This is a known issue with U3.

Title:
A Virtual Machine may unexpectedly reboot when using VMware HA with Virtual Machine Monitoring on ESX 3.5 Update 3

Symptoms:
Virtual Machines may unexpectedly reboot after a VMotion migration to an ESX 3.5 Update 3 Host OR after a Power On operation on an ESX 3.5 Update 3 Host, when the VMware HA feature with Virtual Machine Monitoring is active.

Problem Statement:
A virtual machine may reboot itself when the following conditions exist:

• Virtual Machine is running on ESX 3.5 Update 3 Host, either by virtue of VMotion or a Power On operation, and
• Host has VMware HA enabled with the "Virtual Machine Monitoring" option active.

Resolution:
Note: Before you begin please refer to KB 1003490 for important information on restarting the mgmt-vmware service.

To work around this problem:

*Option 1: Disable Virtual Machine Monitoring *

1. Select the VMware HA cluster and choose Edit Settings from the right-click menu (note that this feature can also be enabled for a new cluster on the VMware HA page of the New Cluster wizard).
2. In the Cluster Settings dialog box, select VMware HA in the left column.
3. Un-Check the Enable virtual machine monitoring check box.
4. Click OK.

1. Disconnect the host form VC (Right click on host in VI Client and select "Disconnect" )
2. Login as root to the ESX Server with SSH.
3. Using a text editor such as nano or vi, edit the file /etc/vmware/hostd/config.xml
4. Set the "heartbeatDelayInSecs" tag under "vmsvc" to 0 seconds as shown here:
   <vmsvc>
   <heartbeatDelayInSecs>0</heartbeatDelayInSecs>
   <enabled>true</enabled>
   </vmsvc>
5. Restart the management agents for this change to take effect. See Restarting the Management agents on an ESX Server (KB 1003490).
6. Reconnect the host in VC ( Right click on host in VI Client and select "Connect" )