It's very rare that I don't come up here in the morning and don't see
failed and strange dumps in my 10 machines backup.

Most of the time, this occurs because some housekeeping/cleaning is
between the time the estimates and backups are done.

Most of the time, the message is there because a file disapeared, or
changed in size.

I can live with it.

Since I've put my amanda indexes and log elsewhere (in
/usr/local/var/amanda) I often get the following kind of messages:

/-- lnx-que-am / lev 2 STRANGE
sendbackup: start [lnx-que-amanda:/ level 2]
sendbackup: info BACKUP=/bin/gtar
sendbackup: info RECOVER_CMD=/bin/gtar -f... -
sendbackup: info end
? gtar: ./usr/local/var/amanda/idx/DailySet1/index/chablis/_/20050921_1.gz.tmp:
Warning: Cannot stat: No such file or directory
? gtar: 
./usr/local/var/amanda/idx/DailySet1/index/lnx-que-wforms1-dev/_WEB__DEP__6i/20050921_1.gz.tmp:
Warning: Cannot stat: No such file or directory
| Total bytes written: 372920320 (356MiB, 2.7MiB/s)
sendbackup: size 364180
sendbackup: end
\--------

Looks like some temporary compressed index files from the current run
should be there but are not. Could there be a problem here ? If not,
is there a way to stop amanda from complaining about missing files ?

Do you see similar behavior ?