HP OpenVMS systems documentation

The configuration in Figure 8-7 incorporates the following strategies, which are critical to its success:

• This configuration has no single point of failure.
• The FDDI interconnect has sufficient bandwidth to serve satellite nodes from the base server configuration.
• All shared storage is MSCP served from the base configuration, which is appropriately configured to serve a large number of nodes.

Multiple-site OpenVMS Cluster configurations contain nodes that are located at geographically separated sites. Depending on the technology used, the distances between sites can be as great as 150 miles. FDDI, asynchronous transfer mode (ATM), and DS3 are used to connect these separated sites to form one large cluster. Available from most common telephone service carriers, DS3 and ATM services provide long-distance, point-to-point communications for multiple-site clusters.

Figure 8-8 shows a typical configuration for a multiple-site OpenVMS Cluster system. Figure 8-8 is followed by an analysis of the configuration that includes:

• Analysis of components
• Advantages

Figure 8-8 Multiple-Site OpenVMS Cluster Configuration Connected by WAN Link

Although Figure 8-8 does not show all possible configuration combinations, a multiple-site OpenVMS Cluster can include:

• Two data centers with an intersite link (FDDI, ATM, or DS3) connected to a DECconcentrator or GIGAswitch crossbar switch.
• Intersite link performance that is compatible with the applications that are shared by the two sites.
• Up to 96 Alpha and VAX (combined total) nodes. In general, the rules that apply to OpenVMS LAN and extended LAN (ELAN) clusters also apply to multiple-site clusters.
  Reference: For LAN configuration guidelines, see Section 4.12.6. For ELAN configuration guidelines, see Section 10.7.8.

The benefits of a multiple-site OpenVMS Cluster system include the following:

• A few systems can be remotely located at a secondary site and can benefit from centralized system management and other resources at the primary site. For example, a main office data center could be linked to a warehouse or a small manufacturing site that could have a few local nodes with directly attached, site-specific devices. Alternatively, some engineering workstations could be installed in an office park across the city from the primary business site.
• Multiple sites can readily share devices such as high-capacity computers, tape libraries, disk archives, or phototypesetters.
• Backups can be made to archival media at any site in the cluster. A common example would be to use disk or tape at a single site to back up the data for all sites in the multiple-site OpenVMS Cluster. Backups of data from remote sites can be made transparently (that is, without any intervention required at the remote site).
• In general, a multiple-site OpenVMS Cluster provides all of the availability advantages of a LAN OpenVMS Cluster. Additionally, by connecting multiple, geographically separate sites, multiple-site OpenVMS Cluster configurations can increase the availability of a system or elements of a system in a variety of ways:
  • Logical volume/data availability---Volume shadowing or redundant arrays of independent disks (RAID) can be used to create logical volumes with members at both sites. If one of the sites becomes unavailable, data can remain available at the other site.
  • Site failover---By adjusting the VOTES system parameter, you can select a preferred site to continue automatically if the other site fails or if communications with the other site are lost.

Reference: For additional information about multiple-site clusters, see HP OpenVMS Cluster Systems.

8.12 Disaster-Tolerant OpenVMS Cluster Configurations

Disaster-tolerant OpenVMS Cluster configurations make use of Volume Shadowing for OpenVMS, high-speed networks, and specialized management software.

Disaster-tolerant OpenVMS Cluster configurations enable systems at two different geographic sites to be combined into a single, manageable OpenVMS Cluster system. Like the multiple-site cluster discussed in the previous section, these physically separate data centers are connected by FDDI or by a combination of FDDI and ATM, T3, or E3.

The OpenVMS disaster-tolerant product was formerly named the Business Recovery Server (BRS). BRS has been subsumed by a services offering named Disaster Tolerant Cluster Services, which is a system management and software service package. For more information about Disaster Tolerant Cluster Services, contact your HP Services representative.

