Hope this makes sense and would appreciate any help as Microsoft Support can't seem to assist with this issue.....
We are migrating from a multi-domain NT4 environment to single domain Win2K Active Directory. We have 3 regional hub sites around the world (Regions A, B & C). We have set up site links from the hub sites to each of the local sites in their respective regions and site links from hub-site to hub-site.
In Region A and B the KCC has chosen the hub sites (Site A and Site B) to be the replication bridgehead server and it has remained that way from when we first brought up the AD servers.
However, in Region C the KCC keeps changing sites. We brought up the hub site's server first and then gradually brought up 20-odd other sites around the region. The KCC swapped the replication bridgehead to site C3 shortly after that site's server was brought on-line. We brought down the server in C3 and the KCC went back to the hub site being the replication bridgehead. When we brought up the C3 server, the KCC went back to that site again. The bridgehead has also been C7 and C13 for short periods of time.
We left C1 being the bridgehead and continued with the roll-out. Then when we brought up a server in Site C19, with the WAN link to that site being the worst in the region (64K). Now Site C19 is the bridgehead server. All replication goes through that site and is killing the WAN link. We brought the server in Site C19 down for 4 days. In that time the KCC chose the Hub Site again. When we brought up the server in Site C19 again the KCC chose that site as the bridgehead.
We have had on-site reports and conference calls with Microsoft engineers. They have advised that the KCC is working correctly and does not take into account bandwidth or server specs. However, what they cannot explain is why this is happening in only one of our three regions. The other two regions the KCC has left the hub site as the bridgehead since the start of the AD roll-out.
We want the bridgehead server to be the hub site but we cannot figure out why this is bouncing around the region. All three region's AD servers have been built exactly the same using strict documentation.
Thanks
jpaf
We are migrating from a multi-domain NT4 environment to single domain Win2K Active Directory. We have 3 regional hub sites around the world (Regions A, B & C). We have set up site links from the hub sites to each of the local sites in their respective regions and site links from hub-site to hub-site.
In Region A and B the KCC has chosen the hub sites (Site A and Site B) to be the replication bridgehead server and it has remained that way from when we first brought up the AD servers.
However, in Region C the KCC keeps changing sites. We brought up the hub site's server first and then gradually brought up 20-odd other sites around the region. The KCC swapped the replication bridgehead to site C3 shortly after that site's server was brought on-line. We brought down the server in C3 and the KCC went back to the hub site being the replication bridgehead. When we brought up the C3 server, the KCC went back to that site again. The bridgehead has also been C7 and C13 for short periods of time.
We left C1 being the bridgehead and continued with the roll-out. Then when we brought up a server in Site C19, with the WAN link to that site being the worst in the region (64K). Now Site C19 is the bridgehead server. All replication goes through that site and is killing the WAN link. We brought the server in Site C19 down for 4 days. In that time the KCC chose the Hub Site again. When we brought up the server in Site C19 again the KCC chose that site as the bridgehead.
We have had on-site reports and conference calls with Microsoft engineers. They have advised that the KCC is working correctly and does not take into account bandwidth or server specs. However, what they cannot explain is why this is happening in only one of our three regions. The other two regions the KCC has left the hub site as the bridgehead since the start of the AD roll-out.
We want the bridgehead server to be the hub site but we cannot figure out why this is bouncing around the region. All three region's AD servers have been built exactly the same using strict documentation.
Thanks
jpaf
