On Sun, 15 Oct 2006, Kern Sibbald wrote:
> On Sunday 15 October 2006 17:48, Michael Brennen wrote:
>>
>> I am starting over with a fresh regress checkout and current
>> bacula cvs. I will pay attention to every detail so as to try
>> and duplicate your test conditions. I may have to change the
>> mtx-changer script to enable the tape offline and inventory but
>> will only uncomment what is already there.
>
> OK. This sounds like a good approach. By the way, it is possible
> that the tape offline is causing some crazy timing or wierd state
> in the changer or in the communications between Bacula and the
> changer. It might be wise to look at it in detail too because
> here I have not enabled tape offline. I suppose I should try it,
> but when I get back -- at least got half of my packing done -- a
> real pain when you cannot take much carry-on.
It has been some months since I enabled the offline, but as I recall
at least the Q47 will not unload a tape that has not first been made
offline. That may be a difference between DDS4 and DLT drives.
A test just now on the Overland shows that it will unload a tape
without the offline. I tested that just by issuing mtx load and
unload commands.
The same test on the Q47 shows that it does require the mt offline.
An mtx unload alone will not unload a loaded tape.
The difference may be whether the changer itself issues an offline
if it thinks a tape is loaded, but adding the mt offline seems
harmless on the Overland though it is not strictly needed.
I have now run the 2drive-incremental-2tape test a couple of times
without error on the Overland. This was with my standard patch to
mtx-changer for the offline and wait_for_drive.
I have now also run the same test a couple of times on the Q47
without error, same patch to mtx-changer.
Drat... I hate needlessly taking up someone else's time.
I think the only substantive difference remaining between this set
and my previous tests is that I am using your DDS-4 configuration
unmodified. My drives are not DDS-4, but the DDS-4 label seems to
be strictly descriptive for humans, that is, Bacula does not
semantically understand that as meaning a certain type of tape.
I will work with this some more on my own to see if I can isolate a
difference in the device configurations that causes the failures.
If I find such I will post it.
At least we have a reasonable conclusion before you leave. :)
Thanks again, and enjoy your vacation in peace.
-- Michael
-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
