HSG80 Array Controller ACS Version 8.4 Configuration and CLI Reference Guide First Edition (April, 1999) Part Number EK-HSG84-RG.
While Compaq Computer Corporation believes the information included in this publication is correct as of the date of publication, it is subject to change without notice.
iii Contents About This Guide Getting Help. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv Compaq Website. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv Telephone Numbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv Precautions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
iv Chapter 2 Controller and Host Concepts Terminology. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2–2 Controller A and Controller B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2–2 This Controller and Other Controller . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2–2 Unresponsive and Surviving Controllers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2–4 Selecting a Failover Mode . .
v Controller and Port Worldwide Names (Node IDs). . . . . . . . . . . . . . . . . . . . . . . . 3–15 Restoring Worldwide Names (Node IDs). . . . . . . . . . . . . . . . . . . . . . . . . . . . 3–17 Unit World Wide Names (LUN IDs) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3–17 Chapter 4 Planning Storagesets Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4–2 Planning and Configuring Storagesets. . . . . . . . . .
vi Reconstruction Policy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4–37 Membership . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4–38 Mirrorset Switches. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4–39 Replacement Policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4–39 Copy Speed. . . . . . . . . . . . . . . . . . .
vii Setting a Port Offline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5–31 Backing Up Power with a UPS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5–31 Changing the Host Default Value . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5–32 Establishing a Local Connection to the Controller . . . . . . . . . . . . . . . . . . . . . . . . 5–33 Shutting Down Your Subsystem . . . . . . . . . . . . . . . . . . . . . . .
viii Moving Storagesets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6–25 Chapter 7 CLI Commands CLI Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7–2 Using the CLI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7–2 Command Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
ix DELETE container-name . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7–53 DELETE FAILEDSET . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7–55 DELETE SPARESET . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7–56 DELETE unit-number . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7–57 DESTROY_PARTITION. . . . . . . . . . . . . . . . . . . .
x SET unit-number . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7–126 SHOW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7–133 SHUTDOWN controller . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7–140 UNMIRROR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7–142 System Profiles Device Profile .
xi Figures The HSG80 Subsystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1–3 HSG80 Array Controller Fibre Channel Optical Cabling . . . . . . . . . . . . . . . . . . . 1–8 Location of Controllers and Cache Modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1–10 HSG80 Controller Operator Control Panel (OCP) . . . . . . . . . . . . . . . . . . . . . . . . 1–11 Cache Module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
xii CLONE Steps for Duplicating Unit Members . . . . . . . . . . . . . . . . . . . . . . . . . . . 4–22 Storageset Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4–31 Storageset Map Template . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4–33 Partitioning a Single-Disk Unit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4–34 Chunk Size Larger than the Request Size . . . . . . . . .
xiii Tables Key to Figure 1–1 The HSG80 Subsystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1–4 Controller Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1–5 Key to Figure 1–2 HSG80 Array Controller Fibre Channel Optical Cabling . . . . 1–8 Key to Figure 1–4 HSG80 Controller Operator Control Panel (OCP) . . . . . . . . . 1–11 Cache Module Memory Configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
xiv Key to Figure 5–7 Example Cabling in Multiple-Bus Failover: Configuration 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5–26 Key to Figure 5–9 Terminal to Local-Connection Port Connection . . . . . . . . . . . 5–34 Unit Switches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6–2 Recall and Edit Command Keys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
xv About This Guide This book describes the features of the HSG80 array controller and configuration procedures for the controller and storagesets running Array Controller Software (ACS) Version 8.4F. Getting Help If you have a problem and have exhausted the information in this guide, you can get further information and other help in the following locations. Compaq Website The Compaq Website has information on this product as well as the latest drivers and Flash ROM images.
xvi About This Guide Precautions Follow these precautions when carrying out the procedures in this book. Electrostatic Discharge Precautions Static electricity collects on all nonconducting material, such as paper, cloth, and plastic. An electrostatic discharge (ESD) can easily damage a controller or other subsystem component even though you may not see or feel the discharge.
xvii Maintenance Port Precautions The maintenance port generates, uses, and radiates radio-frequency energy through cables that are connected to it. This energy may interfere with radio and television reception. Do not leave a cable connected to this port when you’re not communicating with the controller. FCC Precautions This equipment has been tested and found to comply with the limits for a Class A digital device, pursuant to Part 15 of the FCC Rules.
xviii About This Guide Attention! Ceci est un produit de Classe A. Dans un environnement domestique, ce produit risque de créer des interférences radioélectriques, il appartiendra alors à l'utilisateur de prendre les mesures spécifiques appropriées. JAPAN USA This equipment generates, uses, and may emit radio frequency energy.
xix Typographical Conventions Convention ALLCAPS Meaning Command syntax that must be entered exactly as shown and for commands discussed within text, for example: SET FAILOVER COPY=OTHER_CONTROLLER “Use the SHOW SPARESET command to show the contents of the spareset.” Monospaced Screen display.
xx About This Guide Special Notices This book doesn’t contain detailed descriptions of standard safety procedures. However, it does contain warnings for procedures that could cause personal injury and cautions for procedures that could damage the controller or its related components. Look for these symbols when you’re carrying out the procedures in this book: WARNING: A warning indicates the presence of a hazard that can cause personal injury if you do not observe the precautions in the text.
xxi Related Publications The following table lists some of the documents that are related to the use of the controller, cache module, and external cache battery. Document Title Part Number Fibre Channel Storage Switch Service Guide AA-RHBZA-TE / 135268-001 Fibre Channel Storage Switch User’s Guide AA-RHBYA-TE / 135267-001 Fibre Channel Switch Quick Setup Guide AA-RHC0A-TE / 135269-001 Compaq StorageWorks Data Replication Manager HSG80 ACS Version 8.
xxii About This Guide Document Title Compaq StorageWorks RA8000/ESA12000 HSG80 V8.
xxiii Revision History This is a new document. Compaq HSG80 Array Controller ACS Version 8.
1–1 Chapter 1 General Description This chapter illustrates and describes in general terms your subsystem and its major components: the HSG80 array controller, its cache module, and its external cache battery. See the Fibre Channel Switch Documentation that came with the switch kit for specifics about how the switch operates. See the KGPSA PCI-to-Fibre Channel Host Adapter User Guide for information about the adapter that connects the subsystem to your host. Compaq HSG80 Array Controller ACS Version 8.
1–2 General Description The HSG80 Array Controller Subsystem Take a few moments to familiarize yourself with the major components of the HSG80 array controller subsystem. Figure 1–1 shows the components of a typical installation that includes: ■ One BA370 rack-mountable enclosure. ■ Two controllers, each supported by a cache module. ■ Two cache modules, which support nonvolatile memory and dynamic cache policies to protect the availability of its unwritten (write-back) data.
1–3 1 16 2 15 13 14 3 12 2x 4 11 10 9 2x 8 2x 7 5 7 6 CXO6841A Figure 1–1. The HSG80 Subsystem Compaq HSG80 Array Controller ACS Version 8.
1–4 General Description Table 1–1.
1–5 Summary of HSG80 Features Table 1–2 summarizes the features of the controller. Table 1–2.
1–6 General Description Table 1–2.
1–7 Table 1–2. Controller Features (Continued) Feature Supported Maximum number of RAID-5, RAID-1, and RAID-0 storagesets ■ 45 Maximum number of partitions per storageset or individual disk ■ 8 Maximum number of units presented to each host ■ WINNT—16 (8 on each of 2 host adapters). This is a driver limitation. ■ OpenVMS—127 and one Command Console LUN.
1–8 General Description Controller Components Figure 1–2 details the HSG80 Array Controller and its Fibre Channel optical components. 2 3 2x 4 5 1 3X 6 1 2 3 4 5 6 7 10 11 12 8 2x 9 CXO6597A Figure 1–2. HSG80 Array Controller Fibre Channel Optical Cabling Table 1–3.
1–9 Table 1–3.
1–10 General Description Controller Components The HSG80 Array Controller components that you will use most often, such as the maintenance port and the OCP, are conveniently located on the controller’s front panel. The host port and program-card slot are also located on the front panel, making it easy to update the controller’s software or to connect the controller to a different host. Each controller is supported by its own cache module.
1–11 Operator Control Panel The operator control panel (OCP) contains a reset button and six port LED buttons, as shown in Figure 1–4. The reset button flashes about once per second to indicate that the controller is operating normally. The port button LEDs correspond to the controller’s device ports and remain off during normal operation. If an error occurs, the reset button and device fault LEDs will illuminate in a solid or flashing pattern to help you diagnose the problem.
1–12 General Description After you configure your controller, you should periodically check its control panel. If an error occurs, one or more of the device fault LEDs on the control panel will flash in a pattern that will help you to diagnose the problem. See the HSG80 Array Controller ACS Version 8.3 and 8.4 Maintenance and Service Guide for details about troubleshooting your controller.
1–13 Virtual Terminal Display Use the virtual terminal display (VTDPY) utility to troubleshoot communication between the controller and its host, communication between the controller and the devices in the subsystem, and the state and I/O activity of the logical units, devices, and device ports in the subsystem. Disk Inline Exerciser Use the disk inline exerciser (DILX) to investigate the data-transfer capabilities of disk drives.
1–14 General Description microcode updates. Contact the Customer Service Center (CSC) for directions in obtaining the appropriate EMU microcode and installation guide. Clone Utility Use the Clone utility to duplicate the data on any unpartitioned single-disk unit, stripeset, or mirrorset. Back up the cloned data while the actual storageset remains online.
1–15 Cache Module Each controller requires a companion cache module as shown in Figure 1–5. Figure 1–3 on page 1–10, shows the location of a controller’s companion cache module. The cache module, which can contain up to 512 MB of memory, increases the subsystem’s I/O performance by providing read, read-ahead, write-through, and write-back caching. The size of the memory contained in the cache module depends on the configuration of the DIMMs, with the supported combinations shown in Table 1–5.
1–16 General Description 5 4 1 ~ 2 3 2x CXO6161A Figure 1–5. Cache Module Table 1–6.
1–17 External Cache Battery To preserve the write-back cache data in the event of a primary power failure, a cache module must be connected to an external cache battery (ECB) or a UPS. Compaq supplies two versions of ECBs: a single-battery ECB for single controller configurations, and a dual-battery ECB for dual-redundant controller configurations, which is shown in Figure 1–6. 1 2 SH US STAT F OF UT E CH CA ER W PO E CH CA ER W PO US STAT F OF UT SH 4 3 ~ CXO5713A Figure 1–6.
1–18 General Description When the batteries are fully charged, an ECB can preserve 512 MB of cache memory for 24 hours. However, the battery capacity depends upon the size of memory contained in the cache module, as defined in the Table 1–8. Table 1–8.
1–19 Battery Hysteresis When charging a battery, write-back caching will be allowed as long as a previous down time has not drained more than 50 percent of a battery’s capacity. When a battery is operating below 50 percent capacity, the battery is considered to be low, and write-back caching is disabled. CAUTION: Compaq recommends that you replace the ECB every two years to prevent battery failure. NOTE: If a UPS is used for backup power, the controller does not check the battery.
2–1 Chapter 2 Controller and Host Concepts This chapter provides concepts for understanding how the controller works in general, the difference between failover modes and how failover affects configurations, the basics of caching techniques, and an understanding of cache policies. Compaq HSG80 Array Controller ACS Version 8.
2–2 Controller and Host Concepts Terminology When configuring the subsystem you will encounter the following terms and concepts that you must understand: ■ Controller A and controller B ■ This controller and other controller ■ Unresponsive and surviving controller Controller A and Controller B Controllers and cache modules are designated either A or B depending on their location within the storage enclosure as shown in.
2–3 ■ Other controller—the controller that you are not currently connected to or communicating through a terminal program, regardless of whether it is controller A or controller B. The maintenance terminal is not connected to the maintenance port of the “other” controller. Figure 2–2 shows an example of when controller B can be called “this” controller when the maintenance port cable is connected to it. To make controller A into “this” controller, simply move the maintenance port cable to it.
2–4 Controller and Host Concepts Unresponsive and Surviving Controllers When a failover occurs in a dual-redundant pair of controllers, the controller that ceases I/O processing is sometimes referred to as the unresponsive controller. The controller that is failed over to is frequently called the surviving controller.
2–5 Selecting a Failover Mode When you select a failover mode, your selection determines the way in which your controller will be configured in the system. This configuration will take into account how logical units are associated with the controller host ports and how they are accessed by the host. The logical units are said to fail over to the surviving controller, essentially granting it access to the logical units.
2–6 Controller and Host Concepts Shared Port IDs Figure 2–3 shows one example of a Fibre Channel Switched Fabric (FC-SW) configuration in transparent failover mode. In this configuration, the active port attaches to a FC-SW node. The standby port on each controller will assume the worldwide name if the active port fails, so that the standby port will be able to give host access to the units.
2–7 As Figure 2–3 shows, only port 1 on controller A normally is active, which means it presents LUNs 0–99 to the host. Similarly, port 2 on controller B normally is active, which means it also presents LUNs, 100–199 to the host. If one controller fails, the surviving controller automatically changes the standby port to an active port presents the LUNs to the host.
2–8 Controller and Host Concepts FC-SW Host port 1 active Controller A Host port 2 active FC-SW FC-SW Host port 1 active Controller B Host port 2 active FC-SW CXO6862A Figure 2–4. Example of Separate Host Port IDs in Multiple-Bus Failover As Figure 2–4 shows, all ports on both controllers are active and have the ability to present all LUNs, 0–199, to the host. If one pot or one path loses communication, the drivers in the host can access the units through another port or another path.
2–9 Caching Techniques The cache module supports the following caching techniques to increase the subsystem’s read and write performance: ■ Read caching ■ Read-ahead caching ■ Write-through caching ■ Write-back caching Read Caching When the controller receives a read request from the host, it reads the data from the disk drives, delivers it to the host, and stores the data in its cache module.
2–10 Controller and Host Concepts During read-ahead caching, the controller anticipates subsequent read requests and begins to prefetch the next blocks of data from the disks as it sends the requested read data to the host. This is a parallel action. The controller notifies the host of the read completion, and subsequent sequential read requests are satisfied from the cache memory. By default, read-ahead caching is enabled for all disk units.
2–11 Fault-Tolerance for Write-Back Caching The cache module supports nonvolatile memory and dynamic cache policies to protect the availability of its unwritten (write-back) data: Nonvolatile Memory The controller can provide write-back caching for any storage unit as long as the controller’s cache memory is nonvolatile. In other words, to enable write-back caching, you must provide a backup power source to the cache module to preserve the unwritten cache data in the event of a power failure.
2–12 Controller and Host Concepts Table 2–2 shows the cache policies resulting from a full or partial failure of cache module A in a dual-redundant controller configuration. The consequences shown in this table are the same for cache module B. Table 2–2.
2–13 Table 2–2. Cache Policies and Cache Module Status (Continued) Cache Module Status Cache A DIMM or cache memory controller chip failure Cache B Good Cache Policy Unmirrored Cache Mirrored Cache Data integrity: Write-back data that was not written to media when failure occurred was not recovered. Data integrity: Controller A recovers all of its write-back data from the mirrored copy on cache B.
2–14 Controller and Host Concepts Table 2–2. Cache Policies and Cache Module Status (Continued) Cache Module Status Cache A Cache Board Failure Cache B Good Cache Policy Unmirrored Cache Same as for DIMM failure. Mirrored Cache Data integrity: Controller A recovers all of its write-back data from the mirrored copy on cache B. Cache policy: Both controllers support write-through caching only. Controller B cannot execute mirrored writes because cache module A cannot mirror controller B’s unwritten data.
2–15 Table 2–3. Resulting Cache Policies and ECB Status ECB Status Cache A At least 50% charged Cache B At least 50% charged Cache Policy Unmirrored Cache Mirrored Cache Data loss: No Data loss: No Cache policy: Both controllers continue to support write-back caching. Cache policy: Both controllers continue to support write-back caching.
2–16 Controller and Host Concepts Table 2–3. Resulting Cache Policies and ECB Status (Continued) ECB Status Cache A Failed Cache B At least 50% charged Cache Policy Unmirrored Cache Mirrored Cache Data loss: No Data loss: No Cache policy: Controller A supports write-through caching only; controller B supports write-back caching. Cache policy: Both controllers continue to support write-back caching.
2–17 Table 2–3. Resulting Cache Policies and ECB Status (Continued) ECB Status Cache A Failed Cache B Less than 50% charged Cache Policy Unmirrored Cache Mirrored Cache Data loss: No Data loss: No Cache policy: Both controllers support write-through caching only. Cache policy: Both controllers support write-through caching only. Failover: In transparent failover, all units failover to controller B and operate normally.
2–18 Controller and Host Concepts Enabling Mirrored Write-Back Cache Before configuring dual-redundant controllers and enabling mirroring, ensure the following conditions are met: ■ Both controllers are configured with the same size cache, 64 MB, 128 MB, 256 MB, or 512 MB. ■ Diagnostics indicates that both caches are good. ■ Both caches have a battery present, if you have not enabled the CACHE_UPS switch. A battery does not have to be present for either cache if you enable the CACHE_UPS switch.
3–1 Chapter 3 Controller and Host Addressing This chapter provides a basis for understanding how the controller talks to physical devices (on the controller bus) and how the host talks to logical units (on the host bus), which are created from the physical devices. Also described are storagesets, mapping devices and units, setting unit offsets, assigning preferred host paths, and understanding worldwide names and port IDs. Compaq HSG80 Array Controller ACS Version 8.
3–2 Controller and Host Addressing Bus Dynamics Your controller is the intelligent bridge between your host and the devices and the logical units comprising your subsystem. The controller is an integral part of any storage subsystem because it provides a host with high-performance and highavailability access to the storage devices, sometimes called disk drive Storage Building Blocks (SBBs).
