HEP SAN STORAGE PERFORMANCE MONITORING

[Ahmed-Aljiledi]

Hello,,

I have MSA2040 with 2 Pool one RAID6 HDD and the other RAID1+0 HDD, this MSA is attached with 4 ESXi Hosts, I need to monitor the IOPS performance to know if the read and write speed is good or need to be upgraded to SSD...

also, I have bought a new MSA2060 with SSD disks, after getting the IOPS performance result from the MSA2040 I will choose which RAID will be applied on the New MSA.

Thanks a lot..

[ArunKKR]

Hi,

The below commands could be used to monitor the MSA performance:

Run the below commands at least 5 times during a period of high IOs.

show host-port-statistics

Check whether the read latency being reported is less than 15-20ms and write latency less than 5-7ms
If the host ports reports latency continuously, run the below commands as well for 5 times:

show disk-statistics
show controller-statistics

Check whether corresponding disk IOPS is also reporting high values.

If there is continuous latency being reported , you might consider adding SSD disk group to the Pool to improve performance.
Adding SSD disk group to virtual pool requires separate performance tier license.

Adding read cache disk group to virtual pool does not require license,

Having said that, host port latency in MSA command outputs does not necessarily mean that the array is over loaded.
It could be due to multiple reasons like incorrect configuration, Fibre channel path issues (ex: faulty SFP/cables)
It would be good to engage support team to check for any underlying hardware/configuration issues before adding new SSD disks.

MSA 2040 supports linear vdisks as well as virtual Pools.
MSA 2060 supports only virtual pools.

For optimal write performance, parity-based disk groups (RAID 5 and RAID 6) should be created with the “Power of 2” method. This method
means that the number of data (non-parity) drives contained in a disk group should be a power of 2 (Only applicable for virtual pools)

Refer best practices guide page 20:

https://www.hpe.com/psnow/doc/4aa4-6892enw

Check whether the other best practice guidelines has been followed.

Refer to page 26 of the MSA 2060 quickspecs sheet to find performance figures for various RAID ocnfigurations:

https://h20195.www2.hpe.com/v2/getdocument.aspx?docname=a00094629enw

[Ahmed-Aljiledi]

Thank You for the replay, 

this is the output, could you please analyze it.

text file--

https://drive.google.com/file/d/1cefKDip3DGe3nKTy20TK2AVGtS73beAX/view?usp=sharing

[ArunKKR]

Hi,

I could see random spike in write latency in few instances in A2/A3 ports.

This seems to be a direct connect configuration with out a switch in between as load is not evenly spread across MSA host ports.

You may upload the MSA logs at below link to check for any pending firmware updates:

https://msa.ext.hpe.com/

Log a case with HPE support team to check for any hardware/configuration issues.

[Ahmed-Aljiledi]

YES, the servers are connected directly to the MSA not via SAN switch, 

[Ahmed-Aljiledi]

[ArunKKR · Accepted solution]

Hi, 

In that case the server connected to A2 A3 ports would be pushing intermittent high IOs to the MSA.

If ESXI reports latency on the  vms present on host server connected to A2/A3 ports , you may consider moving some load to servers connected to A1 A4 ports.