High CPU usage after adding RDM

[VirtuallyTaken]

Hi,

We have added 2 RDMs from one VM (call this VM1) to an existing VM (call this VM2). Since then, we are experiencing high CPU usage for VM2. We are running Win2K for the guest OS. The CPU is running at 100 percent pretty much all the time.But when you check perfomance monitor or task manager from the guest OS, its quite normal. No high CPU. Has anyone experienced this before? If so, any resolution or why its doing this? Is this completely unrelated to RDMs?

Cheers

[J-D]

Never had this on Windows 2003 or higher. Might be an issue with Windows 2000 but I doubt this.

If you are running 3.0.X I suggest you upgrade...there are many bugs in 3.0

3.5 runs a lot more stable. I hope vSphere will be too but we haven't run this in production yet.

[VirtuallyTaken]

Thanks J-D. Yeah I know what you mean. 3.0.2 was quite horrible. We are going to upgrade this environment tomorrow night and see how it effects it. I was wondering if this could be a disk resignaturing issue? Although, they are just data disks. Anyway, the client has decided to remove the RDMs from VM2 and put it back to the original VM tonight. See what happens tomorrow. Hopefully the upgrade will solve this.

Cheers

[krowczynski]

Hi,

we are also using a lot of servers with rdms and such behavior.

Have you tried this one with Windows 2003 Server?

MCP, VCP

[VirtuallyTaken]

Hi krowczynski,

Yeah we have plenty of environments running RDMs on W2k3 and W2k8 but no problems. See the server that these RDMs were taken from were W2k3 enterprise 32bit and added to the W2k Advanced server So I am just wondering whether its a compatibility or a resignaturing issue with this. Because, I think all the other RDMs were created from scratch and not added from another VM. Maybe this has something to do with it.

See what happens after we do the upgrade.

Cheers

[krowczynski]

It seems to be really a problem with Windows 2000.

MCP, VCP

[VirtuallyTaken]

If thats the case then I dont think there are any RDMs connected to 2000 on any of our clients, otherwise we would have surely heard about it 🐵 Maybe, it is a 2000 thing. Any issue with the guest OS in your environment, as we dont have any? Its just the VM that is complaining.

Cheers

[krowczynski]

Have you checked out if the other haven't really no RDM detected.

MCP, VCP

[VirtuallyTaken]

RDMs were definately detected. The server sees it. We connected the RDMs last friday and all was working ok. The server was not slow. Its just the VM performance stats indicate 100 percent CPU usage. Looking at the historical perf stats of the VM, it indicates that the server has periods of high CPU. Maybe the RDMs have nothing to do with it and its just Windows 2000 server. But then again, we have other windows 2000 servers and they are normal.

[mouyaq]

How have you configured SCSI adapter?? Try to create an independent scsi controller and put it into physical mode.

.

[VirtuallyTaken]

Yeah it just hangs off one scsi controller. But compatibility mode is physical. I might check out what you have suggested and attach a second controller and hang it off that. We are going to upgrade that environment to 35 tonight to see any changes.

[VirtuallyTaken]

The upgrade of their virtual infrastructure to esx35 update4 and vc to update5 didnt fix the issue. The CPU is still showing high cpu usage. We suggested upgrading to w2k3 but client has said thats not an option. Logged a call with vmware.. see what they say.

[VirtuallyTaken]

Spoke to vmware and they said to put the 'Virtualized MMU' to disable, ie. 'Forbid use of these features'. I had done that but the problem stayed for the next 2 days and then mysteriously disappeared.

Now, all the other Windows 2000 servers across different clients are showing the same issue. I have done the same as what VMware had recommended but this time the issue didnt mysteriously disappear. Its been over a week now. I have logged another call with Vmware to see what they say.

Cheers