3–3 Table 3–1. Key to Figure 3–1 Host Bus and Controller Device SCSI Bus Item Description ➀ Host ➁ Controller A ➂ Controller B ➃ Logical units made from storagesets ➄ Host bus ➅ Controller device bus Host Bus When the host communicates to the controller through the host bus, the controller maps that communication to the logical units on the device bus. Even though the physical devices exist, the host accesses only the logical units that you have created from making containers and storagesets.
3–4 Controller and Host Addressing Containers and Storagesets Containers and storagesets are comprised of physical devices. The logical units are created after the devices and storagesets have been added to the storage subsystem. Figure 3–2 shows the configuration for containers, storagesets, and single devices. Frequently, containers are called by the more generic term, storagesets.
3–5 The following are the supported storagesets for the HSG80 array controller: ■ RAID 0 (Stripesets)—physical devices combined in serial to increase transfer or request rates. ■ RAID 1 (Mirrorsets)—physical devices combined in parallel to provide a highly reliable storage unit. ■ RAID 0 + 1 (Striped Mirrorsets)—combined mirrorsets in serial and parallel providing the highest throughput and availability.
3–6 Controller and Host Addressing Logically Connecting the Storage Array to the Host The combination of storage devices is called a storage array. The controller uses a two-step mapping process to logically connect the host to the storage array: ■ Step1—The controller maps the physical devices on its six device buses to storage containers that you have created. ■ Step 2—The controller maps its internal containers to user-created logical units that are directly accessible by the host.
3–7 Place one space between the port number, target number, and the two-digit LUN number when entering the PTL address. An example of a PTL address is shown in Figure 3–3. 1 02 00 LUN 00 (leading zeros are not required) Target 02 (leading zeros are not required) Port 1 Figure 3–3 PTL Naming Convention Figure 3–4 shows the addresses for each device in a typical, extended configuration. Use this figure along with the “Configuration Rules,” page 5–3 to help you work with the devices in your configuration.
3–8 Controller and Host Addressing 51500 41500 61500 61400 61300 61200 51400 41400 6 51300 41300 41200 EMU 5 51200 21500 31500 31400 31300 11500 4 31200 21400 11200 PVA 2 11400 11300 12 3 21300 61100 13 2 21200 51100 61000 50800 51000 40900 40800 EMU 14 60900 41100 41000 30900 30800 PVA 0 EMU Controller A Controller B Cache A Cache B 15 50900 31100 31000 20900 8 1 60800 21100 21000 9 20800 50100 50000 11100 40300 40200 40100 0 40000 11000 30300 30200
3–9 When your controller receives an I/O request, it identifies the host logical unit number, listing all the storageset names in the logical unit. Using the unit number, the controller locates the appropriate device for the I/O request. For example in Figure 3–5, a RAIDset called RAID1 might contain DISK10000, DISK20000, and DISK30000. The controller generates the read or write request to the appropriate device using the PTL addressing convention.
3–10 Controller and Host Addressing Mapping the Containers and Storagesets to the Host with Logical Units In a subsystem, the host communicates to the physical devices through hostaddressable logical units, also called units. These units are created and given a logical unit number (LUN) after adding physical devices, adding containers, and initializing the containers. The host uses these numbers to indicate the source or target for every I/O request it sends to a controller.
3–11 Controller A Port 1 (Active) Port 2 (Standby) Units 0-99 Units 100-199 Controller B Port 1 (Standby) Port 2 (Active) Units 0-99 Units 100-199 CXO6187B Figure 3–6 Controller Port ID and Unit Numbers in Transparent Failover Mode Assigning Unit Numbers in Multiple-Bus Failover Mode In multiple-bus failover mode, the range of assignable units that are accessible from any port on the subsystem is 0–199. Hosts obtain units by reserving the unit for sole access.
3–12 Controller and Host Addressing Controller A Port 1 (Active) Port 2 (Active) Units 0-199 Units 0-199 Controller B Port 1 (Active) Port 2 (Active) Units 0-199 Units 0-199 CXO6454B Figure 3–7 Controller Port ID Numbers and Unit Numbers in Multiple-Bus Failover Mode
3–13 Assigning Unit Offsets Unit offsets are used to define the first unit number that a host connection is able to access on the controller. Many operating systems have a limited addressing range, such as 0–7. Using unit offsets allows hosts to see controller LUNs at 0–7. When assigning unit offsets, keep in mind that the unit offset is set on a host connection basis.
3–14 Controller and Host Addressing Assigning Access Paths Different operating systems may not tolerate other operating systems’ access to their units. To limit or restrict host access to certain units, the HSG80 array controller allows you to specify unit access privileges. You may specify unit access on a unit-byunit basis. You can enable host access to a specific unit by mapping it as accessible through the specified host paths.
3–15 Controller and Port Worldwide Names (Node IDs) A worldwide name—also called a NODE_ID—is a unique 64-bit number assigned to a subsystem by the Institute of Electrical and Electronics Engineers (IEEE) and set by Compaq manufacturing prior to shipping. As shown in Figure 3–9, it refers to the shelf the controllers sit on, and the worldwide name is stored in the controllers’ memory for the benefit of the shelf. The worldwide name assigned to a subsystem never changes.
3–16 Controller and Host Addressing In a subsystem with two controllers in transparent failover mode, the controller port IDs increment as follows: ■ Controller A and controller B, port 1—worldwide name + 1 ■ Controller A and controller B, port 2—worldwide name + 2 For example, using the worldwide name of 5000-1FE1-FF0C-EE00, the following port IDs are automatically assigned and shared between the ports as a REPORTED PORT_ID on each port: ■ Controller A and controller B, port 1—5000-1FE1-FF0C-EE01 ■
3–17 Restoring Worldwide Names (Node IDs) CAUTION: Each subsystem has its own unique worldwide name (node ID). This ID can be found on a sticker, which is located on top of the frame that houses the controllers, the EMU, the PVA, and cache modules. If you attempt to set the subsystem worldwide name to a name other than the one that came with the subsystem, the data on the subsystem will not be accessible. Never set two subsystems to the same worldwide name or data corruption will occur.
4–1 Chapter 4 Planning Storagesets This chapter takes you through the planning steps and procedures for planning and creating storagesets for your subsystem. Compaq HSG80 Array Controller ACS Version 8.
4–2 Planning Storagesets Introduction Storagesets are implementations of RAID technology, also known as a “Redundant Array of Independent Disks.” Every storageset shares one important feature: each one looks like a single storage unit to the host, regardless of the number of drives it uses. You can create storage units by combining disk drives into storagesets, such as stripesets, RAIDsets, and mirrorsets, or by presenting them to the host as single-disk units, as shown in Figure 4–1.
4–3 Unit Unit Mirrorset Unit Stripeset Partitioned storageset RAIDset Unit Striped mirrorset Disk drives Unit Partitioned disk drive Unit CXO5368B Figure 4–1. Units Created from Storagesets, Partitions, and Drives Table 4–1. Controller Limitations for RAIDsets RAIDset Type Limit Total number of RAID5 20 Total number of RAID5 + RAID1 30 Total number of RAID5 + RAID1 + RAID0 45 Compaq HSG80 Array Controller ACS Version 8.
4–4 Planning Storagesets Planning and Configuring Storagesets Use this procedure to plan and configure the storagesets for your subsystem. Use the references in each step to locate details about specific commands and concepts. 1. Create a storageset and device profile. See “Creating a Storageset and Device Profile,” page 4–5, for suggestions about creating a profile. 2. Determine your storage requirements. Use the questions in “Determining Storage Requirements,” page 4–7, to help you. 3.
4–5 Creating a Storageset and Device Profile Creating a profile for your storagesets and devices can help simplify the configuration process. This chapter helps you to choose the storagesets that best suit your needs and make informed decisions about the switches that you can enable for each storageset or storage device that you configure in your subsystem. Familiarize yourself with the kinds of information contained in a storageset profile, as shown in Figure 4–2.
4–6 Planning Storagesets TYPE OF STORAGESET _____ Mirrorset Storageset Name Disk Drives Unit Number Partitions Unit # % __✔_ RAIDset _____ Stripeset accept default values DISK10300, DISK20300, DIS30300 accept default Unit # % Unit # % RAIDset Switches Reconstruction Policy _✔ ___ Unit # % Normal (default) Fast Automatic (default) 64 blocks 128 blocks 256 blocks Other: No (default) Yes, missing: ___ ___ Normal (default) Fast Yes (default) No No (default) Yes Replacement Policy ___ ___ ___
4–7 Determining Storage Requirements Start the planning process by determining your storage requirements.
4–8 Planning Storagesets Choosing a Storageset Type Different applications may have different storage requirements, so you will probably want to configure more than one kind of storageset in your subsystem. All of the storagesets described in this book implement RAID (Redundant Array of Independent Disks) technology. Consequently, they all share one important feature: each storageset, whether it contains two disk drives or ten, looks like one large, virtual disk drive to the host.
4–9 For a comprehensive discussion of RAID, refer to The RAIDBOOK—A Source Book for Disk Array Technology. Using Stripesets to Increase I/O Performance Stripesets enhance I/O performance by spreading the data across multiple disk drives. Each I/O request is broken into small segments called “chunks.” These chunks are then “striped” across the disk drives in the storageset, thereby allowing several disk drives to participate in one I/O request to handle several I/O requests simultaneously.
4–10 Planning Storagesets The relationship between the chunk size and the average request size determines if striping maximizes the request rate or the data-transfer rate. You can set the chunk size or let the controller set it automatically. See “Chunk Size,” page 4–43, for information about setting the chunk size. A major benefit of striping is that it balances the I/O load across all of the disk drives in the storageset.
4–11 For example, if the mean time between failures (MTBF) for a single disk is one hour, then the MTBF for a stripeset that comprises N such disks is l/N hours. As another example, if a single disk’s MTBF is 150,000 hours (about 17 years), a stripeset comprising four of these disks would only have an MTBF of slightly more than four years. For this reason, you should avoid using a stripeset to store critical data.
4–12 Planning Storagesets 1 1 2 3 4 5 6 2 3 4 0 3 0 0 2 3 0 2 0 0 2 0 1 0 0 1 1 0 0 0 0 0 1 2 3 4 5 6 CXO6235B Figure 4–4. Distribute Members across Device Ports Table 4–3.
4–13 ■ Stripesets contain between 2 and 24 members.
4–14 Planning Storagesets Disk 10100 Disk 10000 A A' Disk 20100 Disk 20000 B B' Disk 30100 Disk 30000 C C' Mirror drives contain copy of data CXO5511A Figure 4–5. Mirrorsets Maintain Two Copies of the Same Data Considerations for Planning a Mirrorset Keep these points in mind as you plan your mirrorsets: ■ A controller can support up to 30 storagesets, consisting of mirrorsets and RAIDsets. Mirrorsets that are members of a stripeset count against this limitation (refer to Table 4–1).
4–15 ■ If you’re using more than one mirrorset in your subsystem, you should put the first member of each mirrorset on different buses as shown in Figure 4–6. (The first member of a mirrorset is the first disk drive you add.) When a controller receives a request to read data from a mirrorset, it typically accesses the first member of the mirrorset. Read access depends upon the read source switches, as described in “Read Source,” page 4–40.
4–16 Planning Storagesets ■ Place mirrorsets and RAIDsets on different ports to minimize risk in the event of a single port bus failure. ■ Mirrorset units are set to WRITEBACK_CACHE by default which increases a unit’s performance. ■ A storageset should only contain disk drives of the same capacity. The controller limits the capacity of each member to the capacity of the smallest member in the storageset.
4–17 I/O Request Chunk 1 2 Disk 10000 Chunk 1 4 3 Disk 20000 2 Parity for 3&4 4 Disk 30000 Parity for 1&2 3 CXO5509A Figure 4–7. Parity Ensures Availability; Striping Provides Good Performance Just as with stripesets, the I/O requests are broken into smaller “chunks” and striped across the disk drives until the request is read or written. But, in addition to the I/O data, chunks of parity data—derived mathematically from the I/O data—are also striped across the disk drives.
4–18 Planning Storagesets The relationship between the chunk size and the average request size determines if striping maximizes the request rate or the data-transfer rates. You can set the chunk size or let the controller set it automatically. See “Chunk Size,” page 4–43, for information about setting the chunk size. Considerations for Planning a RAIDset Keep these points in mind as you plan your RAIDsets: ■ A controller can support up to 20 storagesets, consisting of RAIDsets (refer to Table 4–1).
4–19 ■ RAIDsets are particularly well-suited for the following: ❏ Small to medium I/O requests ❏ Applications requiring high availability ❏ High read request rates ❏ Inquiry-type transaction processing ■ RAIDsets are not particularly well-suited for the following: ❏ Write-intensive applications ❏ Applications that require high data transfer capacity ❏ High-speed data collection ❏ Database applications in which fields are continually updated ❏ Transaction processing Using Striped Mirrorsets for Highest
4–20 Planning Storagesets Stripeset Mirrorset1 Mirrorset2 Disk 10100 Disk 20100 Disk 30100 A B C Disk 10000 Disk 20000 Disk 30000 B' C' A' Mirrorset3 CXO5508A Figure 4–8. Striping and Mirroring in the Same Storageset The failure of a single disk drive has no effect on this storageset’s ability to deliver data to the host and, under normal circumstances, it has very little effect on performance.
4–21 Cloning Data for Backup Use the CLONE utility to duplicate the data on any unpartitioned single-disk unit, stripeset, mirrorset, or striped mirrorset in preparation for backup. When the cloning operation is done, you can back up the clones rather than the storageset or single-disk unit, which can continue to service its I/O load. When you are cloning a mirrorset, CLONE does not need to create a temporary mirrorset.
4–22 Planning Storagesets Unit Unit Temporary mirrorset Disk10300 Disk10300 New member Unit Temporary mirrorset Unit Copy Disk10300 Disk10300 New member Clone Unit Clone of Disk10300 CXO5510A Figure 4–9. CLONE Steps for Duplicating Unit Members To clone a single-disk unit, stripeset, or mirrorset: 1. Establish a connection to the controller that accesses the unit you want to clone. 2.
4–23 3. When prompted, enter the unit number of the unit you want to clone. 4. When prompted, enter a unit number for the clone unit that CLONE will create. 5. When prompted, indicate how you would like the clone unit to be brought online: either automatically or only after your approval. 6. When prompted, enter the disk drives you want to use for the clone units. 7. Back up the clone unit. The following example shows the commands you would use to clone storage unit D204.
4–24 Planning Storagesets DISK10300(size=832317) (y,n) [y] ? y MIRROR DISK10300 C_MA SET C_MA NOPOLICY SET C_MA MEMBERS=2 SET C_MA REPLACE=DISK220 DEVICES AVAILABLE FOR CLONE TARGETS: DISK20400 (SIZE=832317) DISK30100 (SIZE=832317) Use available device DISK20400(size=832317) for member
4–25 DISK20000(size=832317) (y,n) [y] ? y MIRROR DISK20000 C_MB SET C_MB NOPOLICY SET C_MB MEMBERS=2 SET C_MB REPLACE=DISK20400 COPY IN PROGRESS FOR EACH NEW MEMBER. PLEASE BE PATIENT... . .
4–26 Planning Storagesets Backing Up Your Subsystem Configuration Your controller stores information about your subsystem configuration in its nonvolatile memory. This information could be lost if the controller fails or when you replace a module in your subsystem. You can avoid reconfiguring your subsystem manually by saving configuration information on one or more of your subsystem disks using the INITIALIZE SAVE_CONFIGURATION command.
4–27 Saving Subsystem Configuration Information to Multiple Disks You can save your subsystem configuration information to as many individual disks as you would like, but you must initialize each using the SAVE_CONFIGURATION switch. Use the following syntax for each: INITIALIZE DISKnnn SAVE_CONFIGURATION Saving Subsystem Configuration Information to a Storageset You can save your subsystem configuration information to a storageset.
4–28 Planning Storagesets Displaying the Status of the Save Configuration Feature You can use the SHOW THIS_CONTROLLER FULL command to find out if the save configuration feature is active and which devices are being used to store the configuration. The display includes a line indicating status and how many devices have copies of the configuration, as shown in the following example.
4–29 CACHE_FLUSH_TIMER = DEFAULT (10 seconds) Mirrored Cache: Not enabled Battery: FULLY CHARGED Expires: WARNING: UNKNOWN EXPIRATION DATE! WARNING: AN UNKNOWN NUMBER OF DEEP DISCHARGES HAVE OCCURRED! NOCCACHE_UPS Extended information: Terminal speed 19200 baud, eight bit, no parity, 1 stop bit Operation control: 00000000 Security state code: 6985 Configuration backup enabled on 1 devices The following example shows sample devices with the SAVE_CONFIGURATION switch enabled: $ SHOW DEVICES FULL Name Typ
4–30 Planning Storagesets Creating a Storageset Map Configuring your subsystem will be easier if you know how the storagesets correspond to the disk drives in your subsystem. You can see this relationship by creating a storageset map like the one shown in Figure 4–10. This storageset map is for a subsystem that contains two RAIDsets, two mirrorsets, and three disk drives in the spareset. Each enclosure also has redundant power supplies.
4–31 Power Supply Spare spare Spare R1 R1 R1 Power Supply Power Supply R2 R2 R2 R2 R2 R2 Power Supply Power Supply Spare Spare Spare Spare Spare Spare Power Supply Power Supply M1 M1 M2 M2 Spare Spare Power Supply Figure 4–10. Storageset Map Compaq HSG80 Array Controller ACS Version 8.
