topic Re: stats command in Array Setup and Networking

Meaning of Output Values in Stats Command

p-wu57 — Thu, 08 May 2014 07:58:49 GMT

Hi , regarding command "stats", I would like to know about the meaning of output.

for example, as below,

__________________________________________________________________________________________________________

/$ stats --diff --vol test-vol --interval 5

Will use per-second statistics

Stats from 2014-04-30,14:45:19 to 2038-01-19,12:14:07

vol reads rdKB writes wrKB seqRead% nsMemHit% nsSsdHit% memKB ssdKB hddKB hddSnapKB

test-vol 666 11506 4 9 1 23 2 0 0 0 0

__________________________________________________________________________________________________________

volume name: test-vol.

reads: IOPS of read ?

writes: IOPS of write ?

nsMemHit%: read cache hit rate in memrory?

nsSsdHit%: read cache hit rate in SSD?

total cache hit rate will be nsMemHit%+nsSsdHit%, right?

if anyone can have a comment?

Thanks in advance

Peggy

Re: stats command

rugby0134 — Thu, 08 May 2014 14:52:55 GMT

I believe your correct.

nsMemhit is the amount of data read back from NVram and/or DIMMs. This would be data that was written in the last few seconds and then requested for read.

nsSSDHit is the amount of data reads from the SSD pools. So the total read cache hit ratio is nsMEmhit + nsSSDHit.

From the stats above, i would guess that the data is large sequential blocks. Large sequential blocks are not moved into SSD cache because the Nimble arrays can read large sequential blocks just as fast from SSD or the SATA pools.

Re: stats command

rugby0134 — Thu, 08 May 2014 16:15:32 GMT

Here's good blog on that topic Glick's Gray Matter: Nimble Performance 101 - CLI Style