SafeKit mutual takeover architecture
Two active servers mirroring each other
With two mirror application modules and two physical servers, it is possible to implement a mutual takeover architecture. Essentially:
- Server 1 runs Application A and acts as a back-up for Server 2
- Server 2 runs Application B and acts as a back-up for Server 1
Normally, each application runs on its own server, with the other server designated as backup.
If one server fails, both applications will be active on the same physical server (see the page SafeKit mirror architecture for a full description). After the failed server restarts, its application will return to run on its default primary server.
Note that during a failure, the remaining server has to be able to handle the combined workload of both applications. If this is feasible, then a mutual takeover cluster is a more economical solution than two separate mirror clusters, because there is no need to invest in back-up servers that will spend most of their time sitting idle waiting for the primary server to fail. If you have several applications to protect, the SafeKit N-to-1 architecture is another alternative to installing multiple separate mirrors.
The mutual takeover architecture requires the SafeKit multi-module feature.
More information:






