Hi,

I am using version 2.5.0p2. on my dump host. One of my clients (same version) 
has a 
filesystem that consistently fails the estimate and dump phases. The same host 
has 
two other filesystems (smaller) that complete without problem. The amandad and 
selfcheck debug from this host shows no indication of problems.

The sendsize debug does show a warning, but I can't find enough information to
correct the problem or be sure that it is the problem. 

Error/Warning message: From backup report ------------

FAILURE AND STRANGE DUMP SUMMARY:
  fa1  amrd0s1f  lev 0  FAILED [disk amrd0s1f, all estimate timed out]
  planner: ERROR Request to fa1 failed: timeout waiting for REP
-------------------------

Error/Warning message: From sendsize.debug  ------------

sendsize[67866]: time 5.776: getting size via dump for amrd0s1f level 0
sendsize[67866]: time 5.777: calculating for device '/dev/amrd0s1f' with 'ufs'
sendsize[67866]: time 5.777: running "/sbin/dump 0Shsf 0 1048576 - 
/dev/amrd0s1f"
sendsize[67866]: time 5.778: running /usr/local/libexec/amanda/killpgrp
sendsize[67866]: time 5.781:   DUMP: WARNING: should use -L when dumping live 
read-write filesystems!
sendsize[67866]: time 5.782:   DUMP: Date of this level 0 dump: Thu Oct  5 
19:36:58 2006
sendsize[67866]: time 5.783:   DUMP: Date of last level 0 dump: the epoch
sendsize[67866]: time 5.784:   DUMP: Dumping /dev/amrd0s1f (/usr) to standard 
output
sendsize[67866]: time 5.857:   DUMP: mapping (Pass I) [regular files]
sendsize[67866]: time 17.022:   DUMP: mapping (Pass II) [directories]
sendsize[67866]: time 17.022:   DUMP: estimated 5457824 tape blocks.
sendsize[67866]: time 17.027: .....
sendsize[67866]: estimate time for amrd0s1f level 0: 11.249
sendsize[67866]: estimate size for amrd0s1f level 0: 5457824 KB
sendsize[67866]: time 17.027: asking killpgrp to terminate
sendsize[67866]: time 18.035: done with amname 'amrd0s1f', dirname '/usr', 
spindle -1
sendsize[67854]: time 18.036: child 67866 terminated normally
-------------------------

I compiled the client with amanda_snapshot, and I can see a .snap directory in 
the filesystem
noted above. 

One question I have is, How and where do you specify "dump -L"?

I get this same warning on one of the other filesystems on this host, but the 
estimate and
dump finish with no problems.

Backups on the host completed when host and server were using (Amanda-2.4.5). 
I appreciate any help you can provide in solving this.

Thanks

        -Mike
 
-- 
Michael Galvez             http://www.people.virginia.edu/~mrg8n
Information Technology Specialist         University of Virginia