Larger Clusters

You can create clusters of up to 16 nodes with Serviceguard. Clusters of up to 16 nodes may be built by connecting individual SPUs via Ethernet; and you can configure up to 8 systems as an Serviceguard cluster using FDDI networking.

The possibility of configuring a cluster consisting of 16 nodes does not mean that all types of cluster configuration behave in the same way in a 16-node configuration. For example, in the case of shared F/W SCSI buses, the practical limit on the number of nodes that can be attached to the same shared bus is four, because of bus loading and limits on cable length. Even in this case, 16 nodes could be set up as an administrative unit, and sub-groupings of four could be set up on different SCSI buses which are attached to different mass storage devices.

In the case of non-shared F/W SCSI connections to an XP series or EMC disk array, the four-node limit does not apply. Each node can be connected directly to the XP or EMC by means of two F/W SCSI buses supporting PV links. Packages can be configured to fail over among all sixteen nodes. For more about this type of configuration, see “Point to Point Connections to Storage Devices,” below.

	NOTE: When configuring larger clusters, be aware that cluster and package configuration times as well as execution times for commands such as cmviewcl will be extended. Some command options help to remedy this situation. For more information, refer to the man page for cmquerycl.

Active/Standby Model

You can also create clusters in which there is a standby node. For example, an eight node configuration in which one node acts as the standby for the other seven could easily be set up by equipping the backup node with seven shared buses allowing separate connections to each of the active nodes. This configuration is shown in Figure 2-10 “Eight-Node Active/Standby Cluster ”.

Figure 2-10 Eight-Node Active/Standby Cluster

Point to Point Connections to Storage Devices

Some storage devices allow point-to-point connection to a large number of host nodes without using a shared SCSI bus. An example is shown in Figure 2-11 “Eight-Node Cluster with XP or EMC Disk Array ”, a cluster consisting of eight nodes with a Fibre Channel interconnect. (Client connection is provided through Ethernet.) The nodes access shared data on an XP 256 or EMC disk array configured with 16 I/O ports. Each node is connected to the array using two separate F/W SCSI channels configured with PV Links. Each channel is a dedicated bus; there is no daisy-chaining.

Figure 2-11 Eight-Node Cluster with XP or EMC Disk Array

Fibre Channel switched configurations also are supported using either an arbitrated loop or fabric login topology. For additional information about supported cluster configurations, refer to the HP Unix Servers Configuration Guide, available through your HP representative.