On Sat, Oct 9, 2010 at 8:21 AM, Gunnarsson, Gunnar
<[email protected]> wrote:
>>OK.  You should adjust the length in your tapetype definition, then - your 
>>tape appears to be at least 550% longer than you have configured.
>
> Can I add the tape length to:
>
> define changer vault {

No, it needs to be in the tapetype.  It's weird, I know, but there are
some difficult problems with combining those two sections.

>> Not sure what you mean by this - what was delayed?
> The holding disk was filled up before vtape was written.

That sounds suspiciously like the flush-* parameters are enabled.  If
not, can you send the corresponding amdump logfile to the list?

>>You probably should be excluding ./devices?
>
> This worked in verions 3.1.1 something is changed 3.2.0beta3 nothing else has 
> changed on the system. I testaed another Solaris client with this version and 
> got the same problem.

Hmm, I don't think we made any changes to how GNU Tar is invoked that
would cause it to no longer be able to read special device files - it
never could do that.  If you still have a sendbackup debug log from
3.1.x, can you compare the tar invocation to the one made by 3.2.0?

Dustin

-- 
Open Source Storage Engineer
http://www.zmanda.com

Reply via email to