VMware Cloud Community
mark_a_k
Contributor
Contributor
Jump to solution

what is considered high KBps for a VM?

I am attempting to analyze some VMs to determine how they are using disk latency. VMware performance charts allow me to gather graphs showing KBps on an individual VM basis. I am seeing very large differences on the existing VMs.

I have two VMs that have an average of 2000 and 300 KBps but hitting peaks of up to 20000 and 3000.

What is considered high KBps?

I know that disk manufacturer and Raid type is a factor here. The EMC tools monitor disk latency in IO for read and writes and i am not sure how to relate to these in KBps.

Thanks for any info here.

Mark

0 Kudos
1 Solution

Accepted Solutions
FredPeterson
Expert
Expert
Jump to solution

Your best bet then is to use ESXTOP at the command line and drill down through the HBA view. First figure out which vmhba is the LUN you want to look at and then drill into the vmhba pathing to the point you can get the world ID's of the VM's on the LUN.

Unfortunately then you'll have to play the game of figuring out which VM's are which world ID's by expanding views and...its just a really annoying process. Hopefully you wouldn't have to do that. I wish VMware had an easy view of the LUN's by VM WITH name. The disk view in esxtop sucks frankly.

If someone knows of a better way of doing this, please chime in.

View solution in original post

0 Kudos
6 Replies
AWo
Immortal
Immortal
Jump to solution

Which EMC tool are you using? With CC or the Performance Manager you are able to see KBps, also.


AWo

\[:o]===\[o:]

=Would you like to have this posting as a ringtone on your cell phone?=

=Send "Posting" to 911 for only $999999,99!=

vExpert 2009/10/11 [:o]===[o:] [: ]o=o[ :] = Save forests! rent firewood! =
0 Kudos
mark_a_k
Contributor
Contributor
Jump to solution

we are using navisphere analyzer

0 Kudos
AWo
Immortal
Immortal
Jump to solution

In Navisphere I can see Kbps metrics for the SP, LUN's and for the disks.


AWo

\[:o]===\[o:]

=Would you like to have this posting as a ringtone on your cell phone?=

=Send "Posting" to 911 for only $999999,99!=

vExpert 2009/10/11 [:o]===[o:] [: ]o=o[ :] = Save forests! rent firewood! =
FredPeterson
Expert
Expert
Jump to solution

When it comes to disk the throughput isn't important, really - unless you need high throughput numbers.

What matters most is latency (and IOPS) and the majority of time latency is due to the number of IOPS coupled with throughput demands on the disk.

Don't worry about your disks throughputs unless latency is an issue. If you think a VM is doing too much disk activity then investigate it the same way you would a physical server.

0 Kudos
mark_a_k
Contributor
Contributor
Jump to solution

the raid group that contains the VMs has high utilization at the moment and i would like to determine if this is being caused by a few of the VMs or are all of the VMs causing it.

0 Kudos
FredPeterson
Expert
Expert
Jump to solution

Your best bet then is to use ESXTOP at the command line and drill down through the HBA view. First figure out which vmhba is the LUN you want to look at and then drill into the vmhba pathing to the point you can get the world ID's of the VM's on the LUN.

Unfortunately then you'll have to play the game of figuring out which VM's are which world ID's by expanding views and...its just a really annoying process. Hopefully you wouldn't have to do that. I wish VMware had an easy view of the LUN's by VM WITH name. The disk view in esxtop sucks frankly.

If someone knows of a better way of doing this, please chime in.

0 Kudos