Hi Alejandro,
Regarding your query you can do 3 level of clustering if you are using UNIX at a Opearting System, Same we are implemented in our project below is the scenario
SAP Business Objects Application:
(a) BOIPS 4.1 SP5
(b) BODS 4.2 SP5
Data Base: Oracle 11G
Opearting System: RHEL 6.0
- Using Cluster SHELL Script at Operating system level we are changing active & passive node for BOIPS Application
- Oracle Local File System is available in both the node (Active & Passive)
- BOIPS & BODS local file system available in both the node (Active & Passive)
- CRON Jobs created on both the node (Active & Passive) but in CRON Jobs we are mapped both the Job Server running in active & passive Node manually we updated CRON Jobs
So if System failover happens in that case first Operating System will star then Oracle File system mounted from Active to Passive then BOIPS Services will start using SHELL script & then BODS Job Server start using SHELL Script.
In any of the case BOIPS, BODS Services will always started on both the node.
This was the best approach for UNIX enviornment.
If you create CRON Jobs please mapped both Job Server Name (Active & Passive Node Job Server name) & BOIPS Name and copy the same SHELL Script in other node, No need to execute from both node only execute from one Node because JOb Server is running on both the node.
Hope this will help!!!!
Thanks,
Daya