On 01/17/2014 01:01 PM, Martin Simmons wrote:
> Are the failing block numbers always a little below 2^32 (like 4294944994 and
> 4294941825 in your messages)? If so, that maybe suggests a compiler bug if
> the sam
Hello,
Every case of this particular error message that I have seen has been
due to data corruption outside of Bacula. Typically this happens when
a disk drive is bad, but since you are running ZFS and its checksums
are good, I can see only several other possibilities:
1. The ZFS code is messed
Hello Steven,
You might take a looks that the patch that Ulrich Leodolter
sent to the bacula-devel list on 1/7/2014 at 2:14pm. It may
or may not do what you want. I am currently looking at that
patch with an eye to integrate it into Bacula, but I don't fully
understand it yet so I am not sure it
> On Fri, 17 Jan 2014 12:14:20 +, Roberts, Ben said:
>
> > Are the failing block numbers always a little below 2^32 (like 4294944994
> > and
> > 4294941825 in your messages)? If so, that maybe suggests a compiler
> > bug if the same source code works when compiled on the older machine (o
> Are the failing block numbers always a little below 2^32 (like 4294944994 and
> 4294941825 in your messages)? If so, that maybe suggests a compiler
> bug if the same source code works when compiled on the older machine (or is it
> the same binary too?).
That is a very good spot, I hadn't picked
17.01.2014 10:46, Vladimir Skubriev ?:
Hi!
I put thing's in order and accidentally deleted mysql database(current
running) with bacula catalog and mysql service database.
But I have backups of bacula catalog database with mysqldump.
But this mysql dump in bacula volumes (
BaculaCatalog000
Are the failing block numbers always a little below 2^32 (like 4294944994 and
4294941825 in your messages)? If so, that maybe suggests a compiler bug if
the same source code works when compiled on the older machine (or is it the
same binary too?).
__Martin
> On Thu, 16 Jan 2014 16:30:59 +00
