Input and Output Repository 1

[hanchilicious]

I have two servers pointing to Input and Output repositories in a common network directory. I would have thought that if a new instance was created for any report via one server, the fact that the other server was looking at the same input/output would reflect the new instance on the other server.

Is it absolutely necessary to cluster the machines in order to do this? Cos I have four machines in total, in two clusters of two. I want all the servers to have the same repositories, but wanted to avoid one big cluster, as two of the four servers are in a remote disaster recovery site I wanted to keep completely distinct.

 

[bdreed35]

When you have more than one instnace of either the Iput FRS or Output FRS, one set will be active and the other passive. This means that one set will be used, and if that set goes down, the other set will become the active FRS.
You can have both sets pointing to the same location so that they use the same files and the transition is seamless.
If they are pointing to seperate locations, you need to have some sort of replication occuring so that the structures are the same. The recommendation is to have the files stored on a RAID system.

~Brian

 

[hanchilicious]

Thanks for your thoughts Brian.

However, these machines are all pointing at the same destination for their Input and Output information, meaning they should be using the same files and reflect this in their object version and instance history.

But what I'm seeing is instances being run from one server all week, but when you look at the backup server, there aren't any new instances since last week - meaning that even though both servers write out to the same output dir - the backup server isn't picking up the new instances.

 

[bdreed35]

I think your not seeing them because the backup system doesn't have any entries for those instances in its version of the system database. Are you running the backup system off of a copy of the main system's system database?

~Brian

 

[bdreed35]

Take a look at this whitepaper:

http://support.businessobjects.com/communityCS/TechnicalPapers/ce10_disaster_recovery_plan.pdf

I think it will help you out.

~Brian

 

[hanchilicious]

Brian,

That's good stuff, buddy.

To answer your initial question, the two databases are distinct from each other - due to their not being in a cluster. The second machine had it's info imported in from the first machine, rather than the db copied.

The white paper indicates that the machines do need to be clustered to achieve this. I guess that'll be the way to go then.

Thanks