The Memory Channel API is not supported for data transfers larger than 8K bytes when loopback mode is enabled in two member clusters configured with MC virtual hub. For more information on loopback mode go to http://www.tru64unix.compaq.com/docs/updates/TCR51A/TITLE.HTM and refer to the 21.Mar.2002 Technical Update titled "MC API Applications May Not Use Transfers Larger Than 8 KB with Loopback Mode Enabled on Clusters Utilizing Virtual Hubs."

For more information about DRM, see the SANworks Data Replication Manager QuickSpec. For more information about the Campus-Wide DisasterTolerant Product, see http://www.compaq.com/solutions/customsystems/ha/tru64_campus_clusters.html.

1. Can be expanded to two TZ89N-AV drives with part number DS-TL892-UA.
2. A DS-TL890-NE/NG MiniLibrary Expansion Unit can be connected to up to three DS-TL891-NE/NG drive units to manage the drives and cartridges in all connected units. A DS-TL800-AA pass-through mechanism is required for the second and third DS-TL891-NE/NG.
3. Can be expanded to two DLT 35/70 drives with part number 120878-B21.
4. The MiniLibrary Expansion Unit can be used to control the drives and cartridges of up to five drive and data units. A MiniLibrary Pass-Through Mechanism, part number 120880-B21, is needed for each additional unit beyond the first drive unit.
5. Can be expanded to two DLT 20/40 drives with part number 128671-B21.
6. An upgrade kit to add one DS-TZ89N-VA tape drive.
7. An upgrade kit to add one 35/70 DLT tape drive.
8. Can be upgraded by the addition of a single or multiple 35/70 DLT tape drives with part number 146209-B21.

The Memory Channel application programming interface (MC API) is not supported in a two-member cluster containing an AlphaServer ES45 and configured with Memory Channel in virtual hub mode. Everything else in the TruCluster product is supported with ES45 in vhub mode, but to deploy an application that uses the MC API in a two-member cluster containing an AlphaServer ES45, a Memory Channel hub must be configured for each rail.

TruCluster Server Version 5.1A requires that additional software subsets be installed. See the TruCluster Server Software Installation manual for more information.

• At least 50 percent free space in root (/), cluster_root#root
• At least 50 percent free space in /usr, cluster_usr#usr
• At least 50 percent free space in /var, cluster_var#var, plus an additional 425 MB to hold the subsets for the new version of the Tru64 UNIX operating system
• At least 50 percent free space in /usr/i18n, cluster_i18n#i18n when used

• Prestoserve NVRAM failover is not supported on shared disk devices.

• Switched Fibre Channel: Eight-member systems may be connected to common storage over Fibre Channel in a fabric (switch) configuration.
• Parallel SCSI: Only four of the member systems may be connected to any one SCSI bus. Multiple SCSI buses may be connected to different sets of members and the sets of members may overlap. Use of a DS-DWZZH-05 UltraSCSI hub with fair arbitration enabled is recommended when connecting four member systems to a common SCSI bus.

• Hard partitions of GS80, GS160, and GS320 can be clustered together either across separate systems or within systems, but each hard partition must have at least one cluster interconnect connection.
• TruCluster Server Version 5.1A requires all members to be connected with either all members using Memory Channel hardware or all members using a private 100 MB or 1000 Mbps full-duplex LAN. Note that there are two variants of Memory Channel (MC):
• ES45 supports single rail Mermory Channel (MC) configured for either 512 MB or 128 MB, and it supports dual rail MC only at 128 MB.
• The use of the MC API is not supported in a two-member cluster containing an AlphaServer ES45 and configured with Memory Channel in virtual hub mode. To use the MC API in a two-member cluster containing an AlphaServer ES45, a Memory Channel hub must be configured for each rail.
• For further information on deploying an ES45 in a TruCluster, please refer to the Cluster Hardware Configuration Technical Update for the AlphaServer ES45 which resides in the repository found at http://www.tru64unix.compaq.com/docs/pub_page/tcr_update.html

Note: For the current status of supported TruCluster HSV110 configurations, please contact TruCluster Product Manager Bob Grosso at bob.grosso@compaq.com or by phone at 1-603-884-1403.

• The TruCluster Server Cluster Hardware Configuration manual provides hardware configuration examples.

• Switched Fibre Channel: Eight-member systems may be connected to common storage over Fibre Channel in a fabric (switch) configuration.
• Parallel SCSI: Only four of the member systems may be connected to any one SCSI bus. Multiple SCSI buses may be connected to different sets of members and the sets of members may overlap. Use of a DS-DWZZH-05 UltraSCSI hub with fair arbitration enabled is recommended when connecting four member systems to a common SCSI bus.

• 340654-001 Fibre Channel Tape Controller
• 152975-001 Fibre Channel Tape Controller II
• 163082-B21 3R-A2673-AA Fibre Channel Modular Data Router (MDR)
• 163083-B21 3R-A2774-AA Fibre Channel Modular Data Router (MDR)
• 218240-B21 3R-A3292-AA Fibre Channel Modular Data Router (MDR)
• 218241-B21 3R-A3312-AA Fibre Channel Modular Data Router (MDR)

Note: For the current status of supported TruCluster HSV110 configurations, please contact TruCluster Product Manager Bob Grosso at bob.grosso@compaq.com or by phone at 1-603-884-1403.

• Hard partitions of GS80, GS160, and GS320 can be clustered together either across separate systems or within systems, but each hard partition must have at least one cluster interconnect connection.
• TruCluster Server Version 5.1A requires all members to be connected with either all members using Memory Channel hardware or all members using a private 100 MB or 1000 Mbps full-duplex LAN. Note that there are two variants of Memory Channel (MC):

Cluster members in a given cluster must all use Memory Channel or all must use LAN, but these interconnects cannot be mixed for cluster communication in the same cluster. Using multiple shared buses and redundant Memory Channel or LAN interconnect hardware promotes no-single-point-of-failure (NSPOF) characteristics for mission-critical applications.

A LAN interconnect can be a direct connection between two cluster members or can employ hubs or switches. In general, any Ethernet adapter, switch, or hub that works in a standard LAN at 100 Mbps or 1000 Mbps should work within a LAN interconnect. (Fiber Distributed Data Interface (FDDI), ATM LAN Emulation (LANE), and 10 Mbps Ethernet are not supported.)

Adapters and switch ports must be configured compatibly with respect to speed (100 Mbps or 1000 Mbps) and operational mode (full-duplex). A maximum of three hops is allowed between cluster members, where a hop means passing from a system, switch, hub, or router, to another system, switch, hub, or router. That is, any combination of up to two hubs, switches, or routers is supported between two cluster members. You must not introduce unacceptable latencies by using, for example, a satellite uplink or a wide area network (WAN) in the path between two components of a LAN interconnect.

A fully redundant LAN interconnect configuration employs two or more Ethernet adapters in a NetRAIN set on each member, with redundant wiring to two or more switches interlinked by two crossover cables. These Ethernet switches must be capable of one of the follow mechanisms for managing traffic across parallel inter-switch links: link aggregation (also known as port trunking), resilient links, or per-port-enabled spanning tree algorithm.