Hitachi Cache Manager User’s Guide Hitachi Universal Storage Platform V Hitachi TagmaStore® Universal Storage Platform Hitachi TagmaStore® Network Storage Controller Hitachi Lightning 9900™ V Series Hitachi Lightning 9900™ FAS T FIND LINK S Document Organization Product Version Getting Help Contents HIT1642-96001
ii Hitachi Cache Manager User’s Guide
Contents Preface ................................................................................................... v Intended Audience .............................................................................................. vi Product Version................................................................................................... vi Document Revision Level ..................................................................................... vi Source Document(s) for this Revision .........
Performing Cache Manager Operations .................................................. 4-1 HRUDCRX Commands ....................................................................................... 4-2 Using Bind to Place or Release Data From the Cache .................................... 4-2 Using COUNTS to Display Cache Volume or Dataset Statistics ........................ 4-5 Using STATUS to Display the Cache Status...................................................
Preface This document describes and provides instructions for using the Hitachi Cache Manager software for the Hitachi RAID storage systems. Cache Manager enables the user to perform Cache Manager Residency operations from the mainframe host. Please read this document carefully to understand how to use this product, and maintain a copy for reference purposes.
Intended Audience This document is intended for system administrators, Hitachi Data Systems representatives, and Authorized Service Providers who are involved in installing, configuring, and operating the Hitachi RAID storage systems. This document assumes the following: • The user has a background in data processing and understands RAID storage systems and their basic functions.
Document Organization The following table provides an overview of the contents and organization of this document. Click the chapter title in the left column to go to that chapter. The first page of each chapter provides links to the sections in that chapter. Chapter Description Overview of Hitachi Cache Manager Provides an overview of the features and functions of Hitachi Cache Manager. About Cache Manager Operations Provides the Cache Manager specifications and describes Priority mode and Bind mode.
Document Conventions The terms “Universal Storage Platform V” and “USP V” refer to all models of the Hitachi Universal Storage Platform V, unless otherwise noted. This document uses the following typographic conventions: Typographic Convention Description Bold Indicates text on a window, other than the window title, including menus, menu options, buttons, fields, and labels. Example: Click OK. Italic Indicates a variable, which is a placeholder for actual text provided by the user or system.
Convention for Storage Capacity Values Physical storage capacity values (e.g., disk drive capacity) are calculated based on the following values: 1 1 1 1 1 KB = 1,000 bytes MB = 1,0002 bytes GB = 1,0003 bytes TB = 1,0004 bytes PB = 1,0005 bytes Logical storage capacity values (e.g.
Comments Please send us your comments on this document. Make sure to include the document title, number, and revision. Please refer to specific section(s) and paragraph(s) whenever possible. • E-mail: doc.comments@hds.com • Fax: 858-695-1186 • Mail: Technical Writing, M/S 35-10 Hitachi Data Systems 10277 Scripps Ranch Blvd. San Diego, CA 92131 Thank you! (All comments become the property of Hitachi Data Systems Corporation.
1 Overview of Hitachi Cache Manager This chapter provides an overview of the features and functions of the Hitachi Cache Manager software.
Hitachi Cache Manager Cache Manager enables users to perform Cache Residency Manager operations on Hitachi RAID storage systems from the mainframe host console. Cache Residency Manager enables users to store specific data in cache memory to make the data available to the host at front-end access speeds. The disk drives are not accessed. For more information about Cache Residency Manager, please see the Hitachi Cache Residency Manager User’s Guide for the storage system.
• The maximum number of extents is limited: Universal Storage Platform V – No more than 4096 extents per logical volume – No more than 16384 extents for the storage system Universal Storage Platform/Network Storage Controller – No more than 4096 extents per logical volume – No more than 4096 extents for the storage system Lightning 9900V and Lightning 9900 – No more than 1024 extents per logical volume – No more than 1024 extents for the storage system • If a user needs an entire mainframe
User data can be removed from cache, but with the cache extent information maintained in the disk controller in the following cases: • By using the SETCACHE command (provided by the IDCAMS utilities of MVS/ESA, OS/390®, z/OS®) with the REINITIALIZE parameter, • If there is an offline microcode change, • If there is a cache memory failure (partial or module down) during maintenance.
2 About Cache Manager Operations This chapter provides the Cache Manager specifications and describes Priority mode and Bind mode.
Methods to Specify Cache Data Table 2-1 Specifications of Cache Manager Supported Features Lightning 9900V and 9900 Universal Storage Platform V and TagmaStore USP/NSC Disk Systems Supported 9900V Series: DKC610I Note: All systems require an additional memory module and the Cache Residency Manager (FlashAccess) software.
