jeremy wadsack writes: >It's run by inetd (or xinetd) so it is started anytime a new >connection comes in. If it doesn't get anything from the connection it >stops. Otherwise, you could end up filling up the memory / process >space with amandad instances by nudging the port and then not >connecting. >
Ahh, I see. That makes a little more sense. =] >> The pertinent lines on the client's /etc/inetd.conf (which, for the time being, >> happens to be the server as well) are as follows: > >> amanda dgram udp wait amanda /usr/lib/amanda/amandad >> amandad >> amandaidx dgram tcp wait amanda /usr/lib/amanda/amandaidx >> amandaidx >> amidxtape dgram tcp wait amanda /usr/lib/amanda/amidxtape >> amidxtape > >> ...on three lines instead of six, though, obviously. > >Uh, you said before that you were using xinetd. If so, don't you need >to add a config file to /etc/xinet.d/ for the amanda connections? > Yes, sorry for the confusion. I had plans of upgrading to xinetd from inetd on a RedHat 6.2 system. I have scripts for amanda located in /etc/xinetd.d, however at the moment I've decided against switching to xinetd just yet, so, alas, I've added the lines in /etc/inetd.conf. >If you are using inetd, then make sure you restarted or reloaded >inetd. > Oh yes, many times. =] >In either case, check /tmp/amanda for a amandad.*.debug file to see if >the connection is even starting amandad. If the file exists, it may >give you more insight to where the problem is. > Actually, if I HUP the inetd process, the /tmp/amandad*debug files are not even created. If I go back to my errorful ways of starting amandad by hand, it creates the file, but of course that's not the correct way to start amandad. On top of that, I can't telnet to port 10080 (connection refused), and I do not have any sort of packet filtering or tcpwrappers on the machine at the moment. -Dave -- david a. brooks * systems administrator * stayonline.net * voice: .. 770/933-0600 x217 * email: .. [EMAIL PROTECTED] * :wq!