I am real stuck on the option available for fault-tolerance for Business Objects XI R2 input/output FRS where no SAN is available.
Here the scenario.
I got two servers with the full BOXI R2 Stack installed.
Both CMS are clustered together
Both FRS on both servers are pointing to their own D:\BOStorage
As you all know the first FRS that start is active and the other one is set up passive.
Therefore what i am trying to do is work out a way to provides two way file replication on both servers on D:\BOStorage
Why two-way? the reason why i need two way is because if server 1 FRS goes down, then server 2 FRS will be active. The data will start writing to server 2 d:\BOStorage, however i also need server 1 d:\BOStorage to keep in-sync with server 2.
I had a look at Windows Server 2003 DFS, it only provide one way replication. Is it possible to configure two one-way replication.
Is there any product out there that works two-way replication in real-time.
You can have two or more FRS services running in your BOE cluster (whereby only one of those FRS services will be active, if it fails one of the passive FRS services will become active automagically), but all FRS services must point to the same file location!
The clustering of the filestore must be transparent to the BOE cluster.
Similar as database clustering (for hosting the CMS databse) must be transparent to the BOE cluster.
Your best option is a SAN (Storage Area Network).
Note:
If you had two separate BOE clusters each with its own filestore location, but one of the clusters being used as a disaster recovery site you would be fine. Assuming the disaster recovery site is not active (meaning offline). You would then replicate your CMS database (on a daily basis for example) to your disaster recovery site and at the same time replicate your filestore to the disaster recovery site.
If this is for Windows , SAN drive (FRS) can not be mapped to two windows servers at the same time. Unless you use Windows OS Clustering instead of XI R2 Clustering . NAS might be a better option on Windows Servers for FRS Fault tolerance.
Agreed, the SAN approach is the most advantageous. If you do not replicate using some sort of tool, such as EMC replication then you have a single point of failure at the SAN level. Now, the likeliness of the SAN crashing is slim but can happen. It really depends on your projects funding. The Suresync product seemed to be very cost effective for us, not to mention we do not have 2 SANs for storage.