Yesterday I rebuilt my Ubuntu 9.10 64 bit machine to 10.04 64 bit. All went well except for the nVidia driver which crapped out after I had finished installing 90% of my software. A couple of uninstall/reinstalls and a few good kicks and it is working. Much improved over 9.10.

So this morning I fired up an XP VMware guest in player 3.0.1 build-227600 newly installed on the Ubuntu host. I built the guest last year on Workstation 6.5 on Ubuntu 8.04 and have been running it on 9.10 for the last 6 months with no issues. However, when run on the 10.04 host the VM pegged 2 of the 8 host CPUs and slowed to a crawl. It did not do much for host performance either.

I found a couple of links discussing the issue. They are listed below:

http://www.loadyourself.com/linux/hi...files-on-ntfs/

http://it.thelibrarie.com/weblog/?p=315

The second one did the trick for me. Basically I did the following:

- downloaded the Intel network adapter driver here http://support.intel.com/support/net.../cs-006120.htm
- copy the driver to a flash drive which will be accessible to the guest
- added this line to the vmx file for the guest - started the guest - note that it has not network capability at this point
- copied the Intel driver from the flash drive to the guest and installed it.

All is now well. Thanks to the original poster. I will leave this thread open for a few days for comments and then mark it resolved.

Ken

I spoke too soon The fix seemed to work for a little while then back to the same problem. I copied a backup copy of the XP VM into place and it worked OK - for a while. Seems to be related to suspending the host with the VM running (perhaps) need to do some more testing.

Ken

Well here is what I have found so far... It I start from a cleanly rebooted host and then start the XP VM I get a message It seems to make no difference if I choose to enable or not. The VM works OK.

n.b. The guest is configured to use 2 CPUs.

If I suspend the host while the VM is running and then bring the host back the VM is pegged at close to 200% CPU (2 CPUs at 100%) and it stays that way.

If I shut down the VM then suspend and bring back the host and restart the VM I see the same maxed out CPU situation.

Just for the halibut I switched the VM to 1 CPU. The problem has gone away and the performance of the VM seems to be as good as with 2 CPUs - although I have not done any formal benchmarking.

Ken

Identical issue with 64-bit Lucid host, 32-bit XP guest and vmWare 7.1: single-CPU VMs (tried 2 different ones and a variety of settings) run normally, multi-CPU VMs are dog slow (10x at least). Will investigate further to see if this is really related to host suspend/resume -- thus far I was indeed running on a host that's been suspended a few times before.

Max, Try setting the guest to 1 cpu. The performance does not seem to suffer greatly and the "peg cpu after suspend" issues seems to go away.

Ken

Oh, I tried that right away when I encountered this story of yours and it did indeed help. Here's the interesting part: after I did a clean reboot, the issue did not yet reproduce again. Tried suspending the laptop with VM running as well as without -- still going strong on 2 [virtual] CPUs. This bug is fixing to be a hard one to sell to vmWare

...and half a dozen suspends/resumes later, the issue resurfaces. It's there all right, but intermittent

Well I do not know how, why or by whom but this issue appears to be fixed. Had occasion to add some horsepower to an XP VM for a particular task. Rebooted the host, bumped up the VM to 4 CPUs and completed the task at hand. Just for the halibut I shut down the VM, suspended and brought back the host and launched the VM. No CPU hogging - Hooray!

Ken