VMware server 2.02 - guest suddenly powers off

Joshwaaa · 06-21-2010, 06:19 PM

Hi All,

Centos 5.5 - 2.6.18-194.el5 i686 athlon i386 GNU/Linux
VMware-server-2.0.2-203138.i386

AMD Athlon(tm) XP 1800+, 1.5 ghz, 1gig RAM

(This host currently has 1 guest which is running a simple X-less DNS & SMTP server that chews up a massive 62mhz and less then 128mb ram at any time.)

I have a problem where the guest suddenly powers off, the console error says: *** VMware Server internal monitor error *** - vcpu-0:VMM fault: regs=0x2e98, exc=3, eip=0x8f443.

This was happening randomly every couple of hours, some times after 4 or more, sometimes after 2.

I had heard that some weird things can happen when the host & guest time/date is not in sync. So I went away and installed NTP, host & guest then had the same time. I thought I has solved the problem because the guest was running for almost 2 days, ...but then it happend again ... (see error1log.txt @ 2010-06-21 20:01:05.112)

Do I need to sync the hw-clock time????

I then tried to power it on again a couple of times (see error2log.txt) and got a double fault: *** VMware Server internal monitor error *** - vcpu-0

oubleFault @ 0x4020:0xb58a9 (0x1000,0x2edc) (0x0, 0x0)

I have done a fair bit of googling on this and it seems to be a common problem across vmware server, esx, esxi where the guest suddenly powers off with a vcpu fault. Unfortunately I could not find any answers to the problem....

Maybe its just a bug and perhaps I need to write a cron script to just power the guest on if the state = off.....

First time posting, so please let me know if you need anything further.

Any help would be appreciated!!!

Cheers,

Joshwaaa

unSpawn · 07-17-2010, 06:58 PM

At least one thread and bug report elsewhere (Gentoo) suggests this to be a problem more often encountered running AMD. http://communities.vmware.com/message/1565011#1565011 says only RHEL-5.3 / Centos-5.3 are supported. But you probably already read that.

//NTLB

Joshwaaa · 07-21-2010, 03:53 AM

Thanks unSpawn,

yeah I found that out

... I have written a dodgy perl script to restart my vm if it happens.

I have synced my vm and vmware server with ntp and it dosent seem to be happening as much ... could be coincidence but I have read some strange things happen when they are out of sync.

Cheers,

Joshwaaa