Has any one setup a Pull server cluster? I’m looking to have two servers with a shared database for compliance data. DFS can handle the config and modules but I’m unsure how to do the database. Can the .edb be placed on a SAN?
You’d change the connection string to point to a clustered SQL server. And the actual pull server would be a web farm, not a cluster per se, usually.
Vague but informative. I was not sure it would work with SQL. I will have to talk to the DBA.
Unfortunately, as the compliance server isn’t actually, er, documented… everything about it is a bit vague. I don’t personally know anyone who’s changed the connection string (in its .config file), but it’s supposedly possible. And is definitely the only way to point it to something other than the Windows Internal Database, or to make it cluster-friendly.
What we have done is to place the Windows Internal Database (i.e. Devices.mdb ) on a file share that is then accessed by two pull servers which are part of an F5 load balanced cluster. Seems to be working although the compliance information returned by the compliance OData service isn’t accurate. That is, it is missing one of the two servers that are configured to get its configuration from a the two node pull server cluster, and the one server this is listed comes back with the wrong IP address. The compliance information seems to be very flaky - a shame because it would be very useful to us.
There’s a trick to how the compliance component works, and a reason why sharing an MDB can be problematic. The service actually keeps the information in-memory, and it only “dumps” to the database under specific circumstances.
There’s a reason Microsoft hasn’t documented ANYTHING about this component - it’s probably going to change a good bit in an upcoming release. I’m not sure I consider the current version entirely “done,” personally.