VMware Cloud Community
dhaddad1
Contributor
Contributor

Poor performance with Dell PERC H700

I'm having performance issues host specifications are:

CPU Capacity: 18.08

Memory: 15.99

Storage: 399 GB

Plenty of available resources on this box. It is running 2 guest OS's both windows server 2008 r2. Im seeing significantly degraded performance vs the same specs running the guest directly.

Upon reviewing the log i noticed this, seemed strange:

CropperCapture[28].jpg

RAID 6 Array, on a H700. Thoughts? How can i further troubleshoot? I'm a bit new to vsphere web client FYI

Tags (3)
0 Kudos
5 Replies
Wh33ly
Hot Shot
Hot Shot

The image might look a bit strange but that is explainable, you probably took the view from "Past Month or something similar. You see an increase in counters in the last 7 days and even more from the last 24 hours.This is because the statistics are archived and averaged, so you see more "entry points" from the last 24 hours, less entry points in the past week, and even more in a monthly view.

By default vCenter uses four collection intervals.

IntervalFrequency
1 Day  5 Minutes
1 Week30 Minutes
1 Month2 Hours
1 Year 1 Day

So in 1 day you have each 5 minutes a sample with data, that will make 288 data entry points with performance data.

After the day passed these samples will be averaged out to 30 minutes sample, so you only have 48 data entry's left from the "past day" etc.


The further you look back, the less likely you can actually spot "short" spikes which caused a problem.

But to help you with your problem, we could use a bit more background information.

What kind of degraded performance do you notice ?

When does it happen ?

How are your VM's configured ?  (CPU/MEM, VMware tools installed?)

At first view I don't see why 2 VM's can't run properly, might be something as simple as VMware tools not installed which can cause your session become sluggish and seem to be a performance problem.

0 Kudos
dhaddad1
Contributor
Contributor

Performance issue: Server is a web\SQL server for staging an app. When running the app on my local machine, i see much better performance then when it is running on the staging server, and the stage server has seriously higher specs then my local box. It appears to be all the time, i have not noticed slower \ faster periods

VM Configuration:

CPU: 8 Cores 4, No reservation or limit, HT sharing: Any MMU: Auto

Memory: No limit

HD: 200GB, LSI SAS

CD \ Floppy disconnected

VMTools Installed

The machines current ussage says CPU 119, Mem 737MB, Storage: 110GB

Attached are the host graphs from 1 day timeframe:

CropperCapture[30].jpgCropperCapture[31].jpgCropperCapture[32].jpgCropperCapture[33].jpg

Host Info, realtime:

STORE2 Average Read Latency: 0.2 Max 4

STORE 2 Average Write Latency: 11.07 Max 51

The store is a PERC H700 RAID10 Array with 6 15k RPM drives

0 Kudos
vlho
Hot Shot
Hot Shot

Hi dhaddad1,

have you raid controller with cache and battery?

Look here:

https://communities.vmware.com/message/2284963#2284963

0 Kudos
dhaddad1
Contributor
Contributor

The host runs a PERC H700 it has cache and battery. When i can take it down for service this evening I will capture the array settings. Do you agree that there is a HD performance issue? Is it a host issue and not a configuration issue on the virtual machine?

0 Kudos
dhaddad1
Contributor
Contributor

Vlho,

I'm having similar problems across the 3 R710s, they all have slightly different configurations. I took one offline so i could troubleshoot it further, here are the machines specifications:

Dell R710

2x Xeon 2.26 (16 total cores) = 18.08ghz

16 GB RAM (this one is a bit lean on ram, others have 32GB)

6x 10k SAS, RAID10 = 816 GB Array

2x gigabit eth

RAID: PERC6/i, 6.3.1-0003, 1.22.32-1371, Version 2.04.00, Adaptive Read, Write-Back enabled, Force write-back with no battery enabled as well

Just another example of "poor performance" this machine has 2 gigabit connections attempting to migrate a 80GB VM from another host onto this host takes just under 2 hours.

0 Kudos