Question: Is it better to use H/W RAID over S/W RAID when implementing HP OpenView NNM, and why?
Question: Is it better for Backup/Disaster Recovery to design redundancy by attaching H/W RAID to each OpenView server instead of sharing the H/W RAID with a primary and backup server initially, and why?
Specifics: I have just taken over a new project in which the design has been approved for 2 Sun 220R Servers with only one H/W RAID 5 configuration Sun Storage A1000 device ordered for the primary server. The initial plan has the primary server being placed into production physically attached to the A1000. The backup server will be a "cloned" image of the primary server and basically turned off sitting in the rack.
The current design has the RAID 5 Sun A1000 having its disks partitioned to include all NNM configuation and database directories
/var/opt/OV /etc/opt/OV /opt
and the CiscoWorks directories that are planned for growth.
In the event of the primary server crashing, the backup server is currently planned to be configured with the same IP address as the primary server, and will just be connected to the network (FastEthernet), and the SCSI cable disconnected from the primary server to the A1000 H/W RAID device and re-connected from the backup server to the A1000.
I believe that it would be better to order another A1000 unit for the backup 220R, and mirror the production server with a full backup of both the 220R and A1000. This will enable the entire unit to be taken off-site for Disaster Recovery (to be implemented at a later date).