Hi all

We're backing up ten hosts using Amanda; one of them intermittently gets
timeouts while all others are OK.

The errors look like:

[in the log file]
FAIL planner bastion /dev/da2s1c 20071205 0 [Request to bastion timed out.]

[in the amdump file]
planner: time 30.231: error result for host bastion disk /dev/da2s1c:
Request to bastion timed out.

There are six partitions being backed up on the system in question; if I
comment out one or more the disklist entries, the backups run
successfully. They've also run successfully in the past on all six
partitions (this problem started a couple of months ago).

The last part of the debug log on the client server looks like this:

--------------------------------------------------------------------------------
amandad: time 0.028: got packet:
--------
Amanda 2.4 REQ HANDLE 002-E83A0280 SEQ 1196885402
SECURITY USER amanda
SERVICE noop
OPTIONS features=fffffeff9ffe0f;
--------

amandad: time 0.029: sending ack:
----
Amanda 2.4 ACK HANDLE 002-E83A0280 SEQ 1196885402
----

amandad: time 0.030: bsd security: remote host fermi user amanda local
user operator amandad: time 0.033: amandahosts security check passed
amandad: time 0.033: running service "noop"
amandad: time 0.033: sending REP packet:
----
Amanda 2.4 REP HANDLE 002-E83A0280 SEQ 1196885402
OPTIONS features=fffffeff9ffe0f;
----

amandad: time 0.034: got packet:
----
Amanda 2.4 ACK HANDLE 002-E83A0280 SEQ 1196885402
----

amandad: time 0.034: pid 89078 finish time Wed Dec  5 20:10:00 2007
--------------------------------------------------------------------------------

I've increased etimeout from 1800 to 3600 but there's been no improvement.

I'm looking for any ideas to either resolve this problem or to gather more
information to find the cause.

Thanks,
Keith