Pages

Thursday 11 February 2010

Understanding VMware Fault Tolerance

VMware Fault Tolerance (FT) provides continuous availability for applications by creating a live shadow instance of a virtual machine that is in virtual lock-step with the primary instance. By allowing instantaneous failover between the two instances in the event of hardware failure, VMware Fault Tolerance eliminates even the smallest of data loss or disruption.


VMware Fault Tolerance (FT) works by creating an identical copy of a virtual machine. One copy of the virtual machine, called the primary, is in an active state, receiving requests, serving information and running applications. Another copy, called the secondary, receives the same input that is received by the primary.

In an FT environment, one virtual machine runs as a primary and FT runs a secondary virtual machine on a different ESX host. The secondary virtual machine shares the primary's virtual disks. The virtual machines are kept in lock-step via logging information sent over a private network connection. The primary is the sender of this logging information and the secondary only listens. FT is based on VMware Record/Replay technology.


In the event that the primary's ESX host fails, the secondary virtual machine takes over without interrupting applications.
VMware FT provides more continuity than VMware HA because FT does not require a virtual machine restart and the secondary virtual machine immediately comes online with all, or almost all state information preserved.
Virtual machines protected by FT are not handled by VMware HA for restart priority. It is considered disabled in the restart priority.


Determining Node Failure
VMware FT uses network heartbeats to determine when primary and backup hosts are down. Backup goes live and becomes the new primary if it declares the current primary dead.
FT uses an atomic operation on the shared VMFS to distinguish between a failed host from a network failure.

VMware FT Migration Transition States
VMware FT tracks failover operations with a variety of states. For more information on transition states, see - VMware Fault tolerance migration transition states (1010634).

No comments:

Post a Comment