Array Performance and Data Protection
1819804 Members
2876 Online
109607 Solutions
New Discussion

Nimble Read Memory Hit and Reach Cache Hit difference

 
SOLVED
Go to solution
Harshakss
Occasional Visitor

Nimble Read Memory Hit and Reach Cache Hit difference

Can someone please clarify the diffence between Random Read memory and cache hit difference which is showing in HPE Infosight

 

 

 

 

3 REPLIES 3
Candhade
HPE Pro

Re: Nimble Read Memory Hit and Reach Cache Hit difference

Nimble Read Memory Hit and Reach Cache Hit difference - I am not sure what is Reach Cache Hit in Nimble.

Can you please upload the snip of the infosight where you see this?

Cache Hit is applicable to CS/HF-series Arrays. Frequently accessed data and copy of data of the volume on which volume pinning is enabled is written to the 3/6 SSD drives in the base/expansion shelves. When the Host send read requested and if the requested data does not exist in NVDIMM/DRAM, controller will fetch it from the SSD instead of going to the NL-SAS drive. If the data is successfully fetched from the SSD and served to the host, then it is a cache-Hit.
So cache-Hit indicates how volume pinning feature and 3/6 SSD (cache drives) have helped in performance.

If the requested data exists on NVDIMM/DRAM when Host send Read IO, controller serves the host without going to the drives. This is considered as Read Hit.
I guess Read memory Hit you are talking about is the number of read IOs which are served to the host from the NVDIMM/DRAM of the controller without having to go to the drives to fetch the data.

When a read request is sent to the array, the HPE Nimble OS will retrieve the data from the fastest area possible.

HPE Nimble Storage OS starts by checking in NVDIMM. If a write was just sent to the NVDIMM, and is still there, it will be retrieved very quickly. If it is not in NVDIMM, HPE Nimble Storage OS will check DRAM, because that is the next fastest location the read can be retrieved from. If the data is not located in the DRAM it will check the Flash Cache - if found, it will validate the checksum, decompress and return the data. If the read is still not located HPE Nimble Storage OS will reference the index to locate the read on disk. Lastly if HPE Nimble Storage OS determines that the data found on is cache worthy, it will send a copy to the flash cache.


I work for HPE

Accept or Kudo

Harshakss
Occasional Visitor

Re: Nimble Read Memory Hit and Reach Cache Hit difference

Please find the below image from HPE Infosight. It is under Pool -> Performance tab.

image.png

Muru123
HPE Pro
Solution

Re: Nimble Read Memory Hit and Reach Cache Hit difference

Hi, The typo in the Read Memory hit created the confusion. Thanks for the screenshot.

 

When a read request is sent to the array, the HPE Nimble OS will retrieve the data from the fastest area possible. It would be NVDIMM/DRAM -> Flash Cache -> Drives

 

Random Read Memory hit - displays the percentage of read requests that have been received and served by the memory (NVDIMM/DRAM)

 

Random Read Cache hit - displays the percentage of read requests that have been received and served by the flash cache.

 

Hope this answers your query.


While I am an HPE Employee, all of my comments (whether noted or not), are my own and are not any official representation of the company

Accept or Kudo