VMWare Server problem

Hi Guys,

Here is the output from hostd.log on a VMWare server, sitting on top of CentOS. VM's running on this machine are shutting down sporadically - maybe one every 24 hours? Can anyone make any sense from the log file?

Versions: VMware server 2.0.1 Centos 4.8

Guest operating systems that have crashed so far are Windows XP & Centos 5.4.

Cheers, Dave

From serverfault SuperBOB

• Hi Dave

  At the first glance I would tell there is something wrong with your data partition (/data/data4). What kind of storage are you using? local disks, raid, nfs? Do you have any snapshots created for the vm's?

  If possible try to enable verbose logging somehow. I didn't find any VMware KB for the VMware Server 2 but there should be a configuration file in /etc or /opt.

  As always the question: why are you using VMware Server and not VMware ESXi? VMware Server 2 is not under development anymore and it's pretty slow if you compare it to ESXi.

  greetings grub

  SuperBOB : Hi Grub, Cheers for the reply - just settling into a new role, will be migrating them over to ESXi or Xen in due course. Its local disks, non-raided. No snapshots. Ill try to increase the logging and check for any disk errors around the times of the shutdowns. Cheers, Dave
  From grub

• How are you for memory? This can happen if you run out.

  SuperBOB : Hi Chris, Load on the host is high, here's a quick copy paste of the memory available & used. Mem: 8053316k total, 7134768k used, There is always plenty of swap available though, Cheers, Dave

  Chris Thorpe : Is there anything in the rest of the logging that mentions a 'Signal 6'? If so, check this out: http://communities.vmware.com/thread/242151?tstart=0 (not CentOS but poss. related. Core of it seems to be a disk controller incompatibility). You may alternatively be hitting this: http://bugs.centos.org/view.php?id=3884

  SuperBOB : I have indeed found some Signal 6 errors, thanks researching that more now.
  From Chris Thorpe