Disk Reporting Issues: RDM's

castlerobertd

Fellow THACk'ers,

Having some issues getting accurate disk data on clustered SQL servers using RDM's (Pure Storage). When there is a fail-over the disks will move to the new active node and im not sure if thats causing confusion in Orion. After each fail-over I make sure to list resources and rediscover to make sure I can see the disks on the node's summary page, but still not 100% accurate. (Ive attached a screenshot). Is there any difference in Orion between monitoring RDM and VMFS and should I be doing something different? Were currently polling these nodes via WMI and id assume using Agent would be more accurate, just waiting on change management approval...yay. Any advice? 

RDMs.PNG

sturdyerde

Monitoring clusters usually works well if you follow this high-level approach: 

• Discover and monitor each node of the cluster
• On each node, click 'List Resources' and deselect the disks
• Discover and monitor the cluster name object (computer account of the cluster)
• On the CNO, click 'List Resources' and deselect the typical compute resources (CPU, memory) but make sure that the disks ARE selected

Using this approach, the disks will be monitored wherever the cluster resources are active, and the actual node's will be monitored directly for everything else.

Finally, you would also apply and SQL Server application monitoring templates to the CNO, 
and monitor the Windows Failover Clustering Service on all nodes.

EDIT: Just wanted to clarify specifically that this approach works for our clusters that use RDMs (file) and that use VMFS (SQL).