I've had this a couple of times now in the last few weeks. I wanted to try out the Linux KVM virtualization hypervisor that comes with Fedora 11 since VMware Server 2 is giving my fits on F11 x86_64 so I installed qemu and all the supporting packages via yum.

I realized that I can't do some things with KVM that I can with VMware Server so I'm putting up with the signal 6 crashes every week or so and went back to VMware Server 2.0.1 for Linux. In order to get that running I had to apply some patches to get the kernel modules compiled but that's been done for a while.

BUT, I forgot to uninstall the KVM RPMs so when I reboot the kvm_intel and kvm kernel modules are loaded. KVM and VMware don't play nice together so when you try to boot a VMware VM with those modules loaded the startup process hangs at 95% and eventually times out.

The VMware logs don't really give much help in pointing out the problem. All you really get is something in /var/log/vmware/hostd.log that looks something like this:


[2009-08-20 10:38:43.138 'Libs' 140336295930128 info] connect to /var/run/vmware/root_0/1250779071591734_12917/testAutomation-fd: File not found
[2009-08-20 10:38:43.638 'Libs' 140336295930128 info] connect to /var/run/vmware/root_0/1250779071591734_12917/testAutomation-fd: File not found
[2009-08-20 10:38:43.638 'Vmsvc' 140336303999248 info] Failed to do Power Op: Error: (3008) Cannot connect to the virtual machine
[2009-08-20 10:38:43.639 'vm:/home/vmware/myws/myws.vmx' 140336303999248 warning] Failed operation
[2009-08-20 10:38:43.639 'ha-eventmgr' 140336303999248 info] Event 5 : Failed to power on myws on myhost.domain.com in ha-datacenter: A general system error occurred


If you see something like this in your hostd.log files, make sure that the KVM kernel modules aren't loaded. I removed them as such:


modprobe -r kvm_intel
modprobe -r kvm


Once you do that your VMs should start fine.

Happy VMing,
Flux.