4–32 Planning Storagesets To create a storageset map: 1. Copy the template from Figure 4–11. 2. Establish a local or remote connection to one of the controllers in your subsystem. 3. Show the devices that are assigned to the controller. Use the following command: SHOW DEVICES 4. Locate each device assigned to the controller and record its location on your copy of the cabinet template. Use the following command: LOCATE device_name The LOCATE command causes the device’s LED to flash continuously. 5.
4–33 Power Supply Power Supply Power Supply Power Supply Power Supply Power Supply Power Supply Power Supply Figure 4–11. Storageset Map Template Compaq HSG80 Array Controller ACS Version 8.
4–34 Planning Storagesets Planning Partitions Use partitions to divide a storageset or disk drive into smaller pieces, which can each be presented to the host as its own storage unit. Figure 4–12 shows the conceptual effects of partitioning a single-disk unit. Partition 1 Partition 2 Partition 3 CXO-5316A-MC Figure 4–12. Partitioning a Single-Disk Unit You can create up to eight partitions per disk drive, RAIDset, mirrorset, stripeset, or striped mirrorset.
4–35 An unpartitioned storage unit has more capacity than a partition that uses the whole unit because each partition requires five blocks of administrative metadata. Thus, a single disk unit that contains one partition can store n-5 blocks of user or application data. See “Guidelines for Partitioning Storagesets and Disk Drives,” page 4–35, for information on manually partitioning a storageset or single-disk unit.
4–36 Planning Storagesets Choosing Switches for Storagesets and Devices Depending upon the kind of storageset or device you are configuring, you can enable the following options or “switches”: ■ RAIDset and mirrorset switches ■ Initialize switches ■ Unit switches ■ Device switches Enabling Switches If you use StorageWorks Command Console to configure the device or storageset, you can set switches from the command console screens during the configuration process.
4–37 RAIDset Switches You can enable the following switches to control how a RAIDset behaves to ensure data availability: ■ Replacement policy ■ Reconstruction policy ■ Membership Replacement Policy Specify a replacement policy to determine how the controller replaces a failed disk drive: ■ POLICY=BEST_PERFORMANCE (default) puts the failed disk drive in the failedset then tries to find a replacement (from the spareset) that is on a different device port than the remaining operational disk drives.
4–38 Planning Storagesets ■ RECONSTRUCT=NORMAL (default) balances the overall performance of the subsystem against the need for reconstructing the replacement disk drive. ■ RECONSTRUCT=FAST gives more resources to reconstructing the replacement disk drive, which may reduce the subsystem’s overall performance during the reconstruction task.
4–39 Mirrorset Switches You can enable the following switches to control how a mirrorset behaves to ensure data availability: ■ Replacement policy ■ Copy speed ■ Read source Replacement Policy Specify a replacement policy to determine how the controller replaces a failed disk drive: ■ POLICY=BEST_PERFORMANCE (default) puts the failed disk drive in the failedset then tries to find a replacement (from the spareset) that is on a different device port than the remaining operational disk drives.
4–40 Planning Storagesets ■ COPY=FAST allocates more resources to reconstructing the replacement disk drive, which may reduce the subsystem’s overall performance during the reconstruction task. Read Source Specify the read source to determine how the controller reads data from the members of a mirrorset: ■ READ_SOURCE=LEAST_BUSY (default) forces the controller to read data from the “normal” or operational member that has the least-busy work queue.
4–41 Device Switches When you add a disk drive or other storage device to your subsystem, you can enable the following switches: ■ Transportability ■ Transfer rate Transportability Indicate whether a disk drive is transportable when you add it to your subsystem: ■ NOTRANSPORTABLE disk drives (default) are marked with StorageWorksexclusive metadata. This metadata supports the error-detection and recovery methods that the controller uses to ensure data availability.
4–42 Planning Storagesets Transportable devices have these characteristics: ❏ Can be interchanged with any SCSI interface that does not use the device metadata, for example, a PC. ❏ Cannot have write-back caching enabled. ❏ Cannot be members of a shadowset, storageset, or spareset. ❏ Do not support forced errors.
4–43 Initialize Switches You can enable the following kinds of switches to affect the format of a disk drive or storageset: ■ Chunk size (for stripesets and RAIDsets only) ■ Save configuration ■ Destroy/Nodestroy After you initialize the storageset or disk drive, you cannot change these switches without reinitializing the storageset or disk drive.
4–44 Planning Storagesets Increasing the Request Rate A large chunk size (relative to the average request size) increases the request rate by allowing multiple disk drives to respond to multiple requests. If one disk drive contains all of the data for one request, then the other disk drives in the storageset are available to handle other requests. Thus, in principle, separate I/O requests can be handled in parallel, thereby increasing the request rate. This concept is shown in Figure 4–13.
4–45 Increasing the Data Transfer Rate A small chunk size relative to the average request size increases the data transfer rate by allowing multiple disk drives to participate in one I/O request. This concept is shown in Figure 4–14. Chunk size = 128k (256 blocks) A1 Request A A2 A3 A4 CXO-5172A-MC Figure 4–14. Chunk Size Smaller than the Request Size Applications such as CAD, image processing, data collection and reduction, and sequential file processing tend to require high data-transfer rates.
4–46 Planning Storagesets Maximum Chunk Size for RAIDsets Do not exceed the chunk sizes shown in Table 4–5 for a RAIDset. (The maximum chunk size is derived by 2048/(d – 1) where d is the number of disk drives in the RAIDset.) Table 4–5.
4–47 ■ SAVE_CONFIGURATION allows the controller to use 256K of each device in a storage unit to save the subsystem’s configuration. The controller saves the configuration every time you change it or add a patch to your controller. If the controller should fail, you can recover your latest configuration from the storage unit rather than rebuild it from scratch.
4–48 Planning Storagesets ■ If you previously configured storagesets with the SAVE_CONFIGURATION option, you do not need to initialize them again after you reconfigure your devices with a new controller. ■ When you replace a controller, make sure the replacement controller does not contain any configuration data. If the controller is not new, use the CONFIGURATION RESET command to purge any existing configuration.
4–49 Destroy/Nodestroy Specify whether to destroy or retain the user data and metadata when you initialize a disk drive that has been previously used in a mirrorset or as a single-disk unit. NOTE: The DESTROY and NODESTROY switches are only valid for striped mirrorsets and mirrorsets. ■ DESTROY (default) overwrites the user data and forced-error metadata on a disk drive when it is initialized. ■ NODESTROY preserves the user data and forced-error metadata when a disk drive is initialized.
5–1 Chapter 5 Configuring an HSG80 Array Controller This chapter explains how to configure an HSG80 array controller and the modules that support its operation in a StorageWorks subsystem. Compaq HSG80 Array Controller ACS Version 8.
5–2 Configuring an HSG80 Array Controller Single Controller “Configuring a Single Controller,” page 5–6 “Connecting a Single Controller to the Host,” page 5–8 Continue creating units until the configuration is complete.
5–3 Introduction Unless you specifically requested a preconfigured subsystem, you will have to configure your controller and its subsystem before you can use them. Figure 5–1 shows the configuration process of how to configure the controller followed by configuring storagesets and units.
5–4 Configuring an HSG80 Array Controller Configuring and Cabling an HSG80 Array Controller You can use these procedures to configure and cable your controller in one of three configurations: ■ “Configuring a Single Controller,” page 5–6 ❏ “Connecting a Single Controller to the Host,” page 5–8 ■ “Configuring in Transparent Failover Mode,” page 5–12 ❏ “Connecting in Transparent Failover Mode to the Hosts,” page 5–14 ■ “Configuring in Multiple-Bus Failover Mode,” page 5–20 ❏ “Connecting in Multiple-Bus
5–5 Tips As you configure the controller, keep these points in mind: ■ Cabling—You can either configure the controller with a maintenance port cable and no optical host bus cables connected to the host, or you can configure with the optical host bus cables connected: ❏ SWCC—You can configure the controller using the SWCC graphical user interface. See the Command Console for HSG80 Solutions Getting Started Guide for more details.
5–6 Configuring an HSG80 Array Controller Configuring a Single Controller Follow these steps to configure an HSG80 array controller: NOTE: This procedure has been written for first-time configuring. However, you can adapt the procedure when you reconfigure the controllers. For replacing and upgrading the controllers, see the HSG80 Array Controller ACS Version 8.3 and 8.4 Maintenance and Service Guide. 1.
5–7 7. Set the time on the controller, which provides a baseline for replacing the external battery, using the following command: SET THIS_CONTROLLER TIME=DD-MMM-YYYY:HH:MM:SS NOTE: All values are numbers, except for MMM, which uses the first three letters of the month. When setting hours, minutes, and seconds—HH, MM, and SS—you must use 24-hour time. 8. Set up the battery discharge timer: a.
5–8 Configuring an HSG80 Array Controller Connecting a Single Controller to the Host For a detailed description of the controller and cables, see Chapter 1, “General Description.” Using Two Switches 4 1 2 5 3 5 4 CXO6881A Figure 5–2. Example Cabling for a Single Configuration using Two Switches Table 5–1.
5–9 Before cabling, note the following: ■ Ports ❏ The active ports are port 1 and port 2, controller A. ❏ Standby ports do not exist in single-controller configurations when a failover mode has not been selected. ❏ Once logical units are added, the active ports will have access to the units.
5–10 Configuring an HSG80 Array Controller Using One Switch 4 5 1 2 3 CXO6880A Figure 5–3. Example Cabling for a Single Configuration using One Switch Table 5–2.
5–11 Before cabling, note the following: ■ Ports ❏ The active ports are port 1 and port 2, controller A. ❏ Standby ports do not exist in single-controller configurations when a failover mode has not been selected. ❏ Once logical units are added, the active ports will have access to the units.
5–12 Configuring an HSG80 Array Controller Configuring in Transparent Failover Mode You can configure dual-redundant controllers to your hosts with one or more switches. Follow these steps to configure an HSG80 array controller: NOTE: This procedure has been written for first-time configuring. However, you can adapt the procedure when you reconfigure the controllers. For replacing and upgrading the controllers, see the HSG80 Array Controller ACS Version 8.3 and 8.4 Maintenance and Service Guide. 1.
5–13 7. Put “this controller” into transparent failover mode. Use the following syntax: SET FAILOVER COPY = THIS_CONTROLLER The “other controller” inherits “this controller’s” configuration, then restarts.Wait for it to return to normal operation before continuing. See details about failover modes in “Selecting a Failover Mode,” page 2–5. 8.
5–14 Configuring an HSG80 Array Controller 11. Enter any other optional CLI commands for your configuration. See “Optional Steps,” page 5–29 and Chapter 7, “CLI Commands.” 12. Display details about the controller you configured. Use the following command: SHOW THIS_CONTROLLER FULL See “SHOW,” page 7–133.
5–15 5 1 3 6 4 2 6 5 CXO6837A Figure 5–4. Example Cabling in Transparent Failover using Two Switches Table 5–3.
5–16 Configuring an HSG80 Array Controller The following steps show how to connect a pair of dual-redundant controllers to the host using two switches. Before cabling, note the following: ■ Ports ❏ The active ports are port 1, controller A and port 2, controller B. ❏ The standby ports are port 2, controller A and port 1, controller B. ❏ Once logical units are added, the active ports will have access to the units. ❏ For a more detailed explanation of active vs.
5–17 5. Connect the host bus cable between the switches and the host adapters. IMPORTANT: Figure 5–4 shows two cables for a host connection. The cables can go either to the same host or they can go to separate hosts. 6. Route and tie the cables as desired. 7. Turn on the subsystem. Some operating systems may require you to start the host to see the devices attached to the new controller. See the HSG80 Array Controller ACS Version 8.X Release Notes for your operating system.
5–18 Configuring an HSG80 Array Controller 5 1 3 6 4 2 5 CXO6836A Figure 5–5. Example Cabling for Transparent Failover with One Switch Table 5–4.
5–19 The following steps show how to connect a pair of dual-redundant controllers to the host using one switch. Before cabling, note the following: ■ Ports ❏ The active ports are port 1, controller A and port 2, controller B. ❏ The standby ports are port 2, controller A and port 1, controller B. ❏ Once logical units are added, the active ports will have access to the units. ❏ For a more detailed explanation of active vs.
5–20 Configuring an HSG80 Array Controller Configuring in Multiple-Bus Failover Mode IMPORTANT: If you are using HSG80 ACS V8.4P code, see the Data Replication Manager HSG80 ACS Version 8.4P Guide to Operations for cabling yourcontroller into the subsystem. Follow these steps to configure an HSG80 array controller: NOTE: This procedure has been written for first-time configuring. However, you can adapt the procedure when you reconfigure the controllers.
5–21 6. On OpenVMS hosts only, change the SCSI version to SCSI-3 mode: SET THIS_CONTROLLER SCSI_VERSION=SCSI-3 IMPORTANT: Once you change to the SCSI-3 version in OpenVMS, you must use a special number called a controller identifier, which makes the controller and the command console LUN visible to the host. See the IDENTIFIER switch with the command “SET controller,” page 7–91. 7. Set the port topology for each port.
5–22 Configuring an HSG80 Array Controller 11. Set up the battery discharge timer: a. Start Frutil using the following command: RUN FRUTIL Frutil displays the following: Do you intend to replace this controller’s cache battery? Y/N b. Enter Y(es). Frutil displays a three-step procedure and prompts you to press return. c. Press return. 12. Enter any other optional CLI commands for your configuration. See “Optional Steps,” page 5–29 and Chapter 7, “CLI Commands.” 13.
5–23 Connecting in Multiple-Bus Failover Mode to the Hosts IMPORTANT: If you are using HSG80 ACS V8.4P code, see the Data Replication Manager HSG80 ACS Version 8.4P Guide to Operations for cabling yourcontroller into the subsystem. You can connect each controller pair to your hosts with two switches. Using Two Switches Follow the procedures below to properly connect your dual-redundant controllers to your host. Figure 5–6 and Figure 5–7 illustrate two different choices on how your setup should appear.
5–24 Configuring an HSG80 Array Controller 5 6 1 4 3 2 6 5 CXO6838A Figure 5–6. Example Cabling in Multiple-Bus Failover: Configuration 1 Table 5–5.
5–25 Before cabling, note the following: ■ Ports ❏ All ports are active. Once units are added, all ports will have access to the same units. ❏ Port 1, controller A and port 2, controller B connect into the same switch. ❏ Port 1 controller B and port 2, controller A connect into the same switch. ❏ Port 1 on both controllers connect into the same switch port on different switches. ❏ Port 2 on both controllers connects into the same switch port on different controllers.
5–26 Configuring an HSG80 Array Controller 7. Turn on the subsystem. Some operating systems may require you to start the host to see the devices attached to the new controller. See the HSG80 Array Controller ACS Version 8.X Release Notes for your operating system. 5 1 3 6 4 2 6 5 CXO6872A Figure 5–7. Example Cabling in Multiple-Bus Failover: Configuration 2 Table 5–6.
5–27 Before cabling, note the following: ■ Ports ❏ All ports are active. Once units are added, all ports will have access to the same units. ❏ Port 1, controller A and port 2, controller A connect into the same switch. ❏ Port 1 controller B and port 2, controller B connect into the same switch. ❏ Port 1 on both controllers connect into the same switch port on different switches. ❏ Port 2 on both controllers connects into the same switch port on different controllers.
5–28 Configuring an HSG80 Array Controller 5. Connect the host bus cable between the switches and the host adapters. IMPORTANT: Cables going into the same host must connect into the same switch port. 6. Route and tie the cables as desired. 7. Turn on the subsystem. Some operating systems may require you to start the host to see the devices attached to the new controller. See the HSG80 Array Controller ACS Version 8.X Release Notes for your operating system.
5–29 Optional Steps Following are some of the more common steps that can be performed anytime during configuring. Changing the CLI Prompt Enter a 1- to 16-character string as the new prompt. For example, you could use the prompt to indicate the controller’s name, such as “HSG1.” Type the following command: SET THIS_CONTROLLER PROMPT = “new prompt” If you are configuring dual-redundant controllers, you can also change the CLI prompt on it.
5–30 Configuring an HSG80 Array Controller First Expansion Enclosure Master Enclosure Second Expansion Enclosure SCSI Target ID = 11 SCSI Target ID = 3 SCSI Target ID = 15 SCSI Target ID = 10 SCSI Target ID = 2 SCSI Target ID = 14 SCSI Target ID = 9 SCSI Target ID = 1 SCSI Target ID = 13 SCSI Target ID = 8 SCSI Target ID = 0 SCSI Target ID = 12 EMU PVA 2 PVA 0 EMU Controller A Controller B Cache A Cache B EMU PVA 3 NOTE: SCSI target IDs 4 and 5 are reserved.
5–31 Setting a Port Offline NOTE: By default, all ports are online and able to access targets and logical units, even when they are designated as standby ports in transparent failover mode. To take a port offline and turn off its access to targets and logical units, use the following command: SET THIS_CONTROLLER PORT_2_TOPOLOGY=OFFLINE See “SET controller,” page 7–91 on how to use the offline switch.
5–32 Configuring an HSG80 Array Controller Changing the Host Default Value By default, the host is Windows NT. However, when two or more of the same host exists, the default will be that host. For example, if a system is configured using three Tru64 UNIX hosts, then the default host will be the Tru64 UNIX operating system.
5–33 Establishing a Local Connection to the Controller You can communicate with a controller locally or remotely. Use a local connection to configure the controller for the first time. Use a remote connection to your host system for all subsequent configuration tasks. See the Quick Setup Guide that came with your platform kit for details. The maintenance port provides a convenient way to connect a PC or terminal to the controller so that you can troubleshoot and configure it.
