Knowledge
This is the same as VCP5-DCV exam blueprint section 6.2
- Verify network configuration
- Use host profiles where possible to ensure consistent configuration
- Use vDS where possible to minimise configuration effort across multiple hosts
- Check configuration of Port Groups / dvPort Groups
- Check Load Balancing and Failover Policies
- Check Security Policies (Promiscuous Mode, Forged Transmits etc)
- Verify that VLAN settings are correct and consistent across a cluster
- Verify a given virtual machine is configured with the correct network resources. KB1003893 is a good resource for this but as a start:
- Check that the network associated with the VM’s vNIC exists, and the spelling is consistent across the infrastructure.
- Check that the “connected” checkbox for the vNIC is checked.
- Ensure the networking is configured correctly within the Guest OS
- Verify that the vSwitch has sufficient ports to support the VM
- Check the uplinks for the vSwitch are consistent (same VLANS)
- Troubleshoot virtual switch and port group configuration issues
- Correct spelling of port group names is important and is case sensitive. Consistency of naming and Security Configuration is vital for the smooth running of the infrastructure.
- Ensure switches are configured correctly (as per point 1) and with sufficient available ports.
- Troubleshoot physical network adapter configuration issues
- Ensure all physical NICs assigned to a vSwitch are configured with the same speed, duplex and VLANs on the physical switch
- If using IP Hash as the load balancing method, ensure Link Aggregation is configured on the switch.
- You can use CDP or LLDP to assist with network troubleshooting, it will identify the switch ports that are connected to each pNIC.
- Identify the root cause of a network issue based on troubleshooting information
- The root cause is likely to fall into one of 4 main areas:
- VM
- Port Group / vSwitch configuration
- Host Uplinks
- Physical Switch configuration
- Use the above notes to assist with determining the area at fault – working from the VM down is probably easiest.
- vmkping –D can be used to ping out through vmknics
- The root cause is likely to fall into one of 4 main areas:
Tools
- vSphere Networking Guide
- vSphere Troubleshooting Guide
- vSphere Command-Line Interface Concepts and Examples
- vSphere Web Client
- vSphere Client
- Some of the above has been pulled from VCP5-DCV study guides such as http://www.virtuallanger.com/2012/01/09/vcp-5-objective-6-2-perform-basic-vsphere-network-troubleshooting/ and http://blog.mwpreston.net/vcp-5/vcp-5-objective-6-2-perform-basic-vsphere-network-troubleshooting/