3par performance test on ESXi

[YujiNexon]

We are considering the introduction of 3par.

It is want advice when to testing on the ESXi.

Test results are reported here, we look forward to a lot of opinion.

[CalvinZito]

I just saw your post and noticed there aren't any responses.  I'm not sure I understand what you're asking - you want advice to test 3PAR on ESXi but what information do you want from your testing?

The integration that 3PAR has with VMware is very deep but if you have specific questions, please ask.

[YujiNexon]

Thank you for your reply.

When you use the 3par with ESX, we had verified how much performance comes out.

Model 8200

Port FC 8Gb 2 port

SSD x 8

SAS x16

I want you to tell me if you know about that some were wondering.

1.

I was able to exert 80000 IOPS or more performance at SSD x8 environment.

ESX of the number is I think that it can exert an even higher performance if more.

Good results than had heard in advance it was obtained.

I think, this is against the test tools that work in a VM were 100% random access,
3par has kept  read cache 80 % and Write cache 25 %.

Why is it that this kind of thing happen?
How it will has cached the 3 par?

2.

Delay viewed from ESX when tested in BBB had to keep 1ms.

Although the performance seemed to have come limit, there was no change to the delay.
I wonder if such a thing can happen Why?

[CalvinZito]

Hi Yuki, I'm not understanding a few things you said so I have a hard time helping with answers. 

You are seeing 80,000 IOPS with your 8 SSDs.  I don't understand "ESX of the number is I think that it can exert an even higher performance if more." Are you asking if you'll get higher performance with VMware?  If that is, I have no way to estimate what your actual performance will be. The HPE storage architects have a tools that can help but I don't have access to them. 

I don't understand your #2 question at all.  What is BBB? Are you saying you hit some max IOPS and the latency stayed constant?  You said that the test is 100% random and read cache is is 80% while write cache is 25%. What is 80% and 25% - cache hits, cache utilization, something else? 

To get more on your performance, you might consider looking at Antemeta 3PAR Vision.  I just did a blog with a demo video that shows the software. There's a free trial license so might be worth checking it out.

I think your best option is to talk to a local resource. I'm happy to try and help but the language barrier is making it difficult to understand what you're seeing.