On Wednesday 11 April 2007 22:13, Ralf Gross wrote: > Kern Sibbald schrieb: > > On Wednesday 11 April 2007 20:28, Ralf Gross wrote: > > > I'm also confused about this order: > > > > > > 1. run before job > > > 2. start job 121 > > > 3. recycle volume 06D124L > > > 4. mark volume 06D124L as used > > > 5. ---> write job #1 to volume 06D124 > > > > > > 08-Apr 00:05 VU0EM005: BeforeJob: run command "/usr/bin/ssh [snip]" > > > 08-Apr 00:07 VU0EM005: Start Backup JobId 121, > > Job=SMTCZB0003.2007-04-08_00.05.00 > > > 08-Apr 00:08 VU0EM005: 3307 Issuing autochanger "unload slot 4, drive 0" > > command. > > > 08-Apr 00:09 VU0EM005: 3304 Issuing autochanger "load slot 5, drive 0" > > command. > > > 08-Apr 00:10 VU0EM005: 3305 Autochanger "load slot 5, drive 0", status is > > OK. > > > 08-Apr 00:10 VU0EM005: 3301 Issuing autochanger "loaded? drive 0" command. > > > 08-Apr 00:10 VU0EM005: 3302 Autochanger "loaded? drive 0", result is Slot 5. > > > 08-Apr 00:10 VU0EM005: Recycled volume "06D124L3" on device "LTO3" > > (/dev/nst0), all previous data lost. > > > 08-Apr 00:10 VU0EM005: Max configured use duration exceeded. Marking > > Volume "06D124L3" as Used. > > > 08-Apr 00:10 VU0EM005: Spooling data ... > > > [snip] > > > 08-Apr 00:12 VU0EM005: Job write elapsed time = 00:02:23, Transfer rate = > > 6.766 M bytes/second > > > 08-Apr 00:12 VU0EM005: Committing spooled data to Volume "06D124L3". > > Despooling 968,636,754 bytes ... > > > > > As I said (or rather implied) perhaps when the volume was recycled the use > > duration expired "before" the First Write Time was properly updated, or maybe > > it wasn't even updated. > > > > If you can specifically show that to be the case, then this would be a bug. > > How would I do this, any hints? I'm running the director with -d 200 > at the moment. Is this enough debug level? Or could I lower the debug > level? BTW, is there a way to add a timestamp to the debug output? > > I 'hope' the same problem will appear again next weekend when the next > volume will be recycled (06D125L3). > > Pool: Differential (trimmed output) > +-------+----------+---------+--------------+------------+-------+-------------------+ > |mediaid|volumename|volstatus|volbytes |volretention|recycle| lastwritten | > +-------+----------+---------+--------------+------------+-------+-------------------+ > | 5 |06D124L3 |Used | 968,583,168| 2,678,400 | 1|2007-04-08 00:12:44| > | 6 |06D125L3 |Used |17,191,802,880| 2,678,400 | 1|2007-03-11 00:33:24| > | 7 |06D126L3 |Used | 6,299,725,824| 2,678,400 | 1|2007-03-19 21:07:48| > | 8 |06D127L3 |Used | 9,633,254,400| 2,678,400 | 1|2007-03-25 00:22:52| > | 19 |06D137L3 |Append | 3,497,001,984| 2,678,400 | 1|2007-04-08 00:18:03| > +-------+----------+---------+--------------+------------+-------+-------------------+ > > > Anything I can check before the next diff. backup on sunday? Any > special debug commands? > > > > > I've used the 'Volume Use Duration' parameter in bacula 1.38.5 for > > > about one year. It always worked as (at least I) expected. > > > > Agreed, but three things: > > > > - Code is in constant change, and sometimes in changing it new bugs are > > introduced. > > > > - Version 2.0.x checks more carefully and more often for expired volume > > parameters than previous versions. > > > > - Maybe you just fell into a new untested case during the recycle. Code can > > work "correctly" for years, then something different happens, and bang things > > go wrong. > > > Of course! If it's a bug I would like to help fix it. If it's a > missconfiguration, I would like to find the error.
The way I would proceed if I had the time to work on this would be to setup a "dummy" backup, either on a separate machine, or with a different Job name on your production machine. I would then write a volume, then recycle it (purge, ...), and at each step do a "llist volume=xxx" and see what happens. You could even set a pretty short volume use duration and see what happens when 1. that use duration is normally exceeded, and 2. what happens just after a recycle and the use duration is exceeded from the previous firstwritten. That would show the evolution of the volume, and might reproduce the "bug". ------------------------------------------------------------------------- Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys-and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV _______________________________________________ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
