Hello,

On 6/29/2006 6:59 PM, Timo Neuvonen wrote:
> A couple of days ago I was writing about tape labeling problems.
> At that time it looked to me, that the problem was related to using my new
> autochanger. Now, after some more debugging, I think the problem exists also
> with a single drive, but the key factor is that the tape must be totally
> empty (eg. brand new). As a workaround, writing an eof to the beginning of
> the tape fixes the problem.

That can always work, because it enables the tape drive to much faster 
recognize the tape format. I'm not sure that this is documented, though.


> 
> Originally, I was trying to label brand new tapes in autochanger:
> 
> *label barcodes
> <clip>
> 3912 Failed to label Volume: ERR=dev.c:678 Rewind error on "VXA320"
> (/dev/nst0). ERR=Input/output error.
> This took about 5 minutes per tape, most of the time the drive just keeps
> displaying "ready". No tape access during the last 4 minutes...

Hmhm... have you checked if mtx-changer uses the function that checks 
that the drive is ready?

> If I do the following (for each tape)
> # mt -f /dev/nst0 rewind
> # mt -f /dev/nst0 weof
> 
> and then run "label barcodes" again, labeling works without any errors, with
> one tape it takes about 1 minute (~30 s to load a tape etc). So, this takes
> much less than the timeout of 5 minutes above.

I know that behaviour from DLT hardware: With an unformated tape, the 
initial tape recognition can take very long.

Your tape drive, I assume, doesn't check the tape format on loading but 
only at the first read access. In that case, you're lucky, because you 
have a very good way to prepare tapes faster... of course, without 
having a quick way to ensure that tape is really empty (in case you 
didn't unwrap it just before).

> I tried to check the documentation, if it says that the tape should start
> with an eof before labeling, but I couldn't find any note like that. Of
> course, if there are no other hidden problems in labeling, the easiest way
> would be to change documentation to require the eof.

Well, it's not required, but it avoids at least one misleading warning 
message. Any problems, I think, beyond that should be handled between 
mtx-changer and the hardware.

> 
> Finally, I re-installed my old Exabyte VXA-2 drive, that I'd been using for
> a couple of years:
> 
> *label
> <clip>
> block.c:944 Read error at file:blk 0:0 on device "VXA2" (/dev/nst0).
> ERR=Input/output error.
> 3000 OK label. Volume=Testxx Device="VXA2" (/dev/nst0)
> Catalog record for Volume "Testxx", Slot 0  successfully created.
> Requesting to mount VXA2 ...
> 3001 Device "VXA2" (/dev/nst0) is mounted with Volume "Testxx"
> *
> 
> So, it looks now that a similar problem appears with the old drive too, but
> now the tape gets labeled anyway.

No problem, but just the above mentioned misleading warning message.

> It's been at least half a year since I've
> used it to label new tapes either, so could it be possible that the newest
> Bacula releases expect more what they should find from the tape before
> labeling it?

Not as far as I know.

 > Both of the drives above were now attached to 1.38.10-3, on FC4
> (same system, just a conf change). It may have been as old as 1.36.3 that
> I've been using last time, when labeling ran ok, with no errors at all (as
> far as I can remember).
> 
> Yes, this may be something Exabyte-spesific problem, but at least I think
> labeling used to work with the same hardware some time ago.
> 
> Ps.
> I've also tried labeling with 1 tape at a time in the autochanger, using
> "label" instead of "label barcodes". It then works in the similar way like
> in the single drive: error message, but tape gets labeled anyway.

Well, one more reason to suspect you should tweak your mtx-changer 
program to ensure the tape drive has finished loading the tape before it 
returns.

Arno

> 
> Regards,
> Timo
> 
> 
> 
> 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
> 

-- 
IT-Service Lehmann                    [EMAIL PROTECTED]
Arno Lehmann                  http://www.its-lehmann.de

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

Reply via email to