[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Bacula-devel] Possible autochanger logic problem in 2.2.10-b1SD

On Wednesday 30 April 2008 21:33:51 David Boyes wrote:
> > Yes, though if the Volume was loaded in a drive at SD start time, it
> is
> > possible that the SD did not recognize it as loaded. For tapes, it
> will
> > check
> > what is in each drive, but for a "disk" it doesn't expect any
> > Volume "mounted", so probably does not check.
> >
> > It this is what happened to you, I would recommend some procedure that
> > unloads
> > all the drives when starting the SD.
> Might be useful to add a SHARED=y/n parameter to the autochanger
> definition that would control whether the controlling SD emptied all
> defined drives at startup -- then you'd always have a defined state at
> initialization. If SHARED=Y, then take the current behavior (since the
> SD that is starting doesn't and can't know what some other SD might be
> doing), otherwise empty all defined drives at SD startup. Default to

Bacula more or less has a way to do what you are suggesting already with "Auto 

The problem I was bringing up (not necessarily his problem) is that if you are 
using a real autochanger, the SD checks what is in each drive, so there is no 
need to unload them.  However, if the drive is a disk (virtual autochanger as 
he is using), then the normal logic is that a disk is never loaded so I am 
not convinced that the SD actually checks what is in the drive.  This is just 
a supposition on my part because I haven't actually looked at the code.


This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
Don't miss this year's exciting event. There's still time to save $100. 
Use priority code J8TL2D2. 
Bacula-devel mailing list

This mailing list archive is a service of Copilotco.