6.4

VMware, Inc. 17
Chapter 1 Introduction
 Reassigns the Principal (Public) IP address to the Secondary server and assigns the Primary its
unique management IP address.
 Makes the newly active server visible on the network. The newly active server begins to intercept and
queue disk I/O operations for the newly passive server.
4 vCenter Server Heartbeat causes the newly passive server to begin accepting updates from the active
server.
5 vCenter Server Heartbeat starts the same protected applications on the new active server. The protected
applications become accessible to users. The managed switchover is complete
Automatic Switchover
Automatic switchover (auto-switchover) is similar to failover (discussed in the next section) but is triggered
automatically when system monitoring detects failure of a protected application.
Like managed switchover, auto-switchover changes the server roles but then stops vCenter Server Heartbeat
on the previously active server to allow the administrator to investigate the cause of the auto-switchover and
verify the integrity of the data.
After the cause for the auto-switchover is determined and problems are corrected, the administrator can use
vCenter Server Heartbeat Console to return the server roles to their original state.
Automatic Failover
Automatic failover is similar to automatic switchover (discussed above) but is triggered when the passive
server detects that the active server is no longer running properly and assumes the role of the active server.
Figure 1-6. Failover
During the automatic failover, the passive server performs the following steps:
1 Apply any intercepted updates currently in the passive server’s receive queue as identified by the log of
update records that are saved on the passive server but not yet applied to the replicated files.
The amount of data in the passive server’s receive queue affects the time required to complete the failover
process. If the passive server’s receive queue is long, the system must wait for all updates to the passive
server to complete before the rest of the process can take place. An update record can be applied only if
all earlier update records are applied, and the completion status for the update is in the passive server’s
receive queue. When no more update records can be applied, any update records that cannot be applied
are discarded.
2 Switch mode of operation from passive to active.
This enables the public identity of the server. The unique management IP address is removed from the
passive server and the shared Principal (Public) IP address is assigned. The passive server becomes active
and available to clients that were connected to the previously active server before the automatic failover
and clients are able to reconnect.