Replication Question

[Code666]

Hi everyone,
I've looked in the FAQ's and searched to no avail. I have a small AD test NW setup at home: 2 Win2k dcs running GC, DHCP, DNS, etc., with 3 clients also running w2k. Everything is working good, except for an oddity. One of my DC's couldn't replicate a particular naming context to the other DC. It was weird, you would see connection objects for both servers in AD Sites & Services, repadmin /showreps would show both servers as replication partners (inbound and outbound), and repadmin /showconn would show the connection objects. Also, here is the event log for DC1:

Event Type: Information
Event Source: NTDS KCC
Event Category: Knowledge Consistency Checker
Event ID: 1308
Date: 1/27/2004
Time: 5:56:55 PM
User: N/A
Computer: DC1
Description:
The Directory Service consistency checker has noticed that 2 successive replication attempts with CN=NTDS Settings,CN=DC2,CN=Servers,CN=bigfirmHQ,CN=Sites,CN=Configuration,DC=bigfirm,DC=biz have failed over a period of 2842 minutes. The connection object for this server will be kept in place, and new temporary connections will established to ensure that replication continues. The Directory Service will continue to retry replication with CN=NTDS Settings,CN=DC2,CN=Servers,CN=bigfirmHQ,CN=Sites,CN=Configuration,DC=bigfirm,DC=biz; once successful the temporary connection will be removed.

Per the KB article on "TS-ing Intra-Site Replication" and looking at DNS I finally resolved the issue. There was no SRV record for DC2 on DC1, more specifically in the DNS console on DC1 the CNAME for DC2 (which had the GUID listed) was missing in the _msdcs folder. I added the record and the above error is gone now. I understand what is happening here and I see how much of a role the SRV records play, but I don't understand exactly what happens from reading the event message. Is DC1 trying to notify DC2 (as replication is notify/pull relationship) and cannot because of the missing SRV or has it notifed the neighboring DC and DC2 can't pull the update down? Or is it the other way around, meaning DC2 can notify DC1 that there is an AD update and DC1 can't get it from DC2? EventID had nothing listed unfortunately with regards to this error. Any help/insight into this is greatly appreciated!

Ed

 

[dowsley]

I believe the problem was DNS only as the servers need to communicate with DNS you need to make sure your servers are pointing to your DNS server and your DNS zones are set to receive secure and unsecure updates by default they are not. You made a manual entry and if DNS was getting updates it will create those entries. You might try opening a command prompt and ping each server by name and your domain to verify they are communicating. another place to check in in Active Directory Sites and Services under your servers you can check communication and force replication. I hope this helpes good luck.

dowsley

 

[Code666]

Dowsley, thanks for the reply. I'll check the settings but I am relatively sure that the DNS zones were set to Dynamic updates and this is how the 2 dc's are set-up DNS-wise:
DC1 -preferred is itself, no alternate
DC2 -preferred is pointed to DC1, alternate is itself.

This is per the MS KB article (don't have it in front of me at the moment)on how to set-up DNS and like I said it seemed to work. I think what happened is that the SRV for DC2 wasn't replicated to DC1, hence my needing to manually add the record. I verified this by trying to ping DC2 by its GUID and it said unknown host. I was able to ping DC1 by its GUID and it was successful. I plan to re-build the domain from scratch to see what happens.

Edward

http://www.petrelworkflow.com