Workstation 5.5.3 build-34685,
Suse 10.2 x86-64 Host,
Vista guest,
Dell 9400 laptop Core 2 Duo 2gig ram.
Very often (so often as to be virtually unusable) VMware will
cause my entire X server to freeze such that the Keyboard and Mouse
are totally unusable.
The guest machine is totally unresponsive and no activity
is shown
The HOST is still running, clock in task bar updating but it
will not respond to the mouse or keyboard.
GKrellm shows processor utilization dropping to near zero,
and bursts of intense disk activity lasting close to 30 seconds
followed by a lull, then more intense disk activity.
Reading other threads, I have my VM locked to a single cpu
(because if allowed to migrate as it wishes processor utilization
jumps to 100% on both cpu). My cpu is locked on high performance
so they will not be throttling up and down.
If I do nothing in the guest, it does not lock up the machine.
But any activity in the guest will lock the Xserver solid (or
prevent any input - gkrellm still shows changing activity so
obviously video output still works.
SSH into the laptop from another machine still works.
After 5 minutes the lock will clear and I can continue
working in the vm for another 2 to 6 minutes.
What is this problem, and when will it be fixed. I see other
references going back almost 6 months on 5 minute freezes.
Regarding updates for dual core machines, my personal recommendation is to get ALL of them. I have seen weird things like Lotus Notes not working correctly, VPN lockups, etc. that were cleared up by downoloading and installing specifically BIOS updates, but I would get all available updates, if possible, as they tend to fix such anomalies.
Oh, I wasn't trying to suggest any of the above were viable workarounds or options.
I didn't really intend to try this over the weekend, but when I went to reload my laptop OS for other reasons I inadvertantly grabbed a Kubuntu 7.04 (Feisty) CD instead of the Kubuntu 6.10 (Edgy) CD and so far I have not been able to get my Windows XP VM to suffer the 5 minute pause issue.
The major difference I believe between the two versions is that the Feisty release is using a 2.6.20 variant instead of the 2.6.17 variant found in Edgy. I also believe a bug was found in the TSC handling in the 2.6.19 series and corrected. Not trying to say it is a solution, just reporting a observation.
This issue has sucked up too much of my time so I just decided to ditch my Thinkpad T60's SMP support and boot Ubuntu from the 2.6.17-11-386 kernel (I just moved it to the top of the list, in front of the 2.6.17-11-generic in grub.lst).
Everything works fine and my win XP VM works fine now. Performance is good enough and can my work done now. Hopefully whatever fix was put into to the 2.6.20 kernels will eventually end up as an update to Edgy?
Not only is this an accurate statement:
"Disabling acpi is NOT an option for any modern acpi based machine. ...scare off evil spirits. "
It is now part of my signature file
Please see Petr's reply in the following posting
http://www.vmware.com/community/thread.jspa?threadID=77895&start=15&tstart=0