Forgive the very quick and simple review, but I want to make sure we are all on the same page...
An IBM SVC virtualizes 3rd party back-end storage devices in order to provide storage to front-end hosts. The back end Storage device maps LUNs to the SVC which it effectively uses as Disks (mdisks) that are aggregated into Pools (mdisk groups) which SVC LUNs (vdisks/volumes) are carved out of. Most of the time, the performance of the back-end Storage devices is largely hidden by the high level of caching employed by the SVC. In cases where the back-end controllers can't keep up with the demand it can cause the front end performance of the SVC and the LUNs carved out of that storage to suffer.
The SRM product doesn't appear to record Latency for Pools (mdisk groups) nor does it gather any back-end response time. The original Solarwinds Storage Manager product gathered the back-end response time of the back-end Storage devices as "External Response Time" at the MDisk level. That product also gathers a large number of performance statistics for all Storage device types at the Pool and Disk level including Response time, IOPs, Throughput and IO Sizes.
I can't stop monitoring with the old Storage Manager for some devices because the missing Statistics would make it much harder to diagnose issues and impossible to provide the reporting I do today.
Please work toward adding the following to the product as soon as possible;
- SVC - External Response Time - It is critical that mdisk external response time be added to the SRM product
- All Devices - Disk - Recording performance statistics at the disk level would be a benefit for any Storage device because it would help in finding hot spots
- All Devices - Pool - Recording Latency statistics would also be helpful for any Storage device to help find overloaded Pools
Feel free to contact me for any clarifications.
