HP Serviceguard Extended Distance Cluster for Linux A.12.00.00 Deployment Guide, March 2014

imposed by the Fibre Channel link for storage and Ethernet for networks. Storage in both data
centers is connected to both the nodes via two FC switches in order to provide multiple paths. This
configuration supports a distance up to 100 kms between datacenter1 and datacenter2.
Figure 4 Two Data Center Setup
Figure 4 shows a configuration that is supported with separate network and FC links between the
data centers. In this configuration, the FC links and the Ethernet networks are not carried over
DWDM links. But each of these links is duplicated between the two data centers, for redundancy.
The disadvantage of having the network and the FC links separate is that if there is a link failure
between sites, the ability to exchange heartbeats and the ability to write mirrored data will not be
lost at the same time. This configuration is supported to a distance of 10 kms between data centers.
Mirroring for the storage is configured such that each half of the mirror (disk set) will be physically
present at one datacenter each. Further, from each of the nodes there are multiple paths to both
of these mirror halves.
Also note that the networking in the configuration shown is the minimum. Added network connections
for additional heartbeats are recommended.
NOTE:
• If you are using the configuration as in Figure 3, you must set the Link Down Timeout
parameter value.
• If you are using the configuration as in Figure 3 with Virtual Machine (VM), you cannot set
the Link Down Timeout parameter value. This can also lead to start of disk having
non-current data during the package failover.
Therefore, HP recommends you to use the configuration as in Figure 4 for VM environment.
For more information about how to set a value for the Link Down Timeout parameter, see
“Setting the Value of the Link Down Timeout Parameter” (page 22).
16 Building an Extended Distance Cluster Using Serviceguard and Software RAID