There are many ways to configure a CI (cluster interconnect) OpenVMS Cluster system. This chapter describes how to configure CI OpenVMS Clusters to maximize both availability and performance. This is done by presenting a series of configuration examples of increasing complexity, followed by a comparative analysis of each example. These configurations illustrate basic techniques that can be scaled upward to meet the availability, I/O performance, and storage connectivity needs of very large clusters.

9.1 CI Components

The CI is a radial bus through which OpenVMS Cluster systems communicate with each other and with storage. The CI consists of the following components:

• CI host adapter
• HSJ or HSC storage controller
  An HSJ or HSC storage controller is optional but generally present.
• CI cables
  For each of the CI's two independent paths (called path A and path B), there is a transmit and receive cable pair.
• Star coupler
  This is a passive device that serves as a common connection point for signals between OpenVMS nodes and HSC or HSJ controllers that are connected to the CI. A star coupler consists of two completely independent and electrically isolated "path hubs." Each CI path hub is extremely reliable because it contains only transformers carrying low-power signals.

Availability and performance can both be increased by adding components. Components added for availability need to be configured so that a redundant component is available to assume the work being performed by a failed component. Components added for performance need to be configured so that the additional components can work in parallel with other components.

Frequently, you need to maximize both availability and performance. The techniques presented here are intended to help achieve these dual goals.

9.2 Configuration Assumptions

The configurations shown here are based on the following assumptions:

1. MSCP serving is enabled.
2. Volume Shadowing for OpenVMS is installed.
3. When performance is being discussed:
   1. CI host adapters are CIPCA or CIXCD.
      Older CI adapter models are significantly slower.
   2. CI storage controllers are HSJ50s.
      Compared with HSJ50s, HSJ40s are somewhat slower, and HSC models are significantly slower.

Configuration 1, shown in Figure 9-1, provides no single point of failure. Its I/O performance is limited by the bandwidth of the star coupler.

Figure 9-1 Redundant HSJs and Host CI Adapters Connected to Same CI (Configuration 1)

9.3.1 Components

This configuration offers the following advantages:

• All nodes have direct access to storage.
• Highly expandable.
• CI is inherently dual pathed.
• No single component failure can disable the cluster.
• If a CI adapter fails, or both its paths are disabled, OpenVMS will automatically fail over all I/O and cluster traffic to the other CI adapter.
• Disks are dual ported between HSJ controllers; automatic disk failover to the other controller if an HSJ fails or if an HSJ loses both paths to a star coupler.
• Redundant storage controllers can be used to provide additional performance by dividing disks between the two storage controllers.
  Disks can be assigned to HSJ storage controllers by the OpenVMS Prefer utility supplied in SYS$EXAMPLES, or by issuing a $QIO call with IO$_SETPRFPATH and IO$M_FORCEPATH modifiers, or by using the HSJ SET_PREFERRED command (less desirable; use only for this configuration).
• Critical disks are shadowed with shadow set members on different SCSI buses.
• Read I/Os are automatically load balanced across shadow set members for performance.
• Lowest cost.

This configuration has the following disadvantages:

• Second CI adapter in each host is unlikely to enhance performance.
• Both HSJs have to share the bandwidth of a single CI.
• Failure of a CI path hub or path cable halves the bandwidth available to all CI components that use the failed component.
• Physical damage to a star coupler or associated cables is likely to disable the entire CI, rendering the cluster unusable.
• Physical damage to the StorageWorks cabinet could render the cluster unusable.

This configuration incorporates the following strategies:

• All components are duplicated.
• Redundant storage controllers are included.
• This configuration has no single point of failure.
• Dual porting and volume shadowing provide multiple copies of essential disks across separate HSJ controllers.
• All nodes have shared, direct access to all storage.
• A quorum disk allows other node to continue if one node fails. (Alternatively, at least three nodes could be used.)

The configuration illustrated in Figure 9-2 with redundant HSJs, host CI adapters, and CIs provides no electrical single point of failure. Its two star couplers provide increased I/O performance and availability over configuration 1.

