Very slow VDI Performance

[Smitty23]

Hi All,

I'm having a very strange problem with my XP Desktops in my VDI Farm. I've got 2 HP DL580 Servers with four-quad processors, and 32GB of RAM. I've got 55 workstations, and 2 VDM servers spread across the 2 ESX Servers. My VDM boxes with 1GB of RAM seem to perform fine, but my workstations have serious lag when performing various tasks in Windows. There is almost no load on my ESX Servers, or on the VM's themselves. The template I used to deploy these VM's was located on other ESX Servers which had an older version of VMWare Tools, but I've upgraded it on the template when I moved it over. All the desktops dont need an update of the tools, but when I look at the VM within VC, it does indicate the VMWare Tools are out of date, even though they show the correct version. Also, each XP VM has 1 processor, and 512MB of RAM.

Any advice is appreciated.

Thanks

[mittim12]

If you are using VDM 2.0 then there is a known display issue with the tools being out of date when the VDM agent is installed in the VM. The tools function properly and are indeed updated but they happen to show out of date. This is fixed in version VDM 2.1.

As far as performance issues, your setup for CPU and Memory sounds like more than enough. You didn't specify any type of disk configuration so i am curious about that. I am also curious as to what test you have done to try and isolate the issue.

If you found this or any other post helpful please consider the use of the Helpful/Correct buttons to award points

[Smitty23]

Thanks mittim. I just realized that the VDM Agent is installed on the template, and it is old because it came from another VM Environment. On my new environment, everything is up to the latest and greatest (Not Update 2), so I will need to get the VDM Agent upgraded. For storage, we have a SAN with a 450GB LUN for the VM's. These are located on its own disk group on the SAN.

EDIT: As for testing, we can barely use the VM in its current state. Whether its through the VI Client, or RDP, its slow as anything

[mittim12]

Just curious but are all 55 VM's on the same LUN? If so what kind of utilization are you seeing on the LUN? Also are you using any type of AV/Firwall software inside the VM? We had a case where Symantec's firewall product was installed on our template and it was butchering performance on all VM's. Once the firewall as disabled performance was back to normal.

If you found this or any other post helpful please consider the use of the Helpful/Correct buttons to award points

[Smitty23]

Yes they are all on the same LUN. The reason why I don't believe this is SAN related is because these VDI's are basically the only thing running on this SAN, and my VDM Servers are on the same LUN, and don't appear to have any issues.

[mittim12]

I know it is difficult to work in the VM's but are you able to view task manager to see if any processes are using up an extreme amount of resources? Any type of ballooning or swapping going on with the VM's?

If you found this or any other post helpful please consider the use of the Helpful/Correct buttons to award points

[Smitty23]

I think I may have found the issue. I was running a VM on my 2nd ESX Server, when my VM stalled. I vmotion'd it to my 1st ESX Server, and it started performing perfectly. Im going to reboot my 2nd ESX Server

Also, within Windows, there is absolutely nothing running, and the VM takes a ridiculous amount of time to boot up

[williambishop]

What arrray/connectivity are you using?

Also, you might want to try to do a vmware tools install, and choose repair. I had problems out of the old tools when moving from 3.0.2 to 3.5

[Smitty23]

I'm running an HP EVA4000, the LUN is RAID 1 and its all Fibre connectivity.

[williambishop]

Try the tools, select repair. Some of my vm's were sluggish until I did this.

W

[Smitty23]

Thanks for the suggestions guys, but now Im realy thinking its an issue with 1 of my ESX Servers. Even running 1 VM on it seems to be a chore. I'm going to rebuild it and see how that fares.

[mittim12]

Just wondering if the rebuild fixed your problem or if your are still having performance issues?

If you found this or any other post helpful please consider the use of the Helpful/Correct buttons to award points

[jonas001]

Ive seen this and its usually bad fiber or HBA. You can have your SAN guy check for errors and retrys on the fiber switch port connecting to your HBA(s) Check your VM for delayed write failures in the event log to verify its disk related. Let me know what you find.