Citrix ByteMobile T1010, Установка, настройка и администрирование

Страница: 80 / 172

3-6 T-Series Traffic Director Installation, Configuration and Administration Guide
Chapter 3 Basic T1000/Unison Load-Balancing Configuration
IP Addresses for High Availability
Bytemobile recommends a high availability (HA) deployment model with two T1000s, as
shown in Figure 3-2. One T1000 works as the primary node accepting all connections, while
the second T1000 works as a standby node. If the primary T1000 (known as T1000-1, or
preferred master ) fails, the secondary T1000 (T1000-2) takes over. Failure may be triggered
by any of the following events:
• Failure of a next-hop router.
• Failure of an external port or link.
• Failure of the primary T1000 itself.
The secondary T1000 monitors the primary by sending periodic healthcheck messages to
determine whether the primary T1000 is accepting connections. If a healthcheck fails, the
secondary node retries the connection for a specified period, after which it determines that
the primary T1000 is not functioning normally. The secondary T1000 then takes over for the
primary (a process called
failover ). After a failover, all subscriber and server connections
must be reestablished, but session persistence rules are maintained as they were before the
failover.
In a high availability deployment, the primary T1000 “owns” all virtual IP addresses. When
a failover occurs, the secondary T1000 becomes primary and then uses gratuitous ARP
(GARP) to advertise the floating IP addresses acquired from the offline T1000. Without
additional configuration, the new primary T1000 will use the MAC addresses of its own
interfaces.
In some instances, certain devices will not accept the GARP messages generated by the
T1000. As a result some external devices will retain the IP to MAC mapping of the offline
T1000, resulting in loss of service. To address this problem, you may configure identical
virtual MAC (VMAC) addresses on both T1000s. Therefore, when failover occurs, ARP
tables on external devices do not need to be updated.
When a failover is triggered, T1000-2 will be declared master in the case of a failure that
affects only T1000-1. Once the failure is addresses, a voluntary
failback is triggered, and
T1000-1 is declared master again.
VLAN Assignments
Note: IP addresses and VLAN IDs in this document are for example purposes only.
Use appropriate values for your network deployment.
In a load balancing deployment, VLANs separate ingress, egress, and management traffic.
The sample network in Figure 3-2 uses the following VLAN IDs:
• Ingress VLANs:
– VLAN ID 100 is the OSN ingress VLAN. It is configured on the T1000 ports as
well as all OSN internal ports. Also used for blade healthcheck purposes.
– VLAN ID 151 is the T1000 ingress VLAN. It is configured on the T1000 ports and
on the wireless-side router.