Virtual Machine periodically fails to respond to ping

Issues:

1)      Users periodically complain of connection problems to a Guest OS.

2)      If you continually ping the Guest OS in question you may notice periods where the Guest OS fails to respond

3)      In the vmware.log file located in the folder containing the guest OS you may file errors similar to the following.

mks| SOCKET 16 recv error 5: Input/output error

mks| SOCKET 16 destroying VNC backend on socket error: 5

mks| SOCKET 17 recv error 5: Input/output error

mks| SOCKET 17 destroying VNC backend on socket error: 5

 

Cause:

Issue 1

According to VMware the problem is caused by an issue in the LSILOGIC scsiport driver.  The condition arises when the SCSI Virtual Adapter is entering standby mode at the same time an interrupt is still being processed.  This may cause the OS to hang for short periods of time.

 

Issue 2

This issue can also be cause by multiple etherchannel bundles being used in the same vSwitch and associated port groups.

 

Solution:

Issue 1

You can configure the LSILogic driver to handle this condition differently using the following.

 

1)      Shutdown the guest OS and Power off the virtual machine

2)      Edit the Virtual Machine vmx file

a.       Insert the following just before the line scsi0.present = "true"

lsilogic.reflectIntrMask = “true”

b.      Save the file

3)      Power on the VM

 

Issue 2

You can only have one etherchannel bundle configured per vSwitch and associated port groups within that vSwitch.  If you have multiple etherchannel bundles assigned to a vSwitch you should remove all but one.

 

Etherchannel bundles can not be used in either an Active/Active or Active/Failover configuration.

 

On the Cisco switches you will typically see error messages related to port flapping if this is the issue.