Hi all,

According to some sources, it is possible to install multiple BF consoles all pointing to 1 rdbms backend. This helps with load sharing and (not to purists) some sort of fail over facility.

Has anyone done this? How did you do it? What is the experience - good, bad or indifferent.

Also, how does one manage RAFW in this case? RAFW stores thing in the filesystem, <installdir>/RAFW need to be synced across BF instances. Oh, BTW let's say, NFS cannot be deployed.

Many thanks

It is certainly possible to install multiple Build Forge servers (Management Consoles) to point at a single back-end database. I have been involved with a number of different folks that have set up this configuration and continue to use this configuration.

This configuration is really about redundancy, not so much about any kind of load balancing or high availability (sort of) for true failover. What I mean by that statement is that you are removing one of the single points of failure in the system, therefore increasing the availability but not to the levels that most people mean when they claim high availability (which is 5 9s of availability).

Because of the distributed nature of the Build Forge system (DB, BF Server, and BF Agents) and the communication within the system, any activities that are in process when a failure to a Build Forge Server (Management Console) occurs will fail BUT those activities can be restarted from the remaining Build Forge Server that is still running - this is due to the loss of communication from the failing Build Forge Server and the Agent actually performing work.

As for bringing RAFW into the discussion, it is certainly possible but there is some more information that is going to need to be gathered before I could really try and produce the steps to perform.

Please feel free to contact me via email or by scheduling a meeting and we can discuss this further within the context of the specific situation.