I have two samba 4 AD DC running using the internal DNS. On one of them, DNS
will fail after a short time (10-15 minutes). Restarting samba on this AD DC
corrects the issue temporarily. This behavior started about 2 weeks ago. We had
not made any changes to either system during this time so it is a complete
mystery. I unfortunately used the latest version of samba from git (4.2) for
creating these systems. I am in the process of building a new 4.1 server and
will join it to the domain as an AD DC (hoping that this will work correctly).
In /etc/krb5.conf, the server with failing dns is also listed as the
admin_server for the realm. No idea what the effect of this is if I can point to
the other server and still resolve dns.
First though, what log files I should even be looking at for the DNS issues.
I can run all the tests for a properly operating DNS and they all return the
correct values (up until DNS fails).
e.g.
host -t SRV _ldap._tcp.ncs.k12.de.us
host -t SRV _kerberos._udp.ncs.k12.de.us
host -t A ncssamba1.ncs.k12.de.us
all return correct information. kinit also works correctly, smbclient -L server
-U% returns the correct information. I am running nslcd on both servers and that
is also working.
Completely lost on what to try to fix this dns issue.
Sincerely,
Dave Hopkins
