VMstore performance

View trends

IOPS

15,736 IOPS

1,269 - 19,073

 

Throughput

321 MBps

57 - 659

 

Latency

1.0 ms

7.5 ms

Flash hit ratio

100 %


Performance Reserves

49% free

auto allocated
free
51% used

Physical Space

9.7 TiB free

other
live data
snapshots
free
284.2 GB used (10%)
3.6 TIB Logical (10.2x space savings)

Protection

0 MB replicated a day

Performance reserves changers

New_DB 26.4% 26.4
MySQL VM0.6% -9.4
Test_411.2%   9.3
RHEV_VM0.8%  -5.9
OpenStack VM16.1%  -4.5
HyperV VM30.6%  -3.9
HyperV VM40.2%  -2.8
XenServer VM11.5% 1.5
Eng Desktop_NewHire6.3% -0.9
Sales Desktop_NewHire0.6%  0.6

Space Changers

HyperV VM30.2 GB 0.2
SWDev_2 57.5 GB 5.0
Test_40.2 GB 0.2
RHEV_VM0.2 GB 0.2
XenServer VM10.2 GB 0.2
MySQL VM0.2 GB 0.2
OpenStack VM10.2 GB  0.2
HyperV VM457.5 GB 5.0
Eng Desktop_NewHire0.2 GB 0.6

  • 13000
  • 9750
  • 6500
  • 3750
Sat 3:00PM Wed 2:50AM Sat 2:50PM

Contribution on Sat 2.47PM

  • New_DB 17.5 %
  • MySQL VM 6.1 %
  • Test_4 4.1 %
  • RHEV_VM 2.5 %
  • OpenStack VM1 5.2 %
  • HyperV VM3 2.3 %
  • HyperV VM4 2.3 %
  • XenServer VM1 2.2 %
  • Eng Desktop_NewHire 2.2 %
  • Sales Desktop_NewHire 2.2 %








Here’s your Tintri dashboard. You’ve just been asked to onboard a new virtualized database. Can you do it without affecting the performance of your other VMs?

Mission: Determine remaining reserves and typical VM behaviors.

Action: What percentage of your reserves are “free”? Click on the number.

Clicking on your % “free” takes you to this view, so you can see how your reserves have been changing over the past 7 days. And on the right, you can see the individual VMs using the MOST performance.

Action: Which VMs are using the most IOPS? Click on IOPS in the menu in the top left corner.

Looks like your greediest VMs are consistently using around 3,000 IOPS. You expect the virtualized database you’re onboarding won’t use any more IOPS than these VMs.

Action: Lets go back to the dashboard and tie this all together. Click on the “Dashboard” tab in the top left menu.

Let’s put it all together. Your VMstore is currently using about 12,000 IOPS, and 49% of your performance reserves are available. Looks like you can easily handle the request to add another virtualized database.

Onboard the virtualized database.

You determined your remaining performance reserves from your dashboard, then drilled in to see the IOPS used by individual VMs.

Summary: You proved you can have total visibility of performance with no support calls, PerfStats or conventional complexity!

RETURN TO TINTRI HUB