I started using Amanda a few weeks ago to backup 7 systems, all is well
except for 3 systems. 

During every nightly dump three boxes fail with the message:

        FAILURE AND STRANGE DUMP SUMMARY:
          mail       hda2 lev 0 FAILED [Estimate timeout from mail]
          mail       hda1 lev 0 FAILED [Estimate timeout from mail]
          dns        hda2 lev 0 FAILED [Estimate timeout from dns]
          dns        hda1 lev 0 FAILED [Estimate timeout from dns]
          app        //fnp/geoff lev 0 FAILED [no backup size line]

One of these is a windows box and it seems to be a problem with Samba,
but I'm not too worried about that for the moment. The other two are
both Linux boxes and the only difference between these two boxes and the
other successful boxes is that they are on the other side of a firewall.

So immediately I assume its the firewall that's the problem, but I have
managed to successfully run a test dump with amanda for one of the two
machines. I set up a test that commented out everything but mail & dns
in the disk file and then mail dumped ok, but dns still failed.

They are both running identical versions of Linux.

I have snipped a section of the log from the mail machine that looks to
be the offending section:

        hda1 0 SIZE 12701
        hda1 1 SIZE 4163
        hda2 0 SIZE 5617335
        hda2 2 SIZE 419676
        ----
                                                                                
        
        amandad: time 142.165: dgram_recv: timeout after 10 seconds
        amandad: time 142.165: waiting for ack: timeout, retrying
        amandad: time 152.165: dgram_recv: timeout after 10 seconds
        amandad: time 152.165: waiting for ack: timeout, retrying
        amandad: time 162.165: dgram_recv: timeout after 10 seconds
        amandad: time 162.165: waiting for ack: timeout, retrying
        amandad: time 172.165: dgram_recv: timeout after 10 seconds
        amandad: time 172.165: waiting for ack: timeout, retrying
        amandad: time 182.165: dgram_recv: timeout after 10 seconds
        amandad: time 182.165: waiting for ack: timeout, giving up!
        amandad: time 182.165: pid 6081 finish time Tue Dec 30 00:35:02
        2003
        
If I had to make a guess, it would be that it's a communication problem
through the firewall, but I am confused by the fact that it does work
sometime in a standalone test mode. I'm hoping that this is a known
problem and that I just have open a port on the firewall or something
similar.

Can anybody cast some light?

Many Thanks,

Geoff