Table of Contents
Advertisement
Chapter 11: System Platform High
System Platform High Availability Failover overview
The System Platform High Availability Failover is an optional feature. This feature offers the
following capabilities:
Node scores
System Platform High Availability Failover uses node score to compute the ability of every
machine to run the resources and decide on which node resources should be running for each
particular situation. In the case there is no issue with the system and resources could be
running on any node, both machines have the same score. Thus System Platform uses a term
of "preferred node" – machine that is supposed to run the resources in the case there is no
issue with the system. This preferred node has a small score benefit. So if both machines are
booted up at the same time, this node will run resources.
No auto-failback
System Platform High Availability Failover does not use a feature called auto-failback that
migrates resources back to the preferred node in case it reappears again and resources are
currently running on the standby node. This is because there is a service impact to switching
servers and if both servers are healthy then running on the preferred node offers no increased
benefit. If the user wants to migrate resources back to the preferred node after a failover or a
switchover, the user can do so using the Manual Switchover option in Failover menu during
the best suitable time.
Note:
Remote reboot (also called STONITH) is disabled by default and not supported in the current
version of System Platform High Availability. Remote reboot should not be configured when
using High Availability.
Expected failover/switchover times
System Platform High Availability Failover uses 30 seconds as a timeout interval of lost ping
replies upon which standby node will declare active node as dead (although it can be not
accessible, not running or blocked). When that period expires, the standby node executes a
takeover (that is, starts resources). Note that System Platform does not provide any web
interface to modify this interval.
In case of manual switchover or in case there is some state of the system that initiates
preemptive failover, the total time between start of the command and time when all resources
Installing and Configuring Avaya Aura
Availability Failover
™
System Platform
April 2010
83
Hide quick links:
Advertisement
Table of Contents
Troubleshooting