5–34 Configuring an HSG80 Array Controller 1 2 3 4 5 6 2 1 CXO6476B Figure 5–9. Terminal to Local-Connection Port Connection Table 5–7.
5–35 CAUTION: The local-connection port described in this book generates, uses, and can radiate radio-frequency energy through cables that are connected to it. This energy may interfere with radio and television reception. Do not leave any cables connected to it when you are not communicating with the controller. Follow these steps to establish a local connection for setting the controller’s initial configuration: 1. Turn off the PC or terminal, and connect it to the controller, as shown in Figure 5–9. a.
5–36 Configuring an HSG80 Array Controller Shutting Down Your Subsystem Follow these steps to shut down your StorageWorks subsystem: 1. On the host, dismount the storage units in your subsystem. 2. Connect a maintenance terminal to one of the controllers in your subsystem. 3. Shut down the controllers. If you have dual-redundant controllers, shut down the “other controller” first, then shut down “this controller.
5–37 Restarting Your Subsystem Follow these steps to restart your subsystem: 1. Plug in the subsystem’s power cord. 2. Turn on the subsystem. 3. Press and hold the reset button on the controller for three seconds, then release it. 4. Check the status of the write-back cache module’s backup battery. If your subsystem has been off for an extended period of time, the battery may be drained.
6–1 Chapter 6 Configuring Storagesets This chapter shows you how to configure storagesets with CLI Commands and the Command Console LUN. It outlines the procedures for partitioning storagesets and disk drives and how to enable autospare. You will also learn how to configure stripesets, mirrorsets, and striped mirrorsets. Compaq HSG80 Array Controller ACS Version 8.
6–2 Configuring Storagesets Configuring Storagesets with CLI Commands One method of configuring storagesets is manual configuration. This method allows you the most flexibility in defining and naming storagesets. See Chapter 7, “CLI Commands.” for complete information about the CLI commands shown in this chapter. Unit Switches You can enable the Unit switches listed in Table 6–1 for the listed storagesets and devices.
6–3 Adding Disk Drives The factory-installed devices in your StorageWorks subsystem have already been added to the controller’s list of eligible devices.
6–4 Configuring Storagesets ■ If the power fails or the bus is reset while HSUTIL is formatting a disk drive, the drive may become unusable. To minimize this possibility, Compaq recommends you use a reliable power source and suspend all non-HSUTIL activity to the bus that services the target disk drive. ■ HSUTIL cannot control or affect the defect management for a disk drive. ■ Do not invoke any CLI command or run any local program that might reference the target disk drive while HSUTIL is active.
6–5 2. DEVICE_CODE_LOAD_DISK 3.
6–6 Configuring Storagesets Configuring a Stripeset See “Using Striped Mirrorsets for Highest Performance and Availability,” page 4–19, for information about creating a profile and understanding the switches you can set for this kind of storage unit. To configure a stripeset: 1. Create the stripeset by adding its name to the controller’s list of storagesets and specifying the disk drives it contains. Use the following command: ADD STRIPESET stripeset-name disknnnnn disknnnnn 2.
6–7 The following example shows the commands you would use to create Stripe1, a three-member stripeset: ADD STRIPESET STRIPE1 disk10000 disk20000 disk30000 INITIALIZE STRIPE1 CHUNKSIZE=128 ADD UNIT D100 STRIPE1 MAXIMUM_CACHED_TRANSFER=16 SHOW STRIPE1 SHOW D100 See Chapter 7, “CLI Commands,” for more information on stripeset switches and values.
6–8 Configuring Storagesets 1–9999. Usually, this number is the same as the unit number set in step 3. SET unit-number identifier=nnnn 4. Verify the mirrorset configuration and switches. Use the following command: SHOW mirrorset-name 5. Verify the unit configuration and switches.
6–9 2. Initialize the RAIDset. Optional: If you want to set the Initialize switches, you must do so in this step. Use the following command: INITIALIZE RAIDset-name switch NOTE: It is recommended that you allow initial reconstruct to complete before allowing I/O to the RAIDset. Not doing so may generate forced errors at the host level. To determine whether initial reconstruct has completed, enter SHOW S or SHOW RAIDSET FULL. 3.
6–10 Configuring Storagesets Chapter 7, “CLI Commands,” contains more information on valid switches and values for configuring a RAIDset. Configuring a Striped Mirrorset See Chapter 4, “Planning Storagesets,” for information about creating a profile and understanding the switches you can set for this kind of storage unit. To configure a striped mirrorset: 1. Create—but do not initialize—at least two mirrorsets. 2. Create a stripeset and specify the mirrorsets it contains.
6–11 6. Verify the unit configuration and switches.
6–12 Configuring Storagesets Configuring a Single-Disk Unit Follow these steps to use a single disk drive as a single-disk unit in your subsystem: 1. Add the disk drive by following the steps in “Adding Disk Drives,” page 6–3. Optionally, you can append Device switch values. If you do not specify switch values, the default values are applied. 2. Initialize the disk drive using the following command: INITIALIZE DISknnn switch 3.
6–13 Partitioning a Storageset or Disk Drive See “Planning Partitions,” page 4–34, for details about partitioning a storage unit. To partition a storageset or disk drive: 1. Add the storageset or disk drive to the controller’s list of storagesets and specify the disk drives it contains. Use the following command: ADD storageset-name DISKnnnnn DISKnnnnn or ADD DISK DIsknnnnn ptl-location Do not split partitioned units across ports. They must be on a single port.
6–14 Configuring Storagesets Use the following command to present partitions to the host: ADD UNIT unit-number storageset-name PARTITION=partition-number switch 6.
6–15 ADD UNIT D1 Raid1 PARTITION=1 ADD UNIT D2 Raid1 PARTITION=2 ADD UNIT D3 Raid1 PARTITION=3 ADD UNIT D4 Raid1 PARTITION=4 SHOW Raid1 . . . Partition number Size Starting Block 1 1915 (0.98 MB) 0 2 1915 (0.98 MB) 1920 3 1915 (0.98 MB) 3840 4 2371 (1.21 MB) 5760 . . . Used by D1 D2 D3 D4 Chapter 7, “CLI Commands,” contains more information on partitioning a storageset or disk drive.
6–16 Configuring Storagesets 2. Verify the contents of the spareset using the following command: SHOW SPARESET Example The following example shows the commands you would use to add DISK60000 and DISK60100 to the spareset. ADD SPARESET disk60000 ADD SPARESET disk60100 SHOW SPARESET Removing a Disk Drive from the Spareset You cannot delete the spareset—it always exists whether or not it contains disk drives.
6–17 Example The following example shows the commands you would use to remove DISK60000 from the spareset. SHOW SPARESET Name SPARESET Storageset spareset Uses disk60000 disk60100 Used by Uses disk60100 Used by DELETE SPARESET disk60000 SHOW SPARESET Name SPARESET Storageset spareset Enabling Autospare With AUTOSPARE enabled on the failedset, any new disk drive that is inserted into the PTL location of a failed disk drive is automatically initialized and placed into the spareset.
6–18 Configuring Storagesets Deleting a Storageset If the storageset you are deleting is partitioned, you must delete each partitioned unit before you can delete the storageset. Use the following steps to delete a storageset: 1. Show the configuration using the following command: SHOW STORAGESETS 2. Delete the unit number shown in the “Used by” column. Use the following command: DELETE unit-number 3. Delete the name shown in the “Name” column. Use the following command: DELETE storageset-name 4.
6–19 Changing Switches for a Storageset or Device You can optimize a storageset or device at any time by changing the switches that are associated with it. See “Choosing Switches for Storagesets and Devices,” page 4–36, for an explanation of the switches. Remember to update the storageset’s profile when you change its switches.
6–20 Configuring Storagesets Changing Initialize Switches The Initialize switches cannot be changed without destroying the data on the storageset or device. These switches are integral to the formatting and can only be changed by reinitializing the storageset. Initializing a storageset is similar to formatting a disk drive; all data is destroyed during this procedure. Changing Unit Switches Use the SET command to change Unit switches that are associated with a unit.
6–21 Configuring with the Command Console LUN The Command Console LUN (CCL) is a type of LUN that allows you to communicate with the controller from the host using StorageWorks Command Console (SWCC) or CLI commands instead of using the maintenance port cable.
6–22 Configuring Storagesets To turn it off, use the following command: SET THIS_CONTROLLER NOCOMMAND_CONSOLE_LUN CAUTION: Disabling the CCL while SWCC is running may result in loss of connection for the StorageWorks Command Console. Turn off SWCC before issuing the command. Finding the CCL Location To see where each CCL is located, use the following commands: SHOW THIS_CONTROLLER or SHOW OTHER_CONTROLLER Look under host port to find the Command Console LUN location.
6–23 Multiple-Bus Failover If SET MULTIBUS_FAILOVER is enabled, all ports will be able to see and access the CCL. In addition, the CCL appears to the host as a direct access device. IMPORTANT: A port will not be able to see the CCL if a host communication has a unit offset higher than the location of the CCL. Transparent Failover If you are in SCSI-2 mode and have enabled the SET FAILOVER command, only one CCL will be enabled. Either host port 1 or host port 2 will see it, but not both.
6–24 Configuring Storagesets CAUTION: Selecting SCSI-3 mode enables access to the CCL by all hosts. If the hosts access the CCL simultaneously, unpredictable consequences can occur. In cases where the CCL can be accessed through multiple paths and LUNs, system administrators of each host must not attempt to access the CCL simultaneously. Adding Storage Units Using the CCL To start configuring storage units, you must first assign unit offsets. See “Assigning Unit Offsets,” page 3–13.
6–25 Moving Storagesets You can move a storageset from one subsystem to another without destroying its data as shown in Figure 6–1. You also can follow the steps in this section to move a storageset to a new location within the same subsystem. CAUTION: Move only normal storagesets. Do not move storagesets that are reconstructing or reduced, or data corruption will result.
6–26 Configuring Storagesets Use the following procedure to move a storageset while maintaining the data it contains: 1. Show the details for the storageset you want to move. Use the following command: SHOW storageset-name 2. Label each member with its name and PTL location. If you do not have a storageset map for your subsystem, you can enter the LOCATE command for each member to find its PTL location.
6–27 7. Add again each disk drive to the controller’s list of valid devices. Use the following command: ADD DISK disk-name PTL-location ADD DISK disk-name PTL-location ADD DISK disk-name PTL-location 8. Recreate the storageset by adding its name to the controller’s list of valid storagesets and specifying the disk drives it contains. (Although you have to recreate the storageset from its original disks, you do not have to add them in their original order.
6–28 Configuring Storagesets (...move the disk drives to their new location...) ADD DISK DISK20000 2 0 0 ADD DISK DISK30000 3 0 0 ADD DISK DISK40000 4 0 0 ADD RAIDSET RAID99 DISK20000 DISK30000 DISK40000 ADD UNIT D100 RAID99 Example The following example moves the reduced RAIDset, R3, to another cabinet. (R3 used to contain DISK20000, which failed before the RAIDset was moved. R3 contained DISK10000, DISK30000, and DISK40000 at the beginning of this example.
7–1 Chapter 7 CLI Commands Command Line Interpreter (CLI) commands allow you to interact with your controller. In the overview, you will learn about CLI and how to use it. The rest of the chapter details the syntax of each command and shows examples of how they are used. Compaq HSG80 Array Controller ACS Version 8.
7–2 CLI Commands CLI Overview The Command Line Interpreter (CLI) is one of the user interfaces through which you control your StorageWorks array controller in the StorageWorks subsystem. The CLI commands allow you to manage the subsystem by viewing and modifying the configuration of the controller and the devices attached to them. You can also use the CLI to start controller diagnostic and utility programs.
7–3 Command Overview The CLI consists of six basic command types: ■ Controller Commands—Configure the controller’s SCSI ID numbers, maintenance terminal characteristics, CLI prompt, and so forth. Controller commands are also used to shut down and restart the controller. ■ Device Commands—Create and configure containers made from physical devices attached to the controller. ■ Storageset Commands—Create and configure complex containers made from groups of device containers.
7–4 CLI Commands Entering CLI Commands Use the following tips and techniques when entering CLI commands: ■ Commands are not case sensitive. ■ For most commands, you only need to enter enough of the command to make the command unique. For example, SHO is the same as entering SHOW. ■ The controller processes each command in sequence. You can continue entering subsequent commands while the controller is processing prior commands.
7–5 Table 7–1. Recall and Edit Command Keys (Continued) Key Function Ctrl/R Recalls the contents of the command line. This is especially helpful if the system issues a message that interrupts your typing. Ctrl/U Deletes all characters on the same line as the cursor. Changing the CLI Prompt You can change the CLI prompt that displays. Use the SET THIS_CONTROLLER PROMPT= command. Enter a 1- to 16- character string as the new prompt.
7–6 CLI Commands ADD CONNECTIONS Adds the specified host connection to the table of known connections. This table is maintained in NVRAM. The maximum table length is 32 connections; if the table contains 32 entries, new connections cannot be added unless some old ones are deleted. There are two mechanisms for adding a new connection to the table, as follows: ■ Physically connecting a host adapter to a controller port.
7–7 NOTE: The default connection name is assigned automatically by the controller when the connection is physically made between a host adapter and a controller port. Default connection names are assigned only by the controller. HOST_ID=nnnn-nnnn-nnnn-nnnn Host_ID is the worldwide name of the host. It is a 16-character hexadecimal number. The hyphens aren’t necessary, but are recommended to avoid mistakes in entering the number.
7–8 CLI Commands In transparent failover mode and normal mode, host connections on controller port 1 have an offset of 0 and host connections on controller port 2 have an offset of 100. These are the default offset values. The relationship between LUN number, unit number, and offset is as follows: ■ LUN number = unit number - offset. ■ Logical unit number or LUN number = the logical unit number presented to the host connection. ■ Unit number = the number assigned to the unit in the ADD UNIT command.
7–9 Examples This will add to the table of known connections, an entry for a connection named George with the indicated host and adapter worldwide names, on port 2 of “this controller.” CLI>ADD CONNECTIONS GEORGE HOST_ID=1000-0000-C920-1234 ADAPTER_ID=1000-0000-C920-5678 CONTROLLER=THIS PORT=2 See also ADD UNIT DELETE connections SET connection-name Compaq HSG80 Array Controller ACS Version 8.
7–10 CLI Commands ADD DISK Names a disk drive and adds it to the controller’s configuration. NOTE: The controller supports a maximum of 72 storage devices, even though more than 72 target IDs are available. Do not exceed the maximum number of devices in the subsystem. Syntax ADD DISK container-name scsi-port-target-lun Parameters container-name Assigns a name to the disk device. This is the name used with the ADD UNIT command to create a single-disk unit.
7–11 Switches NOTRANSPORTABLE (Default) TRANSPORTABLE Indicates whether a disk drive can be accessed exclusively by StorageWorks controllers. If the NOTRANSPORTABLE switch is specified, the controller makes a small portion of the disk inaccessible to the host. This restricted space is used to store information (metadata) that is used to improve data reliability, error detection, and the ability to recover data.
7–12 CLI Commands Examples To add DISK10000 at port 1, target 0, LUN 0, type: ADD DISK DISK10000 1 0 0 To add DISK40200 as a transportable disk drive to port 4, target 2, LUN 0, use: ADD DISK DISK40200 4 2 0 TRANSPORTABLE To add a disk drive named DISK30200 as non-transportable disk to port 3, target 2, LUN 0, and to set the data transfer rate to 10 MHz, enter the following command on one line.
7–13 ADD MIRRORSET Names a mirrorset and adds it to the controller configuration. Syntax ADD MIRRORSET mirrorset-name disk-name1 [disk-nameN] Parameters mirrorset-name Assigns a name to the mirrorset. This is the name used with the ADD UNIT command to identify the mirrorset as a host-addressable unit. The mirrorset name must start with a letter (A through Z) and may consist of a maximum of 9 characters including letters A through Z, numbers 0 through 9, periods (.), dashes (-), or underscores (_).
7–14 CLI Commands Specify COPY=FAST to allow the creation of mirrored data to take precedence over other controller operations. When you specify COPY=FAST, the controller uses more resources to create the mirrored data, and copying takes less time. However, overall controller performance is reduced during copying. Specify COPY=NORMAL when operations performed by the controller should take priority over the copy operation.
7–15 Specify the READ_SOURCE=disk-name of a specific member to which you want the controller to direct all read requests. If the member fails out of the mirrorset, the controller selects the first normal member it finds to satisfy its read requests. Specify READ_SOURCE=LEAST_BUSY to direct read requests to the mirrorset disk with the least amount of work in its queue.
7–16 CLI Commands See also ADD DISK ADD UNIT DELETE container-name INITIALIZE MIRROR REDUCE SHOW mirrorset-name SHOW MIRRORSETS SHOW STORAGESETS UNMIRROR
7–17 ADD RAIDSET Names a RAIDset and adds the RAIDset to the controller’s configuration. Compaq RAIDsets are often referred to as RAID level 3/5 storagesets because they use the best characteristics of RAID level 3 and RAID level 5. The number of members in the storageset is determined by the number of containers specified by the container-name parameter in the command. The data capacity of the RAIDset is determined by the storage size of the smallest member.
7–18 CLI Commands Switches POLICY=BEST_FIT POLICY=BEST_PERFORMANCE (Default) NOPOLICY Set the selection criteria the controller uses to choose a replacement member from the spareset when a RAIDset member fails. Specify POLICY=BEST_FIT to choose a replacement disk drive from the spareset that equals or exceeds the base member size (smallest disk drive at the time the RAIDset was initialized) of the remaining members of the RAIDset.