DKC Software/ Cache Residency Manager Specifications Maximum number of resident extents Logical Volume: ≤ 1024 extents Storage System: ≤ 1024 extents Maximum number of extents for dataset statistics information Logical Volume: ≤ 64 extents Storage System: ≤ 64 extents USP V: Logical Volume: ≤ 4096 extents Storage System: ≤ 16384 extents USP: Logical Volume: ≤ 4096 extents Storage System: ≤ 4096 extents USP V: Logical Volume: ≤ 64 extents LCU: ≤ 64 extents Storage System: ≤ 32640 extents USP: Logical Vo
Overview of Priority Mode and Bind Mode Priority Mode In Priority Mode (normal mode), the total capacity of cache required is: Standard cache + cache required for Cache Manager + Additional cache The main advantage of priority mode is that read data is transferred at host data transfer speed. In priority mode the cache extents are used to hold read data for specific extents on volumes. Write data is write duplexed in normal cache and de-staged to disk using standard algorithms.
Bind Mode In Bind Mode, the total capacity of cache required is: Standard cache + cache required for Cache Manager In bind mode the cache extents are used to hold read and write data for specific extent(s) on volume(s). Any data written to the bind area is not destaged to the disk. To ensure data integrity, write data must be duplexed in the cache area, which consumes a significant amount of the cache.
2-6 About Cache Manager Operations Hitachi Cache Manager User’s Guide
3 Preparing for Cache Manager Operations This chapter provides information on installing and launching Cache Manager, including the HRUDCRX and HRUDCRT programs.
Installing Cache Manager Before you install Cache Manager, you must install Cache Residency on each storage system, make sure that the system option is set to 81 (to support the bound heads counter) and verify that the microcode version supports Cache Manager. The Hitachi Data Systems representative will install the additional cache memory module and set the amount of cache memory. The volume serial number for installation is DCR001.
Note: DCRT is an alias of HRUDCRT. You can invoke Cache Manager in TSO by specifying either HRUDCRT or DCRT as the program name. One issue reported with using DCRT, however, is a missing file for online help. If your installation CD is missing the DCRT help file, avoid using the DCRT alias, or contact HDS for the DCRT help file. HRUDCRX must be registered as authorized libraries, as follows: 1. Register in SYS1.PARMLIB(IKJTSOxx). 2. Add HRUDCRT and the alias DCRT as AUTHCMD NAMES. 3.
Launching Cache Manager Launching HRUDCRX 1. Call the HRUDCRX program. The line count must be from 20 to 99, and the default value is 60. See Figure 3-3 for an example. 2. Use the DD statement to define the volume type. The operand of the statement must specify the unit and volume, and the DISP operand must specify OLD. Note: If more than one DD statement is defined, only the first one will be processed. If you need to include more than one DD statement, you must divide the job step. 3.
7. If you are executing a DSMODE command with either an ENABLE or a DISABLE parameter, in order to log any eccentric areas (those that are not part of a dataset), you must specify DCRLOG DD statement. Note: This command will be ignored if another command or parameter is specified.
The format for comments is as follows: 3-6 • The first column must have a forward slash and an asterisk sign (/*). • Comment statements can be written after a parameter. /*COMMENT*/ Example: STATUS DEV(5837) DKU • Comment statements can be written between parameters. /*COMMENT*/ Example: BIND EA DEV(5837) VTOC • Comment statements can be written after commands. /*COMMENT*/ DEV(5837) DKU Example: STATUS • Comment statements can be written before commands.
Launching HRUDCRT To launch HRUDCRT: 1. The terminal monitor program will issue an ATTACH macro that generates a task and activates the command processor. Register 1 will have a pointer to the command processor parameter list (CPPL). 2. The executed command processor will analyze the operands as follows: – Call the CALLTSSR macro. – Make a parameter control list (PCL), including operand description and abbreviated value. – Hand the analyzed parameter list (PPL) to the operand analyze routine.
Estimating Required Memory To calculate the required memory for HRUDCRX, start with 170 kB, which is the basic requirement. Add to that the variable requirement, which is calculated as follows: (A + B + 2,112 +C * (5D + 1,024)) / 1,024 (kB) A is either 160 (if you are using JOB stream input) or 5 * the block length of the SYSIN dataset. B is either 242 (if you are using SYSOUT output) or 5 * the block length of the SYSPRINT dataset. C is either 1 (if you are specifying a RAIDPARM dataset) or 0.