Figure 9-2 Redundant HSJs and Host CI Adapters Connected to Redundant CIs (Configuration 2)

Configuration 2 has the following components:

Part	Description
Host 1, Host 2	Dual CI capable OpenVMS Alpha or VAX hosts. Rationale: Either host can fail and the system can continue to run. The full performance of both hosts is available for application use under normal conditions.
CI 1-1,CI 1-2, CI 2-1, CI 2-2	Dual CI adapters on each host. Adapter CI 1- n is Host 1's CI adapter connected to CI n, and so on. Rationale: Either of a host's CI adapters can fail, and the host will retain CI connectivity to the other host and to the HSJ storage controllers. Each CI adapter on a host is connected to a different star coupler. In the absence of failures, the full data bandwidth and I/O-per-second capacity of both CI adapters is available to the host.
Star Coupler 1, Star Coupler 2	Two star couplers, each consisting of two independent path hub sections. Each star coupler is redundantly connected to the CI host adapters and HSJ storage controllers by a transmit/receive cable pair per path. Rationale: Any of the path hubs or an attached cable could fail and the other CI path would continue to provide full connectivity for that CI. Loss of a path affects only the bandwidth available to the storage controller and host adapters connected to the failed path. When all paths are available, the combined bandwidth of both CIs is usable.
HSJ 1, HSJ 2	Dual HSJ storage controllers in a single StorageWorks cabinet. Rationale: Either storage controller can fail and the other controller can control any disks the failed controller was handling by means of the SCSI buses shared between the two HSJs. When both controllers are available, each can be assigned to serve a subset of the disks. Thus, both controllers can contribute their I/O-per-second and bandwidth capacity to the cluster.
SCSI 1, SCSI 2	Shared SCSI buses connected between HSJ pairs. Rationale: Either of the shared SCSI buses could fail and access would still be provided from the HSJ storage controllers to each disk by means of the remaining shared SCSI bus. This effectively dual ports the disks on that bus.
Disk 1, Disk 2, . . . Disk n-1, Disk n	Critical disks are dual ported between HSJ pairs by shared SCSI buses. Rationale: Either HSJ can fail and the other HSJ will assume control of the disks the failed HSJ was controlling.
Shadow Set 1 through Shadow Set n	Essential disks are shadowed by another disk that is connected on a different shared SCSI. Rationale: A disk or the SCSI bus to which it is connected, or both, can fail and the other shadow set member will still be available. When both disks are available, both can provide their READ I/O capacity and their READ data bandwidth capacity to the cluster.

9.4.2 Advantages

Configuration 2 offers all the advantages of Configuration 1 plus the following advantages:

• CI is likely to remain fully usable in the event of localized damage to a star coupler cabinet or to the cables connected to it.
• Failure of an HSJ, or an HSJ losing both paths to a star coupler host will result in the other HSJ assuming control of all dual-pathed disks.
• Host with connectivity will automatically MSCP serve unreachable disks to a host that has lost connectivity to an HSJ.
• Dual-pathed disks can be switched to an HSJ with full host connectivity if a host loses connectivity to an HSJ.
  The disks on the HSJ that cannot be reached by a host can be reassigned to the HSJ with full connectivity by operator command, or by a DCL or other program that monitors for connectivity loss. You can assign disks to another storage controller with the Prefer utility supplied in SYS$EXAMPLES, or by issuing a $QIO call with IO$_SETPRFPATH and IO$M_FORCEPATH modifiers, or by powering off the HSJ with reduced connectivity (less desirable).
  The HSJ SET_PREFERRED command is not recommended in this configuration because this command can not be overridden by a host PREFER or IO$_SETPRFPATH modifier. A new SET_PREFERRED command assigning a device to another HSJ will not take effect until the HSJ, to which the device was assigned by a previous SET_PREFERRED command, is power cycled.)

Configuration 2 has the following disadvantages:

• Host CI adapter failure will not cause disks to automatically fail over to an HSJ that still has full host connectivity.
  If a host CI adapter fails and if MSCP serving is enabled, then the other OpenVMS system will begin serving the unreachable HSJ's disks to the host with the failed adapter.
• Physical damage to a star coupler or associated cables is likely to disable the entire CI, rendering the cluster unusable.
• Physical damage to the StorageWorks cabinet could render the cluster unusable.
• Higher cost than configuration 1.

