High Availability and Failover Scenarios

[Eleuthera Lewis]

What are the high availability or best failover practices for the Performance Center, Data Aggregator, and Data Collector? If any one of those is lost, how do you get back to where you were?

I've spent several hours looking and it appears the smoothest, most unoticable transition is if you have everything as a virtual machine, just make a complete copy of that VM.

[David Michel]

The ARCserve High Availability is the solution for the old NetQoS products, but I'm not sure where things are holding with support for IM out of the box.

The DR can be setup over several systems in a RAID type configuration so even if one of the systems supporting the database goes down the database itself will stay up. So with a high availability solution as mentioned above the fail over DA or DC system will come up just like the down primary with the same system name and IP. It can be setup as custom configuration for Linux installations using a simple File Server scenario and choose Block Synchronization for the sync method. just DNS redirection should work fine.

[Eleuthera Lewis]

This post seems to indicate if you lose a Data Collector you can NOT sync the same devices from a new/replacement DC to their existing Repository database entries. So if the DC fails you will have one CAPC entry for the device under the old DC and another CAPC entry for the same device under the new DC? That's why I'm thinking the only way to avoid this is to have a complete image of the DC server stored elsewhere.
99332757

Regarding the Database Repository, is it bandwidth intensive to have cluster nodes 1 & 2 in SiteA, then cluster node 3 in SiteB? I know each node has a copy of the full database, but I'm not aware of if there is any compression with that or how real time that replication is.

From the manuals it looks like the only component with any high availability option is the Data Respository. In the event of a failure of the other components the only option is to restore a database backup.

EDIT:

Looks like this is already a highly requested feature

Idea: High Availability