7–19 Specify NORMAL to balance other controller operations with the reconstruct operation. The controller uses relatively few resources to perform the reconstruct process; therefore, there is little impact on performance. REDUCED NOREDUCED (Default) Permits the addition of a RAIDset missing a member. Specify the REDUCED switch when you add a reduced RAIDset (a RAIDset that is missing a member).
7–20 CLI Commands ADD UNIT D70 RAID8 This example shows how you can create a three-member RAIDset from the members of a reduced four-member RAIDset. Do not initialize the RAIDset again. CAUTION: Data contained on the RAIDset will be erased if you reinitialize the RAIDset.
7–21 ADD REMOTE_COPY_SETS IMPORTANT: This command works in a remote copy environment only using HSG80 ACS V8.4P code, where local (initiator) and remote (target) sites coexist. Creates a remote copy set (RCS) in multiple-bus failover that comprises one unit at the initiator site and one unit at the target site. The units may be a single disk or part of a storageset, mirrorset, or RAIDset. Up to 2 units may be assigned to each remote copy set for a total of 8 copy sets.
7–22 CLI Commands The combined initiator unit-name target path-name must be typed as follows: ■ initiator unit name controller pair name\remote unit name Examples Create a host-addressable unit D1 on the local controller: ADD DISK DISK10000 1 0 0 ADD DISK DISK20100 2 1 0 ADD STRIPESET STRIPE1 DISK10000 DISK20100 INITIALIZE STRIPE1 ADD UNIT D1 STRIPE1 Create a host-addressable unit D21 on the remote controller: ADD DISK DISK30200 3 2 0 ADD DISK DISK40300 4 3 0 ADD STRIPESET STRIPE1 DISK30200 DISK40300 I
7–23 ADD SPARESET Adds a disk drive to the spareset. Syntax ADD SPARESET disk-name Parameter disk-name Indicates the name of the disk drive being added to the spareset. Only one disk drive can be added to the spareset with each ADD SPARESET command.
7–24 CLI Commands ADD STRIPESET Names a stripeset and adds it to the controller configuration. Stripesets are sometimes referred to as RAID level 0 storagesets. The number of members in the stripeset is determined by the number of container-name parameters specified. Syntax ADD STRIPESET stripeset-name container-name1 container-name2 [containernameN] Parameters stripeset-name Assigns a name to the stripeset.
7–25 Examples To create a stripeset named STRIPE1 with three disks: DISK10000, DISK20100, and DISK30200, enter: ADD DISK DISK10000 1 0 0 ADD DISK DISK20100 2 1 0 ADD DISK DISK30200 3 2 0 ADD STRIPESET STRIPE1 DISK10000 DISK20100 DISK30200 To create a stripeset named STRIPE1 and then create a logical unit from it, type: INITIALIZE STRIPE1 ADD UNIT D103 STRIPE1 This example shows how to create a two-member striped mirrorset (a stripeset whose members are mirrorsets), and how to create a logical unit from i
7–26 CLI Commands See also ADD UNIT ADD MIRRORSET DELETE container-name INITIALIZE SHOW STORAGESET SHOW STRIPESET SHOW stripeset-name
7–27 ADD UNIT Creates a logical unit from a device, container, or partition. The controller maps all requests from the host to the logical-unit number as requests to the container specified in the ADD UNIT command. If you add a newly-created storageset or disk to your subsystem, you must initialize it before it can be added as a logical unit. If you are adding a storageset or disk that has data on it that you want to maintain, do not initialize it; it will be added as logical unit.
7–28 CLI Commands In transparent failover mode, adding unit D0 creates a logical unit and presents it as D0 to the host on port 1. Adding unit D100 creates a logical unit and presents it as D100 to the host on port 2. Units must be on a single port. Do not split partitioned units across ports. In multiple-bus failover mode, adding unit D0 creates a logical unit that is presented to both ports on that controller.
7–29 MAXIMUM_CACHED_ TRANSFER_SIZE PREFERRED_PATH NOPREFERRED_PATH READ_CACHE NOREAD_CACHE READAHEAD_CACHE NOREADAHEAD_CACHE WRITE_PROTECT NOWRITE_PROTECT WRITEBACK_CACHE NOWRITEBACK_CACHE RUN NORUN RAIDset ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔ Stripeset ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔ Mirrorset ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔ NoTransportable Disk ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔ Transportable Disk ✔ ✔ ✔ ✔ ✔ ✔ Switch ENABLE_ACCESS_PATH DISABLE_ACCESS_PATH PARTITION Table 7–2.
7–30 CLI Commands If you enable another host ID, previously enabled hosts are not disabled. The new IDs are added. If you wish to enable only certain IDs, disable all access paths (DISABLE_ACCESS_PATH=ALL), then enable the desired IDs. The system will display the following message: Warning 1000: Access IDs in addition to the one(s) specified are still enabled. If you wish to enable ONLY the id(s) listed, disable all access paths (DISABLE_ACCESS_PATH=ALL), then enable the ones previously listed.
7–31 PREFERRED_PATH=OTHER_CONTROLLER PREFERRED_PATH=THIS_CONTROLLER NOPREFERRED_PATH (Default) Used only in Windows NT operating systems. May be set only when dual-redundant controllers are operating in a multiple-bus failover configuration. In a multiple bus failover configuration, the host determines which controller the units are accessed through. The host’s unit-to-controller settings always take precedence over the preferred path assigned to units with this switch.
7–32 CLI Commands READ_CACHE (Default) NOREAD_CACHE Sets the controller’s cache read policy function. Read caching improves performance in almost all situations. Therefore, it is recommended you leave its default setting, READ_CACHE enabled. However, under certain conditions, such as when performing a backup, read caching may not be necessary since only a small amount of data is cached.
7–33 WRITE_PROTECT NOWRITE_PROTECT (Default) Tells the controller whether data contained on the unit can be overwritten. Specify WRITE_PROTECT to prevent the host from writing data to the unit. However, the controller may still write to a write-protected RAIDset to complete a reconstruct operation and metadata, reconstruct data, and copy data may still be written to RAIDsets and mirrorsets. Specify NOWRITE_PROTECT to allow the host to write data to the unit.
7–34 CLI Commands Examples This example shows how to create unit D102 from a single-disk drive named DISK10000 and sets the host’s access to the unit through “this controller.” ADD DISK DISK10000 1 0 0 INITIALIZE DISK10000 ADD UNIT D102 DISK10000 PREFERRED_PATH=THIS_CONTROLLER This example shows how to create unit D107 from a RAIDset named RAID9 and instructs the unit to take advantage of the controller’s write-back caching feature.
7–35 CLEAR_ERRORS CLI Stops the display of current or previous error messages at the CLI prompt. This command does not clear the error conditions, it only stops the display of errors at the CLI prompt. After the cause of the error condition has been corrected, issue the CLEAR_ERRORS CLI command to clear the error message. NOTE: There are three message types: info—general information; warning—user may want to examine, but command will be executed; and error—command will not execute.
7–36 CLI Commands CLEAR_ERRORS controller INVALID_CACHE Clears an invalid cache error and allows the controller and cache to resume operation. If the error is due to an incorrectly-mirrored configuration, the controller indicates mirrored mode status after the error is cleared. Use this command for the following situations: ■ When the controller or cache modules have been replaced, resulting in mismatched data between the controllers.
7–37 Specify DESTROY_UNFLUSHED_DATA to retain the controller information and discard the cached data. Specify NODESTROY_UNFLUSHED_DATA in the following situations: ■ If the controller module has been replaced ■ If the controller’s nonvolatile memory (NVMEM) has lost its contents.
7–38 CLI Commands See also CLEAR_ERRORS CLI CLEAR_ERRORS LOST_DATA CLEAR_ERRORS UNKNOWN CLEAR_ERRORS UNWRITEABLE_DATA
7–39 CLEAR_ERRORS device-name UNKNOWN If a device failure causes the controller to label the device as unknown, the controller does not check the device again to see if it has been repaired or if the error condition has been corrected. You must enter this command so the controller can recognize the device after the cause of the error has been corrected. Use this command to force the controller to recognize a failed device, regardless of the controller’s prior evaluation of the device’s condition.
7–40 CLI Commands CLEAR_ERRORS unit-number LOST_DATA Clears lost data errors on a unit; all partitions on the unit’s container are affected. The controller reports a lost data error on the unit when you remove a write-back cache module or when the cache module contains unflushed data, possibly due to an interruption in the primary power source with no backup power present. The CLEAR_ERRORS LOST_DATA command clears the lost data error but does not recover the lost data.
7–41 See also CLEAR_ERRORS CLI CLEAR_ERRORS INVALID_CACHE CLEAR_ERRORS UNKNOWN CLEAR_ERRORS UNWRITEABLE_DATA Compaq HSG80 Array Controller ACS Version 8.
7–42 CLI Commands CLEAR_ERRORS unit-number UNWRITEABLE_DATA Clears an unwriteable data error on a unit. It affects all partitions on the same container. If a storageset or disk drive fails before its data has been written to it, the controller reports an unwriteable data error. The CLEAR_ERRORS UNWRITEABLE_DATA command removes the data from the cache and clears the unwriteable data error. CAUTION: This command causes data loss.
7–43 See also CLEAR_ERRORS CLI CLEAR_ERRORS INVALID_CACHE CLEAR_ERRORS LOST_DATA CLEAR_ERRORS UNKNOWN RETRY_ERRORS UNWRITEABLE_DATA Compaq HSG80 Array Controller ACS Version 8.
7–44 CLI Commands CONFIGURATION RESET Erases the entire configuration on “this controller,” restores the controller’s default configuration, and shuts down the controller. NOTE: If you plan to use this feature, SAVE_CONFIGURATION must be set when you initialize the container. See “INITIALIZE,” page 7–62. Specify the CONFIGURATION RESET command on “this controller” in nofailover mode only.
7–45 CONFIGURATION RESTORE Copies a controller’s configuration from the disk configuration file into the controller’s non-volatile memory. This command locates the most recent configuration file created on disk and restores it. This command causes a reboot and takes effect immediately. Use this command for a single controller configuration only. Do not use it for controllers in a dual-redundant configuration.
7–46 CLI Commands Syntax CONFIGURATION RESTORE See also CONFIGURATION RESET CONFIGURATION SAVE INITIALIZE
7–47 CONFIGURATION SAVE Forces a current copy of configuration information in a controller’s non-volatile memory into a configuration file on a disk. This allows the user to determine when a copy of the configuration is saved. Use this command to explicitly save a single controller’s configuration. The command takes effect immediately. In a dualredundant configuration, issue this command to both controllers.
7–48 CLI Commands CREATE_PARTITION Divides a non-transportable disk drive storageset into several, separately-addressable storage units. The command marks a specified percentage of a disk drive or storageset to be used as a separately addressable unit. You can divide any nontransportable disk or storageset into a maximum of eight partitions. Each partition can be separately presented to the host. Partitions are not supported in multiple bus failover mode.
7–49 To create a partition, specify a percentage of the container’s total capacity. The entire container is then divided into segments equal to the percentage specified. For example, if SIZE=20, the container is divided into five (1.0/0.2=5) equal segments. The resulting partition is slightly smaller than the size specified because metadata also occupies some of the partition’s allocated space.
7–50 CLI Commands ADD DISK DISK10000 1 0 0 ADD DISK DISK20100 2 1 0 ADD DISK DISK30200 3 2 0 ADD RAIDSET RAID9 DISK10000 DISK20100 DISK30200 INITIALIZE RAID9 CREATE_PARTITION RAID9 SIZE=25 CREATE_PARTITION RAID9 SIZE=25 CREATE_PARTITION RAID9 SIZE=25 CREATE_PARTITOIN RAID9 SIZE=LARGEST ADD UNIT D101 RAID9 PARTITION=1 ADD UNIT D102 RAID9 PARTITION=2 ADD UNIT D103 RAID9 PARTITION=3 ADD UNIT D104 RAID9 PARTITION=4 See also ADD UNIT DELETE unit-number DESTROY PARTITION SHOW
7–51 DELETE connections Deletes a host connection entry from the table of known connections. This command deletes a specified connection from the table of known connections maintained by the controller. The table of known host connections is maintained in the controllers NVRAM. Once a connection is added to the table, it stays there, even if the physical connection between host adapter and controller port is severed.
7–52 CLI Commands Examples Deletes the host connection Server1 from the table of known connections (unless the access path to Server1 is specifically enabled for one or more unit.
7–53 DELETE container-name Deletes a container belonging to the controller’s configuration. You cannot delete a container in use by a higher-level container. For example, you cannot delete a disk belonging to a RAIDset, or a RAIDset belonging to a unit; you must first delete the higher-level container or containers. NOTE: This command does not delete sparesets or failedsets. You cannot delete spareset and failedset containers. See the DELETE FAILEDSET and DELETE SPARESET commands for details.
7–54 CLI Commands Examples To delete a disk drive named DISK10000, type: DELETE DISK10000 To delete a stripeset named STRIPE1, enter: DELETE STRIPE1 To delete a RAIDset named RAID9, use: DELETE RAID9 See also DELETE FAILEDSET DELETE SPARESET UNMIRROR
7–55 DELETE FAILEDSET Removes a disk drive from the failedset. The failedset contains disk drives removed by the controller from RAIDsets and mirrorsets because they failed or were manually removed using the SET command. Enter the DELETE FAILEDSET command before physically removing failed disks from the storage shelf for testing, repair, or replacement. You should consider all disk drives in the failedset defective. Repair or replace disks found in the failedset.
7–56 CLI Commands DELETE SPARESET Removes a disk drive from the spareset. Syntax DELETE SPARESET disk-name Parameter disk-name Identifies the disk drive being deleted from the spareset. Remove only one disk at a time from a spareset.
7–57 DELETE unit-number Deletes a logical unit from the controller configuration. The host cannot address deleted units. If the unit’s write-back caching feature is enabled, the controller flushes the cached data to the unit’s devices before deleting the unit. Before using the DELETE unit-number command, clear any errors with the CLEAR_ERRORS UNWRITEABLE_DATA or CLEAR_ERRORS LOST_DATA commands. Syntax DELETE unit-number Parameter unit-number Identifies the unit number to be deleted.
7–58 CLI Commands DESTROY_PARTITION Marks the area reserved for a partition as available. The freed area is then consolidated with any adjacent free areas. CAUTION: Data contained on a partition is lost when you enter the DESTROY_PARTITION command. You cannot destroy a partition that has been assigned a unit number. First enter the DELETE unit-number command to delete the unit using the partition. After you partition a container, you must initialize it in order to destroy the partitions.
7–59 Example The following example shows how to delete the unit for partition 2 on RAIDset RAID9 and destroy the partition: DELETE D102 DESTROY_PARTITION RAID9 PARTITION=2 See also ADD DISK ADD STORAGESET ADD STRIPESET CREATE_PARTITION DELETE unit-number SHOW Compaq HSG80 Array Controller ACS Version 8.
7–60 CLI Commands DIRECTORY Lists the diagnostics and utilities available on “this controller.” Syntax DIRECTORY Example The example below shows how to display a directory listing: DIRECTORY HSUTIL V84G D CHVSN V84G D CLCP V84G D CLONE V84G D CONFIG V84G D DILX V84G D DIRECT V84G D DSTAT V84G D FRUTIL V84G D FMU V84G D VTDPY V84G D NOTE: CHVSN and DSTAT are not user utilities. They should be used by Compaq authorized service personnel only.
7–61 HELP Displays a brief explanation of how to use the question mark (?) to obtain help on any command or CLI function. You must precede the question mark with a space. Syntax HELP Example To display information regarding the HELP command, type: HELP Help may be requested by typing a question mark (?) at the CLI prompt.
7–62 CLI Commands INITIALIZE Initializes or destroys metadata on a container. During initialization, a small amount of disk space is reserved for controller metadata and is made inaccessible to the host. Disks made transportable do not contain controller metadata. Syntax INITIALIZE container-name CAUTION: The INITIALIZE command destroys all user data on the container unless you enter the NODESTROY switch. The NODESTROY switch is only valid on mirrorsets and striped mirrorsets.
7–63 Parameters container-name Specifies the container to initialize. This is the same name given to the disk or storageset when it was created using the ADD command (for example, ADD DISK, ADD STRIPESET, and so forth). Switches CAPACITY= CYLINDERS= HEADS= SECTORS_PER_TRACK= CAPACITY may be specified 1 to the maximum container size (in blocks); CYLINDERS may be specified 1 to16,777,215; HEADS may be specified 1 to 255; and SECTORS_PER_TRACK may be specified 1 to 255.
7–64 CLI Commands The default chunk size for storagesets with less than 9 members is 256 blocks, or 128 kilobytes (K). The default chunk size for storagesets with more than 9 members is 128 blocks, or 64K. The default values provide optimal storageset performance for a wide variety of applications. A chunk size less than 128 blocks (64K) is not recommended. IMPORTANT: Accept the default chunk size setting for most applications.
7–65 Specify SAVE_CONFIGURATION to store a copy of the controller configuration on the container being initialized. A new controller can receive information from a container containing configuration information saved with the SAVE_CONFIGURATION switch. If you specify SAVE_CONFIGURATION for a multi-device storageset, such as a stripeset, the complete controller configuration information is stored on each disk drive in the storageset.
7–66 CLI Commands SHOW DEVICES FULL Name Type Port Targ Lun Used by -----------------------------------------------------------------------------DISK10000 disk 1 0 0 S2 DEC Switches: RZ28M (C) DEC 1003 NOTRANSPORTABLE TRANSFER_RATE_REQUESTED = 20MHZ (synchronous 10.00 MHZ negotiated) Size: 4108970 blocks Configuration being backed up on this container DISK30300 disk DEC 3 RZ28M 3 0 S2 (C) DEC 1003 Switches: NOTRANSPORTABLE TRANSFER_RATE_REQUESTED = 20MHZ (synchronous 10.
