
Failover
(172.16.0.181) and floating IP 172.16.0.219.
l F/O Group 2 - has subnet 192.168.0/24 with cluster cl02 (192.168.0.211), server sv02
(192.168.0.181) and floating IP 192.168.0.219.
l If the clusters are using spoof and sv03 (192.168.0.182) is added to the server pool for cluster
cl01, this will cause F/O Groups 1 and 2 to be merged into a single F/ O Group that includes
cl01, cl02, sv01, sv02, sv03, and subnets 172.16.0/24 and 192.168.0/24.
e. Then, change the configuration such that the single F/O group is split into 2 F/O Groups.(e.g.,
by deleting sv03 from cl01, above.)
Verify that:
l All clusters can pass traffic.
l Failover occurs as expected if one Peer fails.
l The F/O Groups are rebalanced appropriately when the rebalance command is executed.
Configuring N+1 Failover Between EQ/OS
10 Systems
N+1 Failover is a feature of EQ/OS 10 where the failover configuration consists of multiple active peers ("N") plus
1 passive peer. In this type of failover configuration, the Equalizer clusters are instantiated on all "N"peers and
organized into failover groups. If the passive, or backup peer's connectivity for a failover group's resources is
judged to be "healthier" that the peer on which the group is running, then the group fails over to the passive peer,
which becomes the Primary peer.
N+1 failover provides the ability to configure up to 4 Peers in a failover configuration. All peers will be
"heartbeating" with each other and all synchronizing configuration. If a Peer "fails", for each failover group (F/O)
group affected by the failure, one and only one of the other Peers will take over the F/O Group based on:
1. Which Peer in the F/O group has the best connectivity to servers, routers, etc. In case of a tie amongst one
or more Peers, the Peer with the greatest System ID hex value or " sysid" will take over the F/O group. For
example, if 2 peers have the same level of connectivity with servers, routers, etc, and one "sysid" is
"003048BC2C8A" and the other is "003048D52AA2". The second "sysid" has a higher hex value and will
take over the F/O group.
Note - All flavors of Microsoft© Windows include a hex calculator. A free, downloadable Hex Calculator widget
for Mac OX is also available.
2. The user can subsequently adjust the load across the Peers by managing the preferred Peer for each F/O
Group and executing the rebalance function.
Note - Currently, if a failover event occurs, all F/O Groups are moved when a failover event occurs, even if it
only affects a subset of the F/O Groups. Failover occurs on a F/O Group basis. For example, if an interface goes
down, only the affected F/O Group(s) will be moved.
460
Copyright © 2013 Coyote Point Systems. A subsidiary of Fortinet, Inc.
Comentarios a estos manuales