Dump problems

[Ricardo Oliva]

Hi,

I am currently using Dump / Restore set of scripts for backups over SSH 
and I am running into something very annoying.

All machines (tape host and clients) are Freebsd 4.9-p11 and the back 
ups run fine for most of the time. Every now and just one of the 
machines has this one partition that dump will crash with a:

Dumping machine example, filesystem /dev/da0s2f
   DUMP: Date of this level 0 dump: Wed Nov  3 12:39:19 2004
   DUMP: Date of last level 0 dump: the epoch
   DUMP: Dumping /dev/da0s2f (/usr) to standard output
   DUMP: mapping (Pass I) [regular files]
   DUMP: mapping (Pass II) [directories]
   DUMP: estimated 2764642 tape blocks.
   DUMP: dumping (Pass III) [directories]
   DUMP: 0.02% done, finished in 1:29
   DUMP: dumping (Pass IV) [regular files]
   DUMP: 21.17% done, finished in 0:18
   DUMP: 55.30% done, finished in 0:08
   DUMP: 86.28% done, finished in 0:02
   DUMP: master/slave protocol botched.
   DUMP: The ENTIRE dump is aborted.

So it seems like everything goes fine, the back up starts and then it 
crashes for some reason. Has anyone ran into that lately? I am using 
512 block size and it has been working flawlessly for all the other 
filesystems. I have downed the server and ran fsck on that one 
partition and could not find anything.

The annoying part is that it works sometimes. Today it went OK on a 
Level 6, 2 days ago it went OK with a level 5 back up, and 2 months ago 
it went OK with a level 0 back up, so I can really understand what the 
problem is.

I would really appreciate some pointers. Some additional info, the back 
ups started before and after this one by the scripts work fine, so I do 
not think it is something with the tape-drive. Also, these are live 
filesystems back ups.

Thanks,
--
Ricardo Oliva
Core Systems Administrator
Zoology Department
University of British Columbia
Ph.: 604-822-3882
E-mail: ricardo at zoology.ubc.ca