Paul-
I'm still tracking down some of the things you mentioned I should
check...However, I've just had a brainstorm. This specific server is
bound via winbind to a Windows 2003 Active Directory controller. Now,
while Windows doesn't specify a user specifically named backup (the
amadmin xx version did indeed prove that the build was compiled to use
the backup:backup pair) the domain has a group called "Backup Operators"
and I'm wondering if that isn't getting in the way. If not, the
authentication could be:
I've got all this running through PAM, where the winbind.so module polls
the Domain Controller for its list of users. After a timeout, it
continues with the rest of the nsswitch.conf file. This timeout could
be causing a situation where the amcheck process is shutting down before
the winbind.so has had time to timeout and move on to the local
/etc/passwd entries. Does that sound possible? Any way I can twiddle
things in the xinetd drop file configuration to test this, or some other
config file maybe? Thanks for your input!!
-Kermit Short
Paul Bijnens wrote:
Kermit Short wrote:
Hi Marc, thanks for your ideas! The system does indeed resolve it's
full FQDN both through dig/nslookup, but pinging the IP address does
not resolve the hostname. Not sure if this is an issue?
The output from your netstat command does indeed list the correct
ports and services listening from xinetd:
[EMAIL PROTECTED]:/tmp$ netstat -taun | egrep :1008.
tcp 0 0 0.0.0.0:10082 0.0.0.0:*
LISTEN
tcp 0 0 0.0.0.0:10083 0.0.0.0:*
LISTEN
udp 0 0 192.168.34.15:10080 0.0.0.0:*
I would expect to see amandad listen on ALL interfaces, especially
including localhost!!
Verify xinetd configuration for amandad.