4 Performing Cache Manager Operations This chapter explains the commands and parameters for executing Cache Manager operations. HRUDCRX Commands Using HRUDCRT Commands to Display Information at The Console WARNING: Do not perform the ShadowImage quick restore operation or Volume Migration operations on Cache Manager volumes. These operations swap the internal locations of the source and target volumes.
HRUDCRX Commands HRUDCRX supports the following commands: • BIND either places data into cache or releases it from cache. • COUNTS displays the cache volume or dataset statistics. This command can also reset the counters if necessary. • STATUS displays the cache status information. • DSMODE either defines the tracks for dataset statistics or releases that definition. • HRUDCRX commands can also be combined using JCL.
Figure 4-2 is a sample JCL of a BIND/ENABLE command. (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) //CSUT01 JOB MSGLEVEL=… //CACHE EXEC PGM=HRUDCRX,REGION=1024K //SYSPRINT DD SYSOUT=A //VOL1 DD UNIT=SYSDA,VOL=SER=CSVOL1,DISP=OLD //VOL2 DD UNIT=SYSDA,VOL=SER=CSVOL2,DISP=OLD //SYSIN DD * BIND ENABLE DDNAME(VOL1) VTOC BIND ENABLE DDNAME(VOL2) DATASET(‘USER.
Figure 4-3 is a sample JCL of a BIND/DISABLE command. (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) //CSUT02 JOB MSGLEVEL=(1,1) //CACHE EXEC PGM=HRUDCRX,REGION=1024K //SYSPRINT DD SYSOUT=A //VOL1 DD UNIT=SYSDA,VOL=SER=CSVOL1,DISP=OLD //VOL2 DD UNIT=SYSDA,VOL=SER=CSVOL2,DISP=OLD //SYSIN DD * BIND DISABLE DDNAME(VOL1) VTOC BIND DISABLE DDNAME(VOL2) DATASET(‘USER.
Using COUNTS to Display Cache Volume or Dataset Statistics The COUNTS command allows you to get statistics on cache memory usage and status information. Note: Be sure to reset the counter before accessing cache statistic information. WARNING: If other applications are using the cache statistics information, resetting the counter will interfere with those operations. The COUNTS command will display either volume or dataset statistics.
Figure 4-5 is a sample JCL of a COUNTS command for volume statistics. Cache Manager resets the volume statistics of the current volume, then edits and prints out the volume statistics for all volumes in the storage system.
Figure 4-6 is an example of volume statistics.
Figure 4-7 is a sample JCL of a COUNTS command for dataset statistics. Cache Manager edits and prints out dataset statistics information for the specified datasets, then edits and prints out dataset statistics information for all volumes with the same storage system ID (SSID).
Using STATUS to Display the Cache Status The STATUS command displays volume or dataset status information, including the size of the cache memory, the residence area of the cache extents, and the extent of the dataset statistics information. See the glossary for status definitions. Figure 4-9 illustrates STATUS command parameters. • You can specify target data by DDNAME, (using volume, unit and disp number), VOLUME (volume serial number) or DEVICE (device number).
Figure 4-11 is an example of a cache extent STATUS report.
Figure 4-13 is an illustration of a JCL for volume STATUS. (1) (2) (3) (4) (5) (6) (7) (8) (9) //CSUT13 JOB MSGLEVEL=(1,1) //CACHE EXEC PGM=HRUDCRX,REGION=1024K //SYSPRINT DD SYSOUT=A //RAIDPARM DD DSN=input dataset name, VOL=SER=volume serial number, // UNIT=SYSDA,DISP=SHR //SYSIN DD * STATUS VOLUME(CSVOL2) DKU STATUS VOLUME(CSVOL3) DKC EXTENT /* // (1) (2) (3) (4) (5) (6) Start of job. Execute HRUDCRX. Spool the message dataset. Define a dataset to store the device number.
Figure 4-14 illustrates a sample JCL for a dataset STATUS command. Cache Manager stores, edits and prints the extent information (dataset statistics) for all volumes, including a specified volume with the same storage system ID (SSID).
Figure 4-15 is an example of a dataset STATUS report.
Using DSMODE to Define Tracks for Dataset Statistics Figure 4-17 illustrates DSMODE command parameters. The DSMODE command allows you to define a cache extent for dataset statistics information and then release it. Once the dataset is defined, Cache Manager can display I/O counts for each extent (refer to Figure 4-8). • ENABLE will set dataset statistics information for the specified area, defined either by VTOC or DATASET.
