I just changed the taperalgo parameter in the amanda configuration per
someone's suggestion (regarding a different problem) and have noticed
that 

        taperalgo               largestfit

Appears to work backwards.  It is backing up all the smallest filesystems
first, and putting the biggest ones at the end.  Which means that it hits
tape end on these, and uses nearly 3x as many tapes as before.

Here's an example partial log entry...

FINISH planner date 20041029
STATS driver startup time 6104.079
SUCCESS dumper web013.svk.isite.net / 20041029 1 [sec 1.457 kb 64 kps 43.9 
orig-kb 63]
SUCCESS dumper web031.svk.isite.net / 20041029 1 [sec 1.512 kb 32 kps 21.2 
orig-kb 31]
SUCCESS dumper web012.svk.isite.net / 20041029 1 [sec 1.207 kb 32 kps 26.5 
orig-kb 31]
SUCCESS dumper web014.svk.isite.net / 20041029 1 [sec 1.948 kb 32 kps 16.4 
orig-kb 31]
SUCCESS taper web013.svk.isite.net / 20041029 1 [sec 2.366 kb 64 kps 27.0 {wr: 
writers 3 rdwait 0.000 wrwait 0.007 filemark 2.357}]
SUCCESS dumper anubis.svk.isite.net / 20041029 1 [sec 3.288 kb 64 kps 19.5 
orig-kb 63]
SUCCESS taper web031.svk.isite.net / 20041029 1 [sec 2.354 kb 32 kps 13.6 {wr: 
writers 2 rdwait 0.000 wrwait 0.005 filemark 2.348}]
SUCCESS taper anubis.svk.isite.net / 20041029 1 [sec 2.361 kb 64 kps 27.1 {wr: 
writers 3 rdwait 0.000 wrwait 0.007 filemark 2.352}]
SUCCESS dumper gamera.svk.isite.net /opt 20041029 1 [sec 5.488 kb 64 kps 11.7 
orig-kb 63]
SUCCESS dumper smaug.svk.isite.net / 20041029 1 [sec 5.045 kb 64 kps 12.7 
orig-kb 63]
SUCCESS taper web012.svk.isite.net / 20041029 1 [sec 2.354 kb 32 kps 13.6 {wr: 
writers 2 rdwait 0.000 wrwait 0.005 filemark 2.348}]
SUCCESS taper gamera.svk.isite.net /opt 20041029 1 [sec 2.356 kb 64 kps 27.2 
{wr: writers 3 rdwait 0.000 wrwait 0.007 filemark 2.347}]
SUCCESS dumper web013.svk.isite.net /opt 20041029 1 [sec 1.717 kb 32 kps 18.6 
orig-kb 31]
SUCCESS taper smaug.svk.isite.net / 20041029 1 [sec 2.355 kb 64 kps 27.2 {wr: 
writers 3 rdwait 0.000 wrwait 0.007 filemark 2.347}]
SUCCESS dumper web012.svk.isite.net /usr 20041029 1 [sec 2.776 kb 64 kps 23.0 
orig-kb 63]
SUCCESS taper web014.svk.isite.net / 20041029 1 [sec 2.354 kb 32 kps 13.6 {wr: 
writers 2 rdwait 0.000 wrwait 0.005 filemark 2.348}]
SUCCESS taper web012.svk.isite.net /usr 20041029 1 [sec 2.350 kb 64 kps 27.2 
{wr: writers 3 rdwait 0.000 wrwait 0.007 filemark 2.342}]
SUCCESS taper web013.svk.isite.net /opt 20041029 1 [sec 2.340 kb 32 kps 13.7 
{wr: writers 2 rdwait 0.000 wrwait 0.005 filemark 2.334}]

...etc.

Farther down are 15, 20, and 30 gigabyte filesystems, which it scheduled at
the very last. One per tape, with no small filesystems filling up the tapes.
 
-- 
Joe Rhett
Senior Geek
Meer.net