7–67 LOCATE Indicates the physical location of configured units, storagesets, and devices by flashing the device fault LED on the front of the storage building block (SBB). The device fault LED flashes once per second until turned off with the LOCATE CANCEL command. The LOCATE command can also be used to test the LED itself. The device fault LED on a failed device stays on continuously. When located, the device fault LED on a good device flashes.
7–68 CLI Commands PTL (SCSI-location) Causes the device fault LED on the device at the given SCSI location to flash. See “Mapping the Physical Devices with Device PTL Addressing,” page 3–6 for an explanation of the PTL addressing naming format. Not all devices have a device fault LED. Therefore, they do not appear to respond to the LOCATE command. UNITS Causes the device fault LEDs of all devices used by the units to flash.
7–69 This example shows how to cause the device fault LEDs on all of the devices assigned to disk unit number D102 to flash: LOCATE D102 This example shows how to cause the device fault LEDs on all configured disk devices to flash: LOCATE DISKS This example shows how to turn off the flashing device fault LEDs on all devices: LOCATE CANCEL Compaq HSG80 Array Controller ACS Version 8.
7–70 CLI Commands MIRROR Creates a one-member mirrorset from a single disk. This command is used only on disks configured as units or members of a stripeset, then enter the ADD MIRRORSET command to create a mirrorset from disk drives not already members of higher level containers.
7–71 Switches COPY=FAST COPY=NORMAL (Default) Sets the speed at which the controller copies data to a new member from normal mirrorset members when data is being mirrored to the storageset’s disk drives. Specify COPY=FAST to allow the creation of mirrored data to take precedence over other controller operations. When you specify COPY=FAST, the controller uses more resources to create the mirrored data, and copying takes less time. However, overall controller performance is reduced during copying.
7–72 CLI Commands Specify NOPOLICY to prevent the controller from automatically replacing a failed disk device. This causes the mirrorset to operate in a reduced state until either POLICY=BEST_PERFORMANCE or POLICY=BEST_FIT is selected, or a member is manually replaced in the mirrorset. See “SET mirrorset-name,” page 7–110. Example This example shows how to create a one-member mirrorset from each member of a stripeset.
7–73 See also ADD MIRRORSET REDUCE SHOW MIRRORSETS UNMIRROR Compaq HSG80 Array Controller ACS Version 8.
7–74 CLI Commands POWEROFF Powers off all disk units in a cabinet and turns off the cabinet power. Syntax POWEROFF Switches BATTERY _ON BATTERY_OFF (Default) Instructs the external cache battery (ECB) charger to turn off or remain on. Specify BATTERY_ON to keep the ECB charger on after the POWEROFF command is issued. Specify BATTERY_OFF to turn off the ECB charger after the POWEROFF command is issued.
7–75 SECONDS=nn As soon as the POWEROFF command is entered, all disk units in the cabinet are set to write-through. When the time interval, as represented by nn seconds, has elapsed, an orderly rundown of all units is started. When all units in the cabinet are successfully rundown, the cabinet power is turned off. Table 7–3 shows what action will be taken depending on the switch settings and the results of the attempted flush: Table 7–3.
7–76 CLI Commands In dual-redundant mode, if both controllers can’t be shutdown, then both controllers and their batteries’ chargers remain on.
7–77 REDUCE Removes member disk drives from mirrorsets and decreases the nominal number of members in the mirrorsets. Unlike the SET mirrorset-name REMOVE=disk-name command, the controller does not put reduced members into the failedset. When using the REDUCE command to take a snapshot of a striped mirrorset, you must reduce all mirrorsets with one command. The CLONE utility does this automatically.
7–78 CLI Commands For each reduced mirrorset, there must be at least one remaining normal member after the reduction. If this is not true for all of the disk-names specified, the mirrorset is not reduced. Only normal members can be reduced. A normal member is a mirrorset member whose entire contents are the same as all other normal members within the mirrorset. NOTE: An error is displayed if you attempt to reduce a mirrorset so that there would not be any normal member remaining.
7–79 Name Storageset Uses Used by ------------------------------------------------------------------MIRR1 mirrorset DISK10100 STRIPE1 DISK20100 MIRR2 mirrorset DISK10200 STRIPE1 DISK20200 MIRR3 mirrorset DISK30300 STRIPE1 DISK40200 REDUCE DISK20100 DISK20500 DISK40200 SHOW MIRRORSETS Name Storageset Uses Used by -----------------------------------------------------------------------MIRR1 mirrorset DISK10100 STRIPE1 MIRR2 MIRR3 mirrorset mirrorset DISK10200 DISK30300 STRIPE1 STRIPE1
7–80 CLI Commands RENAME Renames a specified container or a specified host connection. Syntax RENAME old-name new-name Parameters old-name Specifies the existing name of the container or host connection. new-name Assigns the new name for the container or the host connection. See “Command Syntax,” page 7–5, for information regarding container naming rules.
7–81 RESTART controller Flushes all user data from the specified controller’s write-back cache and restarts the controller. Syntax RESTART controller Parameters controller The controller parameter indicates which controller is to be restarted. Specify OTHER_CONTROLLER or THIS_CONTROLLER. Switches IGNORE_ERRORS NOIGNORE_ERRORS (Default) Controls the reaction of the controller based on the status of write-back cache.
7–82 CLI Commands IMMEDIATE_SHUTDOWN NOIMMEDIATE_SHUTDOWN (Default) Instructs the controller whether to flush the write-back cache or not. CAUTION: The IMMEDIATE_SHUTDOWN switch instructs the controller to immediately shutdown, without regard to any data contained within write-back cache. See “Fault-Tolerance for Write-Back Caching,” page 2–11 for considerations when implementing write-back cache. Do not perform any hardware changes until the controller flushes the cache.
7–83 RETRY_ERRORS UNWRITEABLE_DATA Causes the controller to attempt to write previously unwriteable data from the writeback cache to the devices. If a container fails, preventing the data in write-back cache to be written to the container, an unwriteable data error is reported. If possible, correct the condition that caused the unwriteable data and try the write operation again. No data is lost if the retry fails.
7–84 CLI Commands RUN Runs a diagnostic or utility program on “this controller.” Diagnostic and utility programs only run on “this controller.” Syntax RUN program-name Parameter program-name The program-name parameter specifies the name of the diagnostic or utility program to be run. The following programs can currently be run: ■ CHVSN—This is not a user utility. This utility may be used by Compaq authorized service personnel only. ■ CLCP—A utility used to load updated software code or patches.
7–85 CAUTION: Run the DILX utility in the autoconfigure mode only at initial installations. When write operations are enabled, the DILX utility may overwrite existing data.
7–86 CLI Commands Example This example shows how to start the DILX diagnostic program: RUN DILX . . .
7–87 SELFTEST controller Flushes the data from the specified controller’s write-back cache (if present) and shuts down the controller. It then restarts the controller in self-test mode. Press the controller reset (//) button to take the controller out of self-test mode. Syntax SELFTEST controller Parameters controller The controller parameter indicates which controller is to perform the self-test program. Specify OTHER_CONTROLLER or THIS_CONTROLLER.
7–88 CLI Commands IMMEDIATE_SHUTDOWN NOIMMEDIATE_SHUTDOWN (Default) Instructs the controller whether to flush the write-back cache or not. CAUTION: The IMMEDIATE_SHUTDOWN switch instructs the controller to immediately shut down, without regard to any data contained within write-back cache. See “Fault-Tolerance for Write-Back Caching,” page 2–11, for considerations when implementing write-back cache. Do not perform any hardware changes until the controller flushes the cache.
7–89 SET connection-name Changes the operating characteristics of a host connection. The SET connection-name command changes the operating parameters of the specified host connection. A host connection is a specific instance of one host connected to one port of one controller through one host adapter. Syntax SET connection-name Parameters connection-name This is the name of the host connection.
7–90 CLI Commands ■ Unit number = the number assigned to the unit in the ADD UNIT command. This is the number by which the unit is known internally to the controllers. OPERATING_SYSTEM=OS_name Specifies the operating system of the host.
7–91 SET controller Changes parameters on the specified controller. NOTE: After doing a SET command to either controller, Compaq recommends using a SHOW THIS_CONTROLLER and a SHOW OTHER_CONTROLLER command to verify that the changes went into effect. Syntax SET controller Parameter controller Indicates which controller is to be set. Specify OTHER_CONTROLLER or THIS_CONTROLLER. Switches Table 7–4 lists the switches available with this command. Descriptions of the switches follow the table.
7–92 CLI Commands Table 7–4.
7–93 Table 7–4. SET controller Switches (Continued) Switch Values SCSI_VERSION SCSI-2 (default) SCSI-3 TERMINAL_PARITY NOTERMINAL_PARITY odd, even TERMINAL_SPEED 4800, 9600, 19200 TIME dd–mmm–yyy:hh:mm:ss Compaq HSG80 Array Controller ACS Version 8.
7–94 CLI Commands ALLOCATION_CLASS Allocation class is a unique identification number assigned to the controller pair under certain operating systems. The value for an allocation class is 0-4294967295. In DIGITAL Open VMS, this is a 2-byte number (do not specify a number over 65535); for DIGITAL UNIX, it is a 4-byte number. It is reported in response to the SCSI inquiry command and is the same for all units connected to one or both controllers.
7–95 CAUTION: Setting CACHE_UPS without having a UPS or similar backup system in place may result in data loss if power is interrupted. Specify NOCACHE_UPS to instruct the controller to perform regular cache battery checks and evaluate the condition of the cache batteries. Setting the CACHE_UPS switch for either controller sets the CACHE_UPS switch for both controllers. COMMAND_CONSOLE_LUN NOCOMMAND_CONSOLE_LUN (Default) Enables or disables the virtual LUN used with the StorageWorks Command Console.
7–96 CLI Commands MIRRORED_CACHE NOMIRRORED_CACHE (Default) Enables the mirrored-write-back-data cache feature on dual-redundant controllers. When changed, both controllers restart for the new switch setting to take effect. The following tasks are performed when the NOMIRRORED_CACHE switch is specified. Both controllers must be operational before this command is accepted. ■ Data in write-back cache is flushed when cache is configured in non-mirrored mode.
7–97 NODE_ID=nnnn-nnnn-nnnn-nnnn checksum Sets the subsystem worldwide name (node ID). If a situation occurs that requires you to reset the subsystem worldwide ID (node ID), use the name and checksum that appear on the sticker on the frame into which your controller is inserted. CAUTION: Each subsystem has its own unique worldwide name (node ID). If you attempt to set the subsystem worldwide name to a name other than the one that came with the subsystem, the data on the subsystem will not be accessible.
7–98 CLI Commands PORT_1_TOPOLOGY=FABRIC PORT_2_TOPOLOGY=FABRIC Used only in switch (FC-SW) topology. Specifies switch topology for a host port. PROMPT=“new prompt” Specifies a 1- to 16-character prompt displayed when the controller’s CLI prompts for input. Only printable ASCII characters and spaces are valid. The new prompt name must be enclosed within quotes. When changed, the new text entered for this switch takes effect immediately.
7–99 It also specifies how the command console LUN is handled. The command console LUN (CCL) presents to the GUI a virtual LUN through which it communicates with the controller. SCSI-2 specifies that the CCL is not fixed at a particular location, but floats depending on the configuration. SCSI-3 specifies that the CCL is fixed at LUN 0. The SCSI device-type returned to the host is array controller. Changes to this switch take place at the next controller restart.
7–100 CLI Commands Examples This example shows how to change the other controller’s CLI prompt: SET OTHER_CONTROLLER PROMPT=CONTROLLER “B” This example shows how to set the name of the controller pair at the initiator site: SET THIS_CONTROLLER REMOTE_COPY=NEW_YORK NOTE: The other controller in the pair automatically receives the same name.
7–101 SET device-name Changes the transportable characteristics and the maximum data transfer rate between the controller and the specified device. Syntax SET device-name Parameter device-name Specifies the name of the device to change. This can be a previously named device, disk, passthrough device, or container.
7–102 CLI Commands Storagesets cannot be made transportable. Specify NOTRANSPORTABLE for all disks used in RAIDsets, stripesets, mirrorsets, and sparesets. Transportable disks do not contain any metadata or restricted areas on the disk. Therefore, transportable disks forfeit the advantage metadata provides. Transportable disks can be moved to a nonStorageWorks environment with their data intact. If you specify the NOTRANSPORTABLE switch and there is no metadata on the unit, the unit must be initialized.
7–103 SET EMU Sets operating parameters for the environmental monitoring unit (EMU). Syntax SET EMU Switches The SENSOR and FANSPEED switches control both the master and slave EMU settings. The EMU within the primary cabinet (master) instructs the EMUs within the other cabinets to operate at the same SENSOR and FANSPEED settings to which the master EMU is set.
7–104 CLI Commands Table 7–5.
7–105 Select FANSPEED=AUTOMATIC to allow the EMU to control the fan speed for the fans in all connected cabinets. The EMU instructs the fans to operate at high speed when any of the temperature setpoints are exceeded or when one or more fans are not functioning.
7–106 CLI Commands SET FAILEDSET Changes the automatic replacement policy for the failedset. Syntax SET FAILEDSET Switches AUTOSPARE NOAUTOSPARE Specifies the policy to be used by the controller when a disk drive is physically replaced in the failedset. Specify AUTOSPARE to instruct the controller to automatically move devices physically replaced in the failedset into the spareset. Specify NOAUTOSPARE to instruct the controller to leave devices physically replaced in the failedset.
7–107 These steps use DISK10000 as an example. 1. Delete all containers to which the disk belongs. 2. Make the disk transportable. SET DISK10000 TRANSPORTABLE. 3. Initialize the disk. INIT DISK10000 4. Delete the disk. DELETE DISK10000 5. Move DISK10000 to the failedset’s vacant slot.
7–108 CLI Commands SET FAILOVER Configures both controllers to operate in a dual-redundant, transparent failover, configuration. This allows both controllers to access the storage devices, providing controller fault-tolerant data processing. If one of the two controllers fails, the devices and any cache attached to the failed controller become available to and accessible through the other controller.
7–109 Example This example shows how to set the controllers in a dual-redundant configuration and copy the configuration information from “this controller” to “other controller:” SET FAILOVER COPY=THIS_CONTROLLER See also SET MULTIBUS_FAILOVER SET NOFAILOVER SET NOMULTIBUS_FAILOVER Compaq HSG80 Array Controller ACS Version 8.
7–110 CLI Commands SET mirrorset-name Changes the characteristics of a mirrorset, including the addition and removal of members. Syntax SET mirrorset-name Parameter mirrorset-name Specifies the name of the mirrorset to modify. This is the same name given to the mirrorset when it was created with the ADD MIRRORSET command.
7–111 NOTE: No other switches can be set when you specify the MEMBERSHIP switch. If you increase the number of members and specify a replacement policy with the POLICY= switch, the controller automatically adds disk drives from the spareset to the mirrorset until the new number of members is reached, or there are no more suitable disk drives in the spareset.
7–112 CLI Commands NOTE: Normalizing members exist only when you first create a mirrorset or when you clear lost data on a mirrored unit. The controller recognizes the member as normal, and all other original mirrorset members as “normalizing.” New data that is written to the mirrorset is written to all members. The controller copies the data existing before the mirrorset was created on the normal member to the normalizing members.
7–113 Specify POLICY=BEST_PERFORMANCE to choose a replacement disk drive from the spareset with the best performance. The controller attempts to select a disk on a different port than existing mirrorset members. If there is more than one disk drive in the spareset matching the best performance criteria, the controller selects the disk drive that equals or exceeds the base member size of the mirrorset. Specify NOPOLICY to prevent the controller from automatically replacing a failed disk device.
7–114 CLI Commands This example shows how to remove member DISK30000 from mirrorset MIRR1 created above. If the mirrorset has a replacement policy and an acceptable disk drive is in the spareset, the controller automatically adds the spare disk drive to the mirrorset. SET MIRR1 REMOVE=DISK30000 This example shows how to add disk DISK30200 to the mirrorset MIRR1: SET MIRR1 REPLACE=DISK30200 A copy operation begins immediately on DISK30200.
7–115 SET MULTIBUS_FAILOVER Places “this controller” and the “other controller” into a dual-redundant (failover) configuration within a multiple-bus environment. This allows both controllers to access the storage devices and provide greater throughput. If one controller fails, the devices and cache attached to the failed controller become available to and accessible through the remaining controller. Both controllers must be configured for nofailover before you enter the SET MULTIBUS_FAILOVER command.
7–116 CLI Commands Due to the amount of information being passed from one controller to the other, this command may take up to one minute to complete.
7–117 SET NOFAILOVER Reconfigures both controllers to operate in a non-dual-redundant (non-failover) configuration. Immediately after entering this command, remove one controller from the shelf because the sharing of devices is not supported by nonredundant controllers. NOTE: SET NOFAILOVER and SET NOMULTIBUS_FAILOVER have the same effect. Either command exits from transparent or multiple bus failover mode. It is recommended that both controllers be present when this command is carried out.
7–118 CLI Commands Specify NODESTROY_UNFLUSHABLE_DATA to leave the unwritten data intact in the failed controller’s write-back cache. When the failed controller is replaced and placed into service, the write-back cache data is flushed to the appropriate devices. Specify DESTROY_UNFLUSHABLE_DATA to reconfigure the operational controller before replacing the failed controller. The unwritten data of the failed controller may reference devices not present in the new configuration.
7–119 SET NOMULTIBUS_FAILOVER Reconfigures both controllers to operate in a non-dual-redundant (non-failover) configuration. Immediately after entering this command, remove one controller from the shelf because the sharing of devices is not supported by nonredundant controllers. NOTE: SET NOFAILOVER and SET NOMULTIBUS_FAILOVER have the same effect. Either command exits from transparent or multiple bus failover mode. It is recommended that both controllers be present when this command is carried out.