Figure 4-18 illustrates a DSMODE command. Cache Manager specifies dataset statistics information by the tracks of the specified extents. (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) //CSUT24 JOB MSGLEVEL=(1,1) //CACHE EXEC PGM=HRUDCRX,REGION=1024K //SYSPRINT DD SYSOUT=A //VOL1 DD UNIT=SYSDA,VOL=SER=CSVOL1,DISP=OLD //DCRLOG DD DSN=input dataset name, VOL=SER=volume serial number, // UNIT=SYSDA,DISP=SHR //SYSIN DD * DSMODE ENABLE DDNAME(VOL1) VTOC DSMODE ENABLE VOLUME(CSVOL2) DATASET(‘USER.
Figure 4-19 illustrates a DSMODE command. Cache Manager releases dataset statistics information for the specified extent(s) or volume(s).
Using JCL Run Commands Figure 4-20 illustrates a JCL of combined volume commands. Cache Manager edits and prints out the statistics information counter, resets the counter, then resides the VTOC of the specified volumes. Then Cache Manager edits and prints out the extent information.
Figure 4-21 illustrates a JCL of combined dataset commands. Cache Manager edits and prints out the dataset statistics information counter, then sets the dataset statistics information for the VTOC of the specified volumes. Then Cache Manager edits and prints out the cache status and dataset statistics.
Using HRUDCRT Commands to Display Information at the Console HRUDCRT supports the following commands: • COUNTS displays cache volume or dataset statistics. This command can also reset the counters if necessary. • STATUS displays cache status information. • HELP displays definitions of terms that appear on screen. Using COUNTS to Display Cache Volume or Dataset Statistics Figure 4-22 illustrates HRUDCRT COUNTS command parameters. • VOLUME specifies a process target by volume serial number.
Figure 4-23 is an illustration of a HRUDCRT volume COUNTS command.
Figure 4-24 is an illustration of a HRUDCRT dataset COUNTS command.
Using STATUS to Display the Cache Status Figure 4-25 illustrates HRUDCRT STATUS command parameters. See the glossary for further definition of the types of cache status. • VOLUME allows you to specify a process target by volume serial number. • DKU|DKC|CLPR allows you to specify a target range. DKU specifies the target range by a particular volume parameter, and DKC specifies the target range as all volumes in the same storage system ID. You must also define the RAIDPARM DD statement.
Figure 4-27 is an illustration of a HRUDCRT residence STATUS command.
Figure 4-29 is an illustration of a HRUDCRT dataset STATUS command. READY HRUDCRT STATUS VOL(RAID31) CLPR JSE800I-I STATUS ACCEPT JSE001I-I END OF HRUDCRX, RETURN CODE IS 00 ***** CACHE SUBSYSTEM STATUS REPORT (01-06-03) ***** DISK CONTROLLER INFORMATION CU-TYPE SSID DEVICE ADDRESS -------------------------3990-3 X'0080' X'01' STATUS CACHE NVS PINNED DATA ------------------------------------ENABLED ENABLED NONE CLPR NO.
Displaying Help for HRUDCRT Commands Figure 4-30 is an example of the HRUDCRT HELP file. )F FUNCTION THE HRUDCRT COMMAND DISPLAYS STATUS OF CACHE SUBSYSTEMS SPECIFIED. )X SYNTAX HRUDCRT STATUS VOLUME(‘SERIAL’) DKC/DKU EXTENT/NOEXTENT OR HRUDCRT COUNTS VOLUME(‘SERIAL’) DKC/DKU RESET/NORESET REPORT/NOREPORT REQUIRED DEFAULT ALIAS - STATUS/COUNTS VOLUME - DKC NOEXTENT NORESET REPORT - DCRT )O OPERANDS))STATUS - DISPLAYS STATUS OF CACHE SUBSYSTEMS SPECIFIED.
4-26 Performing Cache Manager Operations Hitachi Cache Manager User’s Guide
5 Return Codes, Output Messages, and Error Codes This chapter lists the return and error codes as well as the output messages.
Return Codes HRUDCRX has the following return codes: • 0 indicates that all commands were successfully executed. • 4 indicates that a slight error occurred that doesn’t affect the execution of the commands. The process continues, but an attention message will display. • 8 indicates that an error occurred during the command execution, and the process ended abnormally. The listed commands (if any were specified) will be executed.
Output Messages The output message for HRUDCRX is stored in a sequential dataset. The message includes input control statements, which are printed out with each command followed by the result of the command execution. If you want to print out cache statistics information, specify COUNTS and then REPORT. If you want to print out status information, specify STATUS and then REPORT. The output for HRUDCRT is displayed on a TSO terminal.
Error Codes Table 5-1 Error Number Error Codes Error Message Recommended Action JSE001I-I END OF HRUDCRX, RETURN CODE IS return code HRUDCRX has been executed. The return code has the maximum value of return codes during the execution of HRUDCRX. JSE002I-I [VTOC, dataset name, FROM cchh TO cchh] ON [volume serial number or disk drive unit number] WAS BOUND TO CACHE HRUDCRX specified cache residence by VTOC, DATASET, or CCHH.
