Kern Sibbald wrote:
On Sunday 04 December 2005 18:40, Jesper Krogh wrote:

How can this happen?

Clone_System.2005.12-03_08.22.33 Fatal error: Device /dev/nst0 is busy
writing on another volume.

I have a quite well working setup using 1.36-3 and 25 GB of spool disk
setup. Should Bacula not just wait until the device was done writing to
the volume before trying?


There is a bug report on this error. However, I have been unable to reproduce it. If you are able to distill the problem down to something simple and reproducible, a second bug report would probably be worth while.

It is reproducible here so I'll try to dig up some debugging information.

I've searched the bugtracker, but I cannot find the original bug report.



There is another unrelated feature request.

The spool seems to work ineffectively. Due to network bandwith etc. etc. I would like to run 2-3 concurrent jobs spooling to diske and then interleaving them to tape. (Please correct me if this is not one of the usual use-case for disk-spools ?)

In the situation where 2 concurrent jobs are started with the same speeds (both larger than $spool_size/2) then both build up until the sum of the spools reaches the $spool_size and then both starts requesting to spool to tape. (in the mean time they dont spool to disk, since the disk-spool is full at least until one of the jobs have finished spooling to tape).

I have a disk spool of 25GB, it would be quite more effecient to slice the spool into spool-chunks then:
first job spolling to chunk1, chunk3, chunk5, etc. etc.
second job spooling to chunk2, chunk4, chunk6, etc. etc.

The first that fails to get a chunk would starte spooling the chunk-set to tapes, deleting the chunks as it goes on, thereby freeing up disk-spool and letting the other job continue spooling to disk while the first job spools to tape.
(A round-robin-style disk-spool)

As it is now, both clients wait for spooling for the full time that the the first spooled block takes to write to disk.

There is probably just some feature in the documentation i've overlooked?

Jesper
--
Jesper Krogh, [EMAIL PROTECTED], JabberID: [EMAIL PROTECTED]



-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users

Reply via email to