This chapter describes how to configure and manage the GRHA feature.
GRHA is a high availability solution that works across two datacenters
with a pair of servers connected over a routable network. When the
standby AE Services server is activated, AE Services will start providing service approximately a minute after
the failure detection interval is over. However, all AE Services clients, other than DMCC clients have to re-establish
all monitors/registrations as if they are working with an AE Services server that just came up after a reboot.
 | Important |
|
In the secure mode, before configuring GRHA, new Linux users that
are created on the primary server must be created on the secondary
server as well.
|
 | Note |
|
The IP network
round trip time between two AE Services servers must be less than 100 milliseconds.
GRHA is now supported on the AE Services 7.0 for VMware only. The use of Machine Preserving High
Availability (MPHA) and Fast Reboot High. Availability (FRHA) mode
are not available for AE Services 7.0.
|
To use the GRHA feature, perform the following steps:
Make sure you
have the appropriate information for the GRHA feature. See Configuration worksheet for Geo Redundant High Availability.
Configure and
start GRHA, see Configuring Geo Redundant High Availability.
 | Note |
|
You cannot change
the AE Services IP address or hostname once GRHA is enabled. To change
the hostname or IP address on either AE Services server, you must first disable GRHA.
If you performed
a backup when Geo Redundant High Availability was enabled and then
restored that backup file when GRHA is removed, you must remove the
GRHA configuration from the AE Services Management Console to be able to enable HA. See Removing Geo Redundant High Availability.
You cannot change the AE Services Virtual IP addresses once GRHA is enabled. To change the
Virtual IP addresses on either AE Services server, you must first disable GRHA.
The Virtual IP addresses and both AE Services server IP addresses must be in the same subnet. For two
data centers with different subnets, one possible solution is to use
an extended subnet between the data centers.
|