Well, it's working again. I think that amandad was confused or something. There
were some lost items (shown by the ipcs command) that I recovered with ipcrm.
Also, I had set up /tmp/amanda as a symbolic link to a different partition with
more disk space when I first installed it (and was getting huge debug files), so
I moved it back. I'm thinking perhaps there was a permissions problem that had
decided to magically appear. After making these changes, and restarting inetd,
things were back to normal. Thanks for your ideas, John, they helped me to try
the types of things that finally fixed the problem.
Using amanda 2.4.1p1 and FreeBSD 4.2
--- Eric
"John R. Jackson" wrote:
Amanda has been running smoothly for months, then suddenly last night's dump
failed on all the samba shares (NT boxes) as well as the amanda host and tape
server (also the samba server).
That's all because of amandad not working. Once we get it going again,
this will probably all go away.
I did a HUP to inetd, and the last line of amandad.debug shows:
amandad: error receiving message: timeout
That line must be from after you ran amandad by hand. What happens to
the file if you run amcheck? Does it get updated?
I know you said you went through the FAQ, but could you confirm that if
you do a "netstat -a | grep amanda" that it shows someone listening?
When you did the HUP to inetd, did it log anything (/var/adm/messages
or wherever)?
You might try removing /tmp/amanda and see if it gets recreated.
What version of Amanda are you using? What type of OS?
--- Eric
John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
--
===
Eric Wadsworthemail: [EMAIL PROTECTED]
Conceptual Systems and Software http://www.consys.com
===
