Service Console redundancy for VMware High Availability
Article ID: 344377
Updated On:
Products
Issue/Introduction
An ESX host in a VMware HA cluster considers itself isolated, and powers off all virtual machines running on it (so that other members of the cluster can power on those virtual machines), if:
- it cannot ping any other ESX host in the cluster
- it cannot ping its default gateway
- it does not receive heartbeats from any other hosts in the cluster
Since VMware HA only monitors the service console, based on cluster settings, the ESX host will either shutdown, power off, or leave powered on virtual machines running on it if it loses connection to its service console (due to a bad cable, bad NIC, physical switch port problem, or any other reason). This happens even if the virtual machine network is functioning properly because VMware HA does not monitor the virtual machine network. This leaves many environments vulnerable to have their virtual machines shut down in the event that the service console network goes down.
Environment
VMware ESX 4.0.x
VMware vCenter Server 4.1.x
VMware ESX 4.1.x
VMware ESX Server 3.0.x
VMware ESX Server 3.5.x
VMware vCenter Server 4.0.x
Resolution
To overcome this issue, you can configure an additional service console for redundancy. Having an additional service console on a separate vSwitch, bound to a different vmnic which is attached to a different physical switch, ensures that VMware HA does not power off any virtual machines if the primary service console network goes down. If primary service console network fails, it continues functioning by using the second service console.
For a complete guide to setting up service console redundancy, see VMware HA: Concepts and Best Practices.
For more information about the configuration of a static route for the redundant service console, see Adding a static route to an ESX host for an extra Service Console interface (1005212).
When the secondary Service Console is set up, you must configure multiple VMware HA isolation addresses. For more information about setting multiple isolation response addresses, see Setting Multiple Isolation Response Addresses for VMware High Availability (1002117).
Note: This workaround reduces, but does not eliminate, the event of false positive VMware HA results.
Feedback
