Vmwareservice.exe using 500MB and higer memory

[BrianRowe]

Greetings,

We are running ESXI v4 on a Dell R710 with 24GB of ram, the box has 8 VM's setup on it all running 2003 Web Edition. We hit an issue last week where Vmwareservice.exe was using almost 1GB or ram on a few of the servers. If we restart the VMware Tools Service memory consumption will drop to 5.5mb, after a min or so it goes up to 8mb, from there it slowly continues to climb.

Currently its using the following:

Server1 529MB

Server2 338MB

Server3 333MB

Server4 338MB

Server5 339MB

Server6 339MB

Server7 337MB

Server8 429MB

We have setup a task to restart this service twice a day in order to try and keep the memory usage from making the box unstable. Has anyone else hit this issue? I have not been able to find anything on this aside from a memory leak that was fixed in 3.1

Thanks

Brian

[golddiggie]

There's something installed on all eight of your servers causing this to happen. I just checked three of the 8 servers I'm running on my ESXi 4 u1 host and the VMwareService.exe is using around 7400K on each VM. I ONLY reboot these (Windows) servers when there's an update applied that requires a reboot.

Even after a fresh reboot of the server, the VMwareService.exe process is using about the same amount of memory. The amount of RAM given these VM's ranges from 1GB to 3GB, with each one having 2 vCPU's assigned to them. These are all Windows Server 2003 Standard 32 bit VM's (SP2 installed on each VM)...

What resources have you given these VM's? How many vCPU's, memory, drives, etc... What does the host have for processors? What are you using for VM storage? How up to date is your ESXi host?

What you're doing is definately NOT 'normal'... Time to investigate what's on the VM's that's the true cause of this... If you can't figure it out on your own, reach out to VMware support...

VMware VCP4

Consider awarding points for "helpful" and/or "correct" answers.

[BrianRowe]

Thanks for the reply,

Each VM is given 4 vCPU's, 2GB of ram, they are all 32bit with SP2. The host is a new server that was updated before being put into production (firmware etc)

The host has duel Xeon E5640 2.66 processor's with 12MB cache. The servers are just web servers that were virtualized, all servers were running on Dell1425's for years without issue, just IIS serving up web sites.

The only thing I can think of that could cause an issue would be dependent on what the VMwareService is monitoring, each of these servers has hundreds of IP's assigned, do you know if it tries to monitor any type of traffic on each IP?

[golddiggie]

Start by reducing the vCPU count down to 2 per VM...

VMware VCP4

Consider awarding points for "helpful" and/or "correct" answers.

[ICT-Freak]

I have a couple of VM's with the same issue. It's a VM with 4GB RAM and 4 vCPU's and using 100 IP adresses. I have an open support ticket by VMware and they are working on this case. As a work around, we scheduled a batch file which will restart the VMwareService every day.

Did you find a solution?

[ICT-Freak]

I have tried reducing the vCPU's but this didn't solve the memory leak. I have also tried five different version of the VMware Tools but this also didn't resolve this issue.

[BrianRowe]

I did not find a solution, I have a task running twice a day to restart the service.

[ICT-Freak]

I have heard from VMware that this issue will be solved in Update 3 for vSphere 4.0 and that his issue is already solved in vSphere 4.1. Let's hope VMware will release a stand alone patch too.

I am running a test right now on vSphere 4.1. When I got more information, I'll let you know.

Arne