Understanding and Designing Serviceguard Disaster Recovery Architectures

Support for Cross-Subnet Configurations in Extended Clusters
Beginning with the Serviceguard A.11.18 patches, PHSS_37094 (11i v2) and PHSS_37095 (11i
v3), Cross-Subnet configurations are supported. This allows the nodes in each data center to
configure their heartbeats on subnets that are locally unique to their own data centers. The following
restrictions apply for Cross-Subnet configurations:
• All nodes in the cluster must belong to the same network domain (that is, the domain portion
of the fully-qualified domain name must be the same).
• The nodes must be fully connected at the IP level.
• A minimum of two heartbeat paths must be configured for each cluster node.
• Each heartbeat subnet on each node must be physically routed separately to the heartbeat
subnet on another node; that is, each heartbeat path must be physically separate:
◦ The heartbeats must be statically routed; static route entries must be configured on each
node to route the heartbeats through different paths.
◦ Failure of a single router must not affect both heartbeats at the same time.
• CVM and CFS require link-level traffic communication (LLT) among the nodes, therefore Extended
Clusters cannot be configured in Cross-Subnet configurations with CVM or CFS. As a result,
only Extended Clusters using LVM volume groups are supported in Cross-Subnet configurations.
• Oracle RAC requires a common subnet for the RAC interconnect between nodes, therefore
Extended RAC Cross-Subnet Clusters are not supported.
• Each package subnet must be configured with a standby interface on the local bridged net.
The standby interface can be shared between subnets.
Configuration Requirements for all Extended Distance Clusters
The following configuration requirements are common for all Extended Distance Cluster and
Extended Cluster for RAC configurations:
• Refer to Table 2 (page 51), Table 3 (page 51) and Table 4 (page 53) for information about
the Products, Product revisions and Volume Managers supported in Extended Clusters.
• The network interfaces used for cluster nodes must support DLPI
1
. The network switches used
in the configuration must also support DLPI. DLPI support is not required for Quorum Server
nodes.
• There can be separate TCP/IP networking links and Fibre Channel data links between the
data centers, or both TCP/IP networking and Fibre Channel data can go over common WDM
2
(Wave Division Multiplexing), SONET or SDH links between the data centers. See the section
below “TCP/IP Network and Fibre Channel Data links between the data centers” for more
detail.
• A minimum of two heartbeat paths must be configured for all cluster nodes. The preferred
solution is two separate heartbeat subnets configured in the cluster, each going over a
separately routed
3
network path to the other data centers. Or, there can be a single dedicated
heartbeat subnet with Primary and Standby LANs configured for it, each going over a
separately routed3 network path to the other data centers. For clusters using CVM 3.5, only
a single heartbeat subnet is supported, so you must have both Primary and Standby LANs
configured for the heartbeat subnet on all nodes. Multiple heartbeat subnets are supported
with LVM, SLVM, VxVM, CVM 4.1, CFS 4.1, CVM 5.0, CFS 5.0, CVM 5.0.1, and CFS 5.0.1.
For EC RAC clusters, HP recommends that you to have an additional network for Oracle RAC
cache fusion traffic. It is acceptable to use a single Standby network to provide backup for
both the heartbeat network and the RAC cache fusion network however, it can only provide
failover capability for one of these networks at a time. The RAC CRS heartbeat must go over
Extended Distance Cluster on HP-UX 55