Configuration 2 provides configuration 1 strategies, plus:

• Second CI provides additional redundancy.
• Second CI cables and star coupler are physically separate.

The availability of a CI configuration can be further improved by physically separating the path A and path B CI cables and their associated path hubs. This significantly reduces the probability of a mechanical accident or other localized damage destroying both paths of a CI. This configuration is shown in Figure 9-3.

Figure 9-3 Redundant Components and Path-Separated Star Couplers (Configuration 3)

Configuration 3 is electrically identical to configuration 2. However, the path A and path B cables are physically separate for both CIs. The path A cables for both CI 1 and CI 2 are routed together to star coupler cabinet A, but are connected to different CI path hubs in that cabinet.

Similarly, the path B cables for both CIs go to different CI path hubs in star coupler cabinet B. The path-specific star coupler cabinets and associated path cables should be separated as much as possible. For example, the star coupler cabinets could be installed on opposite sides of a computer room, and the CI cables could be routed so that path A and path B cables follow different paths.

The CI OpenVMS Cluster configuration shown in Figure 9-3 has the following components:

Part	Description
Host 1, Host 2	Dual CI capable OpenVMS Alpha or VAX hosts. Rationale: Either host can fail and the system can continue. The full performance of both hosts is available for application use under normal conditions.
CI 1-1,CI 1-2, CI 2-1, CI 2-2	Dual CI adapters on each host. Adapter CI 1- n is Host 1's CI adapter connected to CI n, and so on. Rationale: Either host's CI adapters can fail and the host will retain CI connectivity to the other host and to the HSJ storage controllers. Each CI adapter on a host is connected to a different star coupler. In the absence of failures, the full data bandwidth and I/O-per-second capacity of both CI adapters is available to the host.
Star Coupler A (Path A Hubs), Star Coupler B (Path B Hubs)	Two CI star couplers, each comprising two independent path hubs. Star Coupler A's path hubs are connected to path A cables for both CIs, and Star Coupler B's path hubs are connected to path B cables for both CIs. Rationale: Mechanical or other localized damage to a star coupler or an attached cable would probably not affect the other CI paths. The other paths and star coupler would continue to provide full connectivity for both CIs. Loss of a path affects only the bandwidth available to the storage controllers and host adapters connected to the failed path. When all paths are available, the combined bandwidth of both CIs is usable.
Path A CI Cables, Path B CI Cables	Each path's hub is connected to the CI host adapters and HSJ storage controllers by a transmit/receive cable pair per path. The path A cables of both CIs are routed together, but their routing differs from the routing of the path B cables.
HSJ 1, HSJ 2	Dual HSJ storage controllers in a single StorageWorks cabinet. Rationale: Either storage controller can fail and the other controller can control any disks the failed controller was handling by means of the SCSI buses shared between the two HSJs. When both controllers are available, each can be assigned to serve a subset of the disks. Thus, both controllers can contribute their I/O-per-second and bandwidth capacity to the cluster.
SCSI 1, SCSI 2	Shared SCSI buses connected between HSJ pairs. Rationale: Provide access to each disk on a shared SCSI bus from either HSJ storage controller. This effectively dual ports the disks on that bus.
Disk 1, Disk 2, . . . Disk n-1, Disk n	Critical disks are dual ported between HSJ pairs by shared SCSI buses. Rationale: Either HSJ can fail and the other HSJ will assume control of the disks that the failed HSJ was controlling.
Shadow Set 1 through Shadow Set n	Essential disks are shadowed by another disk that is connected on a different shared SCSI. Rationale: A disk, or the SCSI bus to which it is connected, or both, can fail and the other shadow set member will still be available. When both disks are available, their combined READ I/O-per-second capacity and READ data bandwidth capacity is available to the cluster.