This one time, at band camp, Kris Vassallo wrote:
>One day last week after AMANDA had been running fine for months it
>just broke and I now am seeing this in my email log daily:

What's your etimeout set to, and how big is the filesystem on
host.domain:/backup/repo and /home ?

I've just had the very same thing in the last week, and after some
fiddling I found it was timing out when trying to get the estimates.  I
boosted etimeout to 3600 and the dump succeeded this morning.

>Amanda Backup Client Hosts Check
>--------------------------------
>Client check: 3 hosts checked in 0.226 seconds, 0 problems found

I was seeing this too, because of course the check doesn't require
seeking over some 95G of data.

I think the etimeout is a bit of a hack, though, and would prefer to see
some ping mechanism in the protocol so that clients can let the server
know that they're still alive and doing stuff.

-- 
jaq AT spacepants DOT org                           
http://spacepants.org/jaq.gpg