7–120 CLI Commands Under some circumstances, the data in a failed controller’s write-back cache may not fail over to the operating controller’s write-back cache. For example, cache data will not failover if the operating controller has a failed cache battery because of the risk of data loss if the power is interrupted. Specify NODESTROY_UNFLUSHABLE_DATA to leave the unwritten data intact in the failed controller’s write-back cache.
7–121 SET RAIDset-name Changes the characteristics of a RAIDset. Syntax SET RAIDset-name Parameters RAIDset-name Specifies the name of the RAIDset to modify. This is the name used with the ADD UNIT command to identify the RAIDset as a host-addressable unit. Switches POLICY=BEST_FIT POLICY=BEST_PERFORMANCE (Default) NOPOLICY Specifies the replacement policy to use when a member within the RAIDset fails.
7–122 CLI Commands RECONSTRUCT=FAST RECONSTRUCT=NORMAL (Default) Sets the speed at which the controller reconstructs the data on the new RAIDset member replacing a failed member. Specify NORMAL to balance other controller operations against the reconstruct operation. The controller uses relatively few resources to perform the reconstruct, and there is little impact on performance. Specify FAST when the reconstruct operation must take precedence over other controller operations.
7–123 REPLACE=disk-name Instructs the controller to add a disk member to an existing RAIDset if the following conditions are met: ■ The replacement policy is set to NOPOLICY. ■ The disk member is not in any configuration, including a spareset. An error is displayed and the command is rejected if the RAIDset is not in a reduced state, if a replacement policy is already specified, or if the disk specified is already being used by a configuration (including a spareset).
7–124 CLI Commands SET remote-copy-set-name IMPORTANT: This command works in a remote copy environment only using HSG80 ACS V8.4P code, where local (initiator) and remote (target) sites coexist. Changes the characteristics of a remote copy set. Syntax SET remote-copy-set-name Parameters remote-copy-set-name Specifies the name of the remote copy set to modify. This is the name used with the ADD REMOTE_COPY_SETS command to identify the remote copy set as a hostaddressable unit.
7–125 Select NORMAL to unlock the remote copy center and to change the error mode so that the unit in the remote copy set can be written to. INITIATOR=initiator unit-name Sets the name of the controller pair as the initiator. REMOVE=target path-name Removes a controller pair from an existing remote copy set. The controller pair specified by target path-name is removed from the remote copy set specified by remote-copy-set-name.
7–126 CLI Commands SET unit-number Changes the characteristics of a unit. Syntax SET unit-number Parameter unit-number Specifies the logical unit number to modify. The unit-number is the name given to the unit when it was created using the ADD UNIT command. Switches Table 7–6 lists all switches for the SET unit-number command and shows which switches can be used with each type of device and storageset. Descriptions of the switches follow the table.
7–127 Container Type ENABLE_ACCESS_PATH DISABLE_ACCESS_PATH MAXIMUM_CACHED_ TRANSFER_SIZE IDENTIFIER NOIDENTIFIER PREFERRED_PATH NOPREFERRED_PATH READ_CACHE NOREAD_CACHE READAHEAD_CACHE NOREADAHEAD_CACHE WRITE_PROTECT NOWRITE_PROTECT WRITEBACK_CACHE NOWRITEBACK_CACHE RUN NORUN Switch Table 7–6 SET unit-number Switches for Existing Containers RAIDset ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔ Stripeset ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔ Mirrorset ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔ NoTransportable Disk ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔ Tr
7–128 CLI Commands If you enable another host ID, previously enabled hosts are not disabled. The new ID is added. If you wish to enable only certain IDs, disable all access paths (DISABLE_ACCESS_PATH=ALL), then enable the desired IDs. The system will display the following message: Warning 1000: Access IDs in addition to the one(s) specified are still enabled. If you wish to enable ONLY the id(s) listed, disable all access paths (DISABLE_ACCESS_PATH=ALL), then enable the ones previously listed.
7–129 PREFERRED_PATH=OTHER_CONTROLLER PREFERRED_PATH=THIS_CONTROLLER NOPREFERRED_PATH (Default) Used only in Windows NT operating systems. May be set only when dual-redundant controllers are operating in a multiple bus failover configuration. In a multiple bus failover configuration, the host determines which controller the units are accessed through. The host’s unit-to-controller settings always take precedence over the preferred path assigned to units with this switch.
7–130 CLI Commands Read caching improves performance in almost all situations, so it is generally recommended to leave it enabled. However, under certain types of conditions, such as when performing a backup, read-caching may not be necessary since only a small amount of data is cached. In such instances, it may be beneficial to disable read cache and remove the processing overhead associated with caching. READAHEAD_CACHE (Default) NOREADAHEAD_CACHE Enables the controller to keep track of read I/Os.
7–131 Specify WRITE_PROTECT to prevent host write operations to the unit. However, the controller may still write to a write-protected RAIDset to satisfy a reconstruct pass or to reconstruct a newly replaced member. However, metadata, reconstruct, and copy writes are still allowed to RAIDsets and mirrorsets. Specify NOWRITE_PROTECT to allow the host to write data to the unit. This allows the controller to overwrite existing data. NOWRITE_PROTECT is the default for transportable disks.
7–132 CLI Commands Example This example shows how to enable write protect and turn off the read cache on unit D102: SET D102 WRITE_PROTECT NOREAD_CACHE See also SHOW UNITS SHOW unit-number
7–133 SHOW Displays information about controllers, storagesets, devices, partitions, and units. The SHOW command may not display some information for devices accessed through the companion controller in a dual-redundant configuration. When information regarding a device or parameter does not appear, enter the same SHOW command from a terminal on the other controller.
7–134 CLI Commands THIS_CONTROLLER OTHER_CONTROLLER device-name Specifies the name of a particular device to be displayed. For example, SHOW DISK20100 displays information about the device named DISK20100. device-type Specifies the type of devices you want to be displayed. Valid choices are: ■ DEVICES—Shows all devices attached to the controller ■ DISKS—Shows all disks attached to the controller EMU Displays information regarding the status of the environmental monitoring unit (EMU).
7–135 unit-number Specifies the name of a particular unit to be displayed. For example, SHOW D102 displays information about the unit named D102. UNITS Displays information for all units configured to the controller. In addition to the unit name you defined for the unit, the information includes the unique 128-bit subsystem unit ID. This ID consists of the controller node ID plus a 64-bit unit ID generated by the subsystem.
7–136 CLI Commands Examples This example shows how to display a listing of disks: SHOW DISKS Name Type Port Targ Lun Used by ----------------------------------------------------------DISK20300 disk 1 0 0 DISK10100 disk 1 1 0 D100 D101 This example shows a full listing of devices attached to the controller: SHOW DEVICES FULL Name Type Port Targ Lun Used by ------------------------------------------------------------------DISK10300 disk 1 3 0 R0 Switches: NOTRANSPORTABLE TRANSFER_R
7–137 This example shows how to display a complete listing of the mirrorset named MIRR1: SHOW MIRR1 Name Storageset Uses Used by -----------------------------------------------------------------------------MIRR1 mirrorset DISK50300 S0 DISK60300 Switches: POLICY (for replacement) = BEST_PERFORMANCE COPY (priority) = NORMAL READ_SOURCE = LEAST_BUSY MEMBERSHIP = 2, 2 members present State: NORMAL DISK60300 (member DISK50300 (member 0) is NORMAL 1) is NORMAL Size: 17769177 blocks Compaq HSG80 Array Cont
7–138 CLI Commands This example shows the full information for a controller: SHOW THIS_CONTROLLER FULL Controller: HSG80 (C) DEC CX00001905 Software R024F-0, Hardware 0000 NODE_ID = XXXX-XXXX-XXXX-XXXX ALLOCATION_CLASS = 767 SCSI_VERSION = SCSI-2 Configured for dual-redundancy with ZG83301871 In dual-redundant configuration Device Port SCSI address 7 Time: NOT SET Command Console LUN is lun 0 (IDENTIFIER = 8888) Host PORT_1: Reported PORT_ID = XXXX-XXXX-XXXX-XXXX PORT_1_PROFILE = PLDA PORT_1_TOPOLOGY =
7–139 CACHE_FLUSH_TIMER = DEFAULT (10 seconds) Mirrored Cache: Not enabled Battery: FULLY CHARGED Expires: WARNING: UNKNOWN EXPIRATION DATE! WARNING: AN UNKNOWN NUMBER OF DEEP DISCHARGES HAVE OCCURRED! NOCCACHE_UPS Extended information: Terminal speed 19200 baud, eight bit, no parity, 1 stop bit Operation control: 00000000 Security state code: 6985 Configuration backup enabled on 1 devices This example shows how to display the current settings for the EMU: SHOW EMU EMU CABINET SETTINGS SENSOR_1_SETPOINT
7–140 CLI Commands SHUTDOWN controller Flushes all user data from the specified controller’s write-back cache (if present) and shuts down the controller. The controller does not automatically restart. All units accessed through the failed controller failover to the surviving controller. Syntax SHUTDOWN controller Parameter controller Indicates which controller is to shut down. Specify OTHER_CONTROLLER or THIS_CONTROLLER.
7–141 IMMEDIATE_SHUTDOWN NOIMMEDIATE_SHUTDOWN (Default) Instructs the controller when to shutdown. CAUTION: The IMMEDIATE_SHUTDOWN switch causes the controller to keep unflushed data in the write-back cache until it restarts and is able to write the data to devices. Do not perform any hardware changes until the controller flushes the cache.
7–142 CLI Commands UNMIRROR Converts a one-member mirrorset back to a non-mirrored disk drive and deletes its mirrorset from the list of known mirrorsets. This command can be used on mirrorsets already members of higher-level containers (stripesets or units). The UNMIRROR command is not valid for disk drives having a capacity greater than the capacity of the existing mirrorset.
A–1 Appendix A System Profiles This appendix contains device and storageset profiles you can use to create your system profiles. It also contains an enclosure template you can use to help keep track of the location of devices and storagesets in your shelves. Compaq HSG80 Array Controller ACS Version 8.
A–2 System Profiles Device Profile Type of Device ___ Platter disk drive ___ Optical disk drive ___ Tape Drive ___ CD-ROM Device Name Unit Number Device Switches Transportability ___ No (default) ___ Yes Initialize Switches Chunk size ___ Automatic (default) ___ 64 blocks ___ 128 blocks ___ 256 blocks ___ Other: Unit Switches Read Cache ___ Yes (default) ___ No Availability ___ Run (default) ___ NoRun Save Configuration ___ No (default) ___ Yes Write Cache ___ Yes (default) ___ No Write Protect
A–3 Storageset Profile Type of storageset _____ Mirrorset _____ RAIDset _____ Stripeset _____ Striped Mirrorset Storageset Name Disk Drives Unit Number Partitions Unit # % Unit # % Unit # Unit # % % RAIDset Switches Reconstruction Policy ___ ___ Normal (default) Fast Best performance (default) Best fit None No (default) Yes, missing: Automatic (default) 64 blocks 128 blocks 256 blocks Other: Yes (default) No ___ ___ Yes (default) No No (default) Yes Write-Back Cache Yes (default) No N
A–4 System Profiles Enclosure Template Power Supply Power Supply Power Supply Power Supply Power Supply Power Supply Power Supply Power Supply
B–1 Appendix B Controller Specifications This chapter contains physical, electrical, and environmental specifications for the HSG80 array controller. Compaq HSG80 Array Controller ACS Version 8.
B–2 Controller Specifications Physical and Electrical Specifications for the Controller Table B–1 lists the physical and electrical specifications for the controller and cache modules. Table B–1 Controller Specifications Hardware Length Width Power HSG80 Array Controller module 12 inches 9.5 inches 23.27 W Write-back Cache, 512 MB 5.5 inches 9.5 inches 2.48 W 8.72 W Current at +5 V Current at +12 V 6.
B–3 Environmental Specifications The HSG80 array controller is intended for installation in a Class A computer room environment. The optimum operating environmental specifications are listed in Table B–2; the maximum operating environmental specifications are listed in Table B–3; and the maximum nonoperating environmental specifications are listed in Table B–4. These are the same as for other Compaq storage devices.
B–4 Controller Specifications Table B–3 Maximum Operating Environmental Specifications Condition Specification Temperature +10° to +40°C (+50° to +104°F) Derate 1.8°C for each 1000 m (1.
GL–1 Glossary This glossary defines terms pertaining to the HSG80 Fibre Channel array controller. It is not a comprehensive glossary of computer terms. 8B/10B A type of byte encoding and decoding to reduce errors in data transmission patented by the IBM Corporation. This process of encoding and decoding data for transmission has been adopted by ANSI. adapter A device that converts the protocol and hardware interface of one bus type into another without changing the function of the bus.
GL–2 Glossary array controller software Abbreviated ACS. Software contained on a removable ROM program card that provides the operating system for the array controller. asynchronous Pertaining to events that are scheduled as the result of a signal asking for the event; pertaining to that which is without any specified time relation. See also synchronous. autospare A controller feature that automatically replaces a failed disk drive.
GL–3 built-in self-test A diagnostic test performed by the array controller software on the controller’s policy processor. byte A binary character string made up of 8 bits operated on as a unit. cache memory A portion of memory used to accelerate read and write operations. CCITT Acronym for Consultive Committee International Telephone and Telegraph. An international association that sets worldwide communication standards, recently renamed International Telecommunications Union (ITU).
GL–4 Glossary command line interpreter The configuration interface to operate the controller software. configuration file A file that contains a representation of a storage subsystem’s configuration. container 1) Any entity that is capable of storing data, whether it is a physical device or a group of physical devices. (2) A virtual, internal controller structure representing either a single disk or a group of disk drives linked as a storageset.
GL–5 data striping The process of segmenting logically sequential data, such as a single file, so that segments can be written to multiple physical devices (usually disk drives) in a round-robin fashion. This technique is useful if the processor is capable of reading or writing data faster than a single disk can supply or accept the data. While data is being transferred from the first disk, the second disk can locate the next segment. device See node and peripheral device.
GL–6 Glossary dual-simplex A communications protocol that allows simultaneous transmission in both directions in a link, usually with no flow control. DUART Dual universal asynchronous receiver and transmitter. An integrated circuit containing two serial, asynchronous transceiver circuits. ECB External cache battery. The unit that supplies backup power to the cache module in the event the primary power source fails or is interrupted. ECC Error checking and correction. EDC Error detection code.
GL–7 failedset A group of failed mirrorset or RAIDset devices automatically created by the controller. failover The process that takes place when one controller in a dual-redundant configuration assumes the workload of a failed companion controller. Failover continues until the failed controller is repaired or replaced. FC–AL The Fibre Channel Arbitrated Loop standard.
GL–8 Glossary FCC Class B This certification label appears on electronic devices that can be used in either a home or a commercial environment within the United States. FCP The mapping of SCSI-3 operations to Fibre Channel. FDDI Fiber Distributed Data Interface. An ANSI standard for 100 megabaud transmission over fiber optic cable. FD SCSI The fast, narrow, differential SCSI bus with an 8-bit data transfer rate of 10 MB/s. See also FWD SCSI and SCSI. fiber A fiber or optical strand.
GL–9 full duplex (adj) Pertaining to a communications method in which data can be transmitted and received at the same time. FWD SCSI A fast, wide, differential SCSI bus with a maximum 16-bit data transfer rate of 20 MB/s. See also SCSI and FD SCSI. GBIC Gigabit Interface Converter. The devices that are inserted into the ports of the Fibre Channel switch and that hold the Fibre Channel cables. GLM Gigabit link module giga A prefix indicating a billion (10 9) units, as in gigabaud or gigabyte.
GL–10 Glossary hot disks A disk containing multiple hot spots. Hot disks occur when the workload is poorly distributed across storage devices which prevents optimum subsystem performance. See also hot spots. hot spots A portion of a disk drive frequently accessed by the host. Because the data being accessed is concentrated in one area, rather than spread across an array of disks providing parallel access, I/O performance is significantly reduced. See also hot disks.
GL–11 I/O interface See interface. I/O module A 16-bit SBB shelf device that integrates the SBB shelf with either an 8-bit single ended, 16-bit single-ended, or 16-bit differential SCSI bus. I/O operation The process of requesting a transfer of data from a peripheral device to memory (or visa versa), the actual transfer of the data, and the processing and overlaying activity to make both of those happen. IPI Intelligent Peripheral Interface.
GL–12 Glossary local terminal A terminal plugged into the EIA-423 maintenance port located on the front bezel of the controller. See also maintenance terminal. logical bus A single-ended bus connected to a differential bus by a SCSI bus signal converter. logical unit A physical or virtual device addressable through a target ID number. LUNs use their target’s bus connection to communicate on the SCSI bus.
GL–13 maintenance terminal An EIA-423-compatible terminal used with the controller. This terminal is used to identify the controller, enable host paths, enter configuration information, and check the controller’s status. The maintenance terminal is not required for normal operations. See also local terminal. member A container that is a storage element in a RAID array. metadata The data written to a disk for the purposes of controller administration.
GL–14 Glossary not permitted to arbitrate or originate frames. An L_Port in non-participating mode may or may not have an AL_PA. See also participating mode. nominal membership The desired number of mirrorset members when the mirrorset is fully populated with active devices. If a member is removed from a mirrorset, the actual number of members may fall below the “nominal” membership. node In data communications, the point at which one or more functional units connect transmission lines.
GL–15 other controller The controller in a dual-redundant pair that is connected to the controller serving your current CLI session. See also this controller. outbound fiber One fiber in a link that carries information away from a port. parallel data transmission A data communication technique in which more than one code element (for example, bit) of each byte is sent or received simultaneously. parity A method of checking if binary numbers or characters are correct by counting the ONE bits.
