Planning for some redundancy in CM07 or CM12?
But what about the idea of redundancy? The worry that if you take the advice and go with one primary and no CAS (or no central for CM07), that all your eggs are in one basket. That if the server goes down, clients can't update or get software. So maybe to primary sites are better? Not so.
What you can do is just add a couple servers that clients talk to. Put the MP\SUP\DP roles onto each box and you're almost there. An MP still needs to go talk to the primary site to get information to give to a client requesting policy. So if you set up a very simple, small, supported SQL replica to the MP, it will have all the information it needs to give clients their policy, even if the primary site is down. Clients can then get policy, scan, patch, run software, send in inventory, etc., all by talking to one of the 2 new servers.
Note that by adding just one MP\SUP\DP box does nothing for redundancy other than to offer uptime for servicing or patching of the primary site. But if that one new box goes down, you're in the same boat; clients can't update or get software. If you go this route, you really want at least 2 of these servers. And preferably in different sites (if you're worried about the outage of a site).
This gets more complicated with more roles, but the concept is the same. Once you get the strategy, it's easy to start thinking about other client communications you might need to keep going.
And if you're wondering if that SQL replica to the primary isn't just as complicated as a primary to a CAS or a central, it's nothing like it. It's wafer thin! In my world of 1000's of packages and collections, that DB replica still amounts to just a 5GB file. Enough to reside in RAM on any 8GB box. And we've never had replication issues since we got off SQL2K many years ago. Set it and forget it.
- Created on .