There shouldn't be more than one connection object unless you enable a special option to create redundant connections. Eventually, AD will figure out that it needs to create a new connection in your scenario.
Thanks, Brian Desmond [email protected] w – 312.625.1438 | c – 312.731.3132 -----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Michael Leone Sent: Tuesday, June 9, 2015 2:34 PM To: [email protected] Subject: [NTSysADM] Fwd: Questions about connections between sites I am still confused about connections between DCs in different sites. Here's my situation: 2 sites (#1 and #2), each with 1 DC in them. 1 site (#3) with 2 DCs 1 site with 6 DCs in it (that's HQ) Site #1: DC has one (1) automatically generated connection to DC in site #2 Site #2: DC has connection to site #1, connection to site #3, and 1 to a DC in site #4 (HQ) that I am about to retire. Site #3: DC#1 has connection to site #2, and connection to other DC in site #3 DC#2 has connection other DC in site #3, and connection to DC in site #4 (HQ) Here's my worry: Site #1 has only that one connection site #2. If somehow that single DC in site #2 goes down, becomes inaccessible, etc, site #1 has no connections to anywhere else (no AD site connections, not necessarily complete loss of IP connectivity across the enterprise). That sure seems like a single point of failure to me. And once I demote that DC in site #4, my connections all seem to be Site #4 has to connection to site #3 (single connection), which then has to connect to site #2 (single connection), before changes finally replicate out to site #1. Shouldn't there be at least 2 connections for every DC? I don't understand why site #1 has only 1 connection. Oh,. I can manually create q connection to HQ (for example), but why isn't the KCC generating a 2nd connection there, as a fail safe ? Seems to be a lot of single generated connections, which doesn't seem that safe to me. What if the link to that site goes down - doesn't it take like 2 hours for AD to finally realize the connection is down, and to generate a new one somewhere else? That's how long it took the other day, when I had a similar situation happen. What am I not understanding here? Shouldn't there be at least 2 connections for every DC (preferably to different sites), in case one of those connections goes dead?