Error Number JSE101I-E Error Message Recommended Action OPERAND ‘operand’ IS INCONSISTENT The set of operands is inconsistent. This operand can not be used with the other specified operand. (S) Terminate the command process, but the next process continues. (Return code is 8.) (P) Correct the error and execute the command again. JSE102I-E JSE103I-E ‘operand’ SPECIFIED BUT REQUIRED OPERAND MISSING The specified operand requires another operand.
Error Number JSE204I-S Error Message Recommended Action NOT ENOUGH SPACE IN CACHE This utility tried to reside the tracks specified by the BIND command, but there is not enough cache memory. (S) HRUDCRX terminates abnormally. (Return code is 12.) (P) After releasing unnecessary resident area, execute the command again. JSE205I-S CACHE ACCESS IS DISABLED This utility tried to execute a BIND or DSMODE command, but the cache storage system is disabled because of an unspecified failure.
Error Number JSE231I-S Error Message Recommended Action ALREADY BOUND TO CACHE BY [BIND or PRIO] MODE The command to reside data using PRIO mode was made to the BIND mode area, or vice versa. (S) HRUDCRX terminates abnormally. (Return code is 12.) (P) Check the mode of the already resident area and correct the parameter, then execute the command again. JSE232I-S EXTENT OVER ON BOUND TO CACHE The number of resident extents exceeds the limitation.
Error Number Error Message JSE2401-E DATASET ‘dataset name’ DOES NOT EXIST IN CATALOG JSE2411-W Recommended Action DATASET TYPE CAN NOT CACHE PROCESS ‘dataset name’ The specified dataset did not exist in specified catalog. (S) HRUDCRX terminates abnormally. (Return code is ‘8’.) (P) Check the specified dataset name and catalog name and then execute the command again. The specified dataset is not supported under CMG environment.
Error Number JSE307I-E JSE500I-U Error Message Recommended Action I/O ERROR job name, step name, disk drive unit number, device type, dd name, order, error type, BBCCHHR, QSAM I/O error has been detected during the output operation of the COUNTS or STATUS command. See “SYNADAF macro” explanation for more detail. INVALID RAIDPARM PARAMETER The specified contents of the dataset in RAIDPARM DD statement has a coding rule error. (S) Terminate the command process, but the next process continues.
Error Number JSE511I-W Error Message Recommended Action ‘disk drive unit number’ IS NOT FOUND The disk drive unit number specified by the dataset in RAIDPARM DD statement does not exist. (S) Terminate the process for this disk drive unit, but the process continues for the following disk drive units. (Return code is 4.) JSE800I-I command name ACCEPT The specified STATUS or COUNTS command has been accepted normally.
6 Troubleshooting This chapter provides troubleshooting information for Cache Manager and instructions for calling technical support.
Troubleshooting The Hitachi RAID storage systems provide continuous data availability and are not expected to fail in any way that would interrupt access to user data. For troubleshooting information, please refer to the User and Reference Guide for the storage system.
Acronyms and Abbreviations CC CPPL CU cylinder cylinder command processor parameter list control unit DASD DCR DFW direct-access storage device Dynamic Cache Residency (another name for Cache Residency Manager) DASD fast write ESA Enterprise Systems Architecture GB gigabyte(s) HH head head IPL initial program load JCL job control language kB kilobyte(s) LCU LVI logical control unit logical volume image (also called device emulation) max MB MVS maximum megabyte(s) Multiple Virtual Storage
Acronyms-2 Acronyms and Abbreviations Hitachi Cache Manager User’s Guide
Glossary This glossary defines some of the special terms used in this document. bound heads Read and write counts from cache. cache extents Cache Residency Manager cache status: cache off error disabled disconnected enabled cache error initializing Areas used for Cache Residency or Cache Manager. Other names for this product are: FlashAccess, Dynamic Cache Residency (DCR). Cache failed to stop. Cache is offline. Cache is disabled for maintenance. Status is normal.
Extent Relation (continued): EQUAL The number of dataset extents managed by the VTOC is the same as the number of datasets from which statistics are derived. EQUALThe number of dataset extents managed by the VTOC is less than the number of datasets from which statistics are derived. LONG The number of dataset extents from which statistics are derived includes extents that do not have datasets. SHORT The number of dataset extents from which statistics are derived includes datasets that do not have extents.
Hitachi Cache Manager User’s Guide
HIT1642-96001