GL–16 Glossary parity RAID See RAIDset. partition A logical division of a container, represented to the host as a logical unit. peripheral device Any unit, distinct from the CPU and physical memory, that can provide the system with input or accept any output from it. Terminals, printers, tape drives, and disks are peripheral devices. point-to-point connection A network configuration in which a connection is established between two, and only two, terminal installations.
GL–17 protocol The conventions or rules for the format and timing of messages sent and received. PTL Port-Target-LUN. The controller’s method of locating a device on the controller’s device bus. PVA module Power Verification and Addressing module. quiesce The act of rendering bus activity inactive or dormant. For example, “quiesce the SCSI bus operations during a device warm-swap.” RAID Redundant Array of Independent Disks.
GL–18 Glossary RAID level 3/5 A DIGITAL-developed RAID storageset that stripes data and parity across three or more members in a disk array. A RAIDset combines the best characteristics of RAID level 3 and RAID level 5. A RAIDset is the best choice for most applications with small to medium I/O requests, unless the application is write intensive. A RAIDset is sometimes called parity RAID. Raid level 3/5 storagesets are sometimes referred to as RAIDsets. RAIDset See RAID level 3/5.
GL–19 request rate The rate at which requests are arriving at a servicing entity. RFI Radio frequency interference. The disturbance of a signal by an unwanted radio signal or frequency. replacement policy The policy specified by a switch with the SET FAILEDSET command indicating whether a failed disk from a mirrorset or RAIDset is to be automatically replaced with a disk from the spareset. The two switch choices are AUTOSPARE and NOAUTOSPARE. SBB StorageWorks building block.
GL–20 Glossary SCSI device ID number A bit-significant representation of the SCSI address referring to one of the signal lines, numbered 0 through 7 for an 8-bit bus, or 0 through 15 for a 16-bit bus. See also target ID number. SCSI ID number The representation of the SCSI address that refers to one of the signal lines numbered 0 through 15. SCSI-P cable A 68-conductor (34 twisted-pair) cable generally used for differential bus connections.
GL–21 storage array subsystem See storage subsystem. storageset (1) A group of devices configured with RAID techniques to operate as a single container. (2) Any collection of containers, such as stripesets, mirrorsets, striped mirrorsets, and RAIDsets. storage subsystem The controllers, storage devices, shelves, cables, and power supplies used to form a mass storage subsystem.
GL–22 Glossary striping The technique used to divide data into segments, also called chunks. The segments are striped, or distributed, across members of the stripeset. This technique helps to distribute hot spots across the array of physical devices to prevent hot spots and hot disks. Each stripeset member receives an equal share of the I/O request load, improving performance.
GL–23 Ultra SCSI A Fast-20 SCSI bus. See also Wide Ultra SCSI. unit A container made accessible to a host. A unit may be created from a single disk drive or tape drive. A unit may also be created from a more complex container such as a RAIDset. The controller supports a maximum of eight units on each target. See also target and target ID number. unwritten cached data Sometimes called unflushed data. See dirty data. UPS Uninterruptible power supply.
GL–24 Glossary write-back caching A cache management method used to decrease the subsystem’s response time to write requests by allowing the controller to declare the write operation “complete” as soon as the data reaches its cache memory. The controller performs the slower operation of writing the data to the disk drives at a later time. write-through caching A cache management method used to decrease the subsystem’s response time to a read.
I–1 Index A AC input module part number, 1–4 Access paths general description, 3–14 Active ports multiple-bus failover, 2–7 transparent failover mode, 2–5 Active vs. standby ports multiple-bus failover mode, 2–7 transparent failover mode, 2–5 Adapters.
I–2 Index to spareset using CLI, 6–15 Addresses providing with the PVA module, 5–29 Addressing controller and host, 3–1 PTL convention, 3–6 ALLOCATION_CLASS SET controller, 7–94 Array Controller.
I–3 Cache policies fault-tolerance for write-back caching, 2–11 Cache, setting flush timer, 7–94 CACHE_FLUSH_TIMER SET controller, 7–94 CACHE_UPS SET controller, 7–94 Caching techniques, 1–6 general description, 2–9 read caching, 2–9 read-ahead caching, 2–9 write-back caching, 2–10 write-through caching, 2–10 CAPACITY CREATE_PARTITION, 7–49 INITIALIZE, 7–63 Caution, defined, xx Change volume serial number utility.
I–4 Index CLEAR_ERRORS controller INVALID_CACHE, 7–36 CLEAR_ERRORS device-name UNKNOWN, 7–39 CLEAR_ERRORS unit-number LOST_DATA, 7–40 CLEAR_ERRORS unit-number UNWRITEABLE_DATA, 7–42 CONFIGURATION RESET, 7–44 CONFIGURATION RESTORE, 7–45 CONFIGURATION SAVE, 7–47 CREATE_PARTITION, 7–48 customizing the prompt, 7–98 DELETE connections, 7–51 DELETE container-name, 7–53 DELETE FAILEDSET, 7–55 DELETE SPARESET, 7–56 DELETE unit-number, 7–57 DESTROY_PARTITION, 7–58 DIRECTORY, 7–60 editing keys, 7–4 getting help, 7–
I–5 CONFIG, running, 7–84 Configuration modifying controller configurations, 7–2 resetting, 7–44 restoring, 7–45, 7–64 saving, 7–47 CONFIGURATION RESET, 7–44 CONFIGURATION RESTORE, 7–45 Configuration rules devices, 5–3 LUN capacity, 5–3 mirrorsets, 5–3 partitions per storageset, 5–3 RAID-5 and RAID-1 storagesets, 5–3 RAID-5 storagesets, 5–3 RAID-5, RAID-1, and RAID-0 storagesets, 5–3 requirements, 5–3 striped mirrorsets, 5–3 stripesets, 5–3 See also Summary of controller features CONFIGURATION SAVE, 7–47 C
I–6 Index controller and cache module location, 1–10 displaying information, 7–133 dual-redundant controller configuration, 5–14, 5–23 dual-redundant controller configurations with mirrored cache, 2–18 ECB diagnostics, 1–18 failover modes supported, 1–5 fault LEDs, 1–11 fibre channel optical cabling illustration of parts, 1–8 part numbers of parts used in configuring, 1–8 parts used in configuring, 1–8 general description, 1–4 host ports, 1–10 local connection, 5–33 location, 1–10 maintenance port, 1–10 m
I–7 SECTORS_PER_TRACK, 7–49 SIZE, 7–48 CREATE_PARTITION container-name SIZE=percent, 7–48 Creating disks, 7–10 mirrorsets, 7–13 partitions, 6–13 RAIDsets, 7–17 single-disk units, 7–34 sparesets, 7–23 storageset and device profiles, 4–5 stripesets, 7–24 units, 7–27 CYLINDERS CREATE_PARTITION, 7–49 INITIALIZE, 7–63 D Data backing up with the Clone utility, 1–14 duplicating with the Clone utility, 1–14 Data center cabinet ECB Y cable, 1–16 to 1–17 Data transfer rate, 4–45 Data-retention-policy CLEAR_ERRORS c
I–8 Index Device profile, A–2 Device protocol, 1–6 Device statistics utility. See DSTAT utility Device switches, 4–36, 4–41 changing switches, 4–36 device transfer rate, 4–42 enabling switches, 4–36 NOTRANSPORTABLE, 4–41 TRANSFER_RATE_REQUESTED, 4–42 transportability, 4–41 TRANSPORTABLE, 4–41 Device targets.
I–9 partitions supported, 5–3 removing from a mirrorset, 7–77 removing from sparesets using CLI, 6–16 removing from the failedset, 7–55 removing from the spareset, 7–56 renaming, 7–80 renaming the volume serial number with the CHVSN utility, 1–14 setting device data transfer rate, 7–11 showing, 7–133 to 7–134 transfer rate, 7–11 Disk inline exerciser general description, 1–13 DISKS SHOW device-type, 7–134 Display.
I–10 Index Erasing metadata, 4–49 Error messages clearing from CLI, 7–35 clearing unwriteable data errors, 7–42 Examples adding disk drives to a spareset, 6–16 cloning a storage unit, 4–23 configuring a mirrorset, 6–8 configuring a RAIDset, 6–9 configuring a single-disk unit, 6–12 configuring a striped mirrorset, 6–11 configuring a stripeset, 6–7 deleting storagesets, 6–18 partitioning a storageset, 6–14 removing disk drives from a spareset, 6–17 Exercisers DILX, 1–13 See also Utilities and exercisers Ext
I–11 assigning access paths, 3–14 bus dynamics, 3–2 connecting to the storage array, 3–6 mapping containers and storagesets to the host with logical units, 3–10 number of heterogeneous hosts supported, 1–5 protocol supported, 1–5 Host bus general description, 3–3 Host bus interconnect, 1–6 Host path preferring in multiple-bus failover mode, 2–8 Host port number supported, 1–5 Host ports location, 1–10 LUNs in multiple-bus failover mode, 2–8 LUNs in transparent failover mode, 2–6 Host-assisted failover.
I–12 Index INITIALIZE container-name, 7–62 Initialize switches, 4–43 chunk size, 4–43 CHUNKSIZE, 4–43 DESTROY, 4–49 destroy/nodestroy, 4–49 NODESTROY, 4–49 save configuration, 4–46 SAVE_CONFIGURATION, 4–46 Logical units.
I–13 Membership RAIDset switches, 4–38 Metadata erasing, 6–17 retaining, 6–17 MIRROR, 7–70 COPY, 7–71 POLICY, 7–71 MIRROR disk-name mirrorset-name, 7–70 Mirrored write-back cache enabling, 2–18 MIRRORED_CACHE SET controller, 7–96 Mirrorset unit number in OpenVMS, 6–8 Mirrorset switches, 4–39 COPY, 4–39 POLICY, 4–39 READ_SOURCE, 4–40 Mirrorsets actual number of members, 7–77 adding to configuration, 7–13 changing switches, 6–19 choosing a replacement member, 7–18 configuring using CLI, 6–7 converting back t
I–14 Index NODESTROY, 4–49 INITIALIZE, 7–64 NODESTROY_UNFLUSHABLE_DAT A SET NOMULTIBUS_FAILOVER, 7–119 NODESTROY_UNFLUSHED_DATA CLEAR_ERRORS controller INVALID_CACHE, 7–36 NOIDENTIFIER SET controller, 7–95 SET unit-number, 7–128 NOIGNORE_ERRORS RESTART controller, 7–81 SELFTEST controller, 7–87 SHUTDOWN controller, 7–140 NOIMMEDIATE_SHUTDOWN RESTART controller, 7–82 SELFTEST controller, 7–88 SHUTDOWN controller, 7–141 NOMIRRORED_CACHE SET controller, 7–96 Nonvolatile memory fault-tolerance for write-back
I–15 for RAIDsets, 4–37 for storage units, 6–2 initialize, 4–43 Other controller explained, 7–2 OVERRIDE_BAD_FLUSH POWEROFF, 7–74 Overwriting data, 4–49 P Part numbers 256-MB cache upgrade, 1–16 64-MB cache upgrade, 1–16 AC input module, 1–4 BA370 rack-mountable enclosure, 1–4 cooling fan, 1–4 dual-battery ECB, 1–4 ECB, 1–4 ECB Y cable, 1–17 BA370 enclosure, 1–16 data center cabinet, 1–16 EMU, 1–4 fibre channel optical cabling parts used in configuring the controller, 1–8 GLM optical, 1–8 I/O module, 1–4
I–16 Index PORT_1_TOPOLOGY SET controller, 7–97 PORT_2_ALPA SET controller, 7–97 PORT_2_TOPOLOGY SET controller, 7–97 Ports See also Device ports, Host ports Power source enabling write-back caching, 2–11 Power supply part number, 1–4 Power, verification, and addressing module.
I–17 adding while missing a member, 7–19 changing characteristics, 7–121 changing switches, 6–19 choosing chunk size, 4–43 configuring using CLI, 6–8 deleting, 7–53 description, 4–2, 4–16 displaying information, 7–133 initializing, 7–62 maximum chunk size, 4–46 maximum membership, 4–18 planning, 4–18 removing a member, 7–122 renaming, 7–80 replacing a member, 7–123 showing, 7–133 specifying replacement policy, 7–121 switches, 4–37 Read caching enabled for all storage units, 2–9 general description, 2–9 Rea
I–18 Index NOIGNORE_ERRORS, 7–81 NOIMMEDIATE_SHUTDOWN, 7–82 Restarting subsystem, 5–37 Restarting the subsystem, 5–37 Restoring configuration, 7–45 RETRY_ERRORS unit-number UNWRITEABLE_DATA, 7–83 Revision history, xxiii RUN, 7–84 ADD UNIT, 7–32 CHVSN, 7–84 CLCP, 7–84 CLONE, 7–84 CONFIG, 7–84 DILX, 7–84 DIRECT, 7–85 DSTAT, 7–85 FMU, 7–85 FRUTIL, 7–85 HSUTIL, 7–85 SET unit-number, 7–130 VTDPY, 7–85 RUN program name, 7–84 S Save configuration, 4–46 SAVE_CONFIGURATION, 4–46 INITIALIZE, 7–64 Saving configurat
I–19 NOTERMINAL_PARITY, 7–99 PORT_1_ALPA, 7–97 PORT_1_TOPOLOGY, 7–97 PORT_2_ALPA, 7–97 PORT_2_TOPOLOGY, 7–97 PROMPT, 7–98 SCSI_VERSION, 7–98 TERMINAL_PARITY, 7–99 TERMINAL_SPEED, 7–99 TIME, 7–99 SET device-name, 7–101 NOTRANSPORTABLE, 7–101 TRANSFER_RATE_REQUESTED, 7–101 TRANSPORTABLE, 7–101 SET EMU, 7–103 FANSPEED, 7–104 SENSOR_N_SETPOINT, 7–103 SET FAILEDSET, 7–106 AUTOSPARE, 7–106 NOAUTOSPARE, 7–106 SET FAILOVER, 7–108 SET FAILOVER COPY=controller, 7–108 SET mirrorset-name, 7–110 COPY, 7–110 MEMBERSHIP,
I–20 Index controller behavior at restart, 7–82 controller behavior at shutdown, 7–141 controller behavior selftest, 7–88 controller cache flush timer, 7–94 controller cache UPS policy, 7–94 controller configuration handling, 7–64 controller error handling at self-test, 7–82, 7–87 controller error handling at shutdown, 7–140 data retention policy, 7–36 device data transfer rate, 7–11, 7–101 failedset autospare feature, 7–106 fan speed, 7–104 full display, 7–135 mirrorset copy data, 7–71 mirrorset copy spe
I–21 NOIMMEDIATE_SHUTDOWN, 7–141 Shutting down subsystem, 5–36 Shutting down the subsystem, 5–36 Single controller configuring, 5–6 connecting, 5–8 Single-battery ECB part number, 1–4 Single-controller configuration ECB, 1–17 Single-disk unit unit number in OpenVMS, 6–12 Single-disk units backing up, 4–21 configuring with CLI, 6–12 displaying switches, 6–19 SIZE CREATE_PARTITION, 7–48 Sparesets adding disk drives using CLI, 6–15 adding to configuration, 7–23 AUTOSPARE, 6–17 removing a disk drive, 7–56 remo
I–22 Index striped mirrorsets, 4–2 stripesets, 4–2, 4–9 See also Configuration rules StorageWorks array controller, 7–2 Striped mirrorset unit number in OpenVMS, 6–10 Striped mirrorsets configuring using CLI, 6–10 description, 4–2, 4–19 maximum number of physical devices, 1–7, 5–3 planning, 4–20 Stripeset unit number in OpenVMS, 6–6 Stripesets adding to configuration, 7–24 configuring using CLI, 6–6 deleting, 7–53 description, 4–2, 4–9 displaying information, 7–133 initializing, 7–62 maximum number of mem
I–23 Terminology of the controller, 2–2 Testing controllers, 7–87 This controller explained, 7–2 removing from dual-redundant controller configuration, 7–117, 7–119 starting diagnostic or utility programs, 7–84 This Controller and Other Controller general description, 2–2 This controller, defined, xix TIME SET controller, 7–99 Tip, defined, xx Topology supported, 1–5 Transfer rate how chunk size affects, 4–43 setting device, 7–11, 7–101 switch, 4–42 TRANSFER_RATE_REQUESTED, 4–42 ADD DISK, 7–11 SET device-n
I–24 Index VTDPY, 1–13 See also Config utility See also HSUTIL Troubleshooting and maintaining the controller utilities and exercisers, 1–12 Typographical conventions, xviii U Unit how a unit is online in multiple-bus failover, 2–7 Unit offsets general description, 3–13 Unit switches changing, 6–20 overview, 6–2 Units adding to configuration, 7–27 assigned and accessed in multiple-bus failover mode, 2–7 assigned and accessed in transparent failover mode, 2–6 assigning access paths, 3–14 assigning unit nu
I–25 Clone utility, 1–14 CONFIG utility, 1–13 DILX, 1–13 DSTAT, 1–14 FMU, 1–12 FRUTIL, 1–14 HSUTIL, 1–13 VTDPY, 1–13 V Write-back caching enabled for all disk units, 2–10 fault-tolerance, 2–11 general description, 2–10 setting the flush timer, 7–94 WRITEBACK_CACHE ADD UNIT, 7–33 SET unit-number, 7–131 Write-through caching general description, 2–10 Virtual terminal display, 1–13 Volume serial number generating a new one with the CHVSN utility, 1–14 renaming with the CHVSN utility, 1–14 VTDPY general des
