This section describes the preparation and installation of Adabas on IBM z/VM systems.
To use the information in this section, you need to be familiar with z/VM and CP concepts such as directories and virtual machines. See the Glossary for an explanation of terms referenced.
Note:
See section VM/GCS for
VM/GCS-related information to support some Software AG products such as Entire
Net-Work. Adabas itself no longer supports
VM/GCS.
The following list provides an overview of the Adabas installation procedure on a z/VM system.
| Step | Description | Additional Information |
|---|---|---|
| 1 | Define virtual machines for the Adabas environment. | Two virtual machines are required for operating
Adabas. At least two more virtual machines are recommended:
For specific information on the CP directory requirements for each of these virtual machines, refer to the individual sections describing the virtual machines. |
| 2 | Define minidisk space for the Adabas library disk. | If a DBA virtual machine has been defined, the library minidisk should be defined by an MDISK statement in the CP directory for the DBA virtual machine; otherwise, it should be defined in the CP directory of the Adabas nucleus machine. For detailed information about the Adabas library minidisk see the section Adabas Library Requirements. |
| 3 | Define minidisk space for the Adabas database. | Some effort should be made to place these minidisks on different channels, both virtual and real. For detailed information about the Adabas database minidisks, see the section Disk Space Requirements for the Database. |
| 4 | Load the Adabas release tape and subsequently install a database. | This can be done using either the online procedure (INSTPROD) as described in the following section and in section Using INSTPROD (Online Install), or manually; however, the online procedure is recommended. |
Note:
If you plan to use the Adabas Recovery Aid (ADARAI) utility, specific
changes must first be made to the PROFILE and
ADFdbid EXECs. Neither INSTADA nor the ADAMAINT EXEC
are currently capable of making the changes needed for running ADARAI. For more
information, see the section
VM/ESA or
z/VM in the ADARAI utility documentation.
The following steps are required to load the release tape and install Adabas using the online procedure:
to load the release tape and install Adabas using the online
procedure:
LOGON to and then DISCONNECT from the ID table manager virtual machine.
LOGON to the Adabas nucleus virtual machine. If a database administrator (DBA) virtual machine has been defined, DISCONNECT from Adabas; otherwise, ATTACH a tape drive to Adabas as drive 181 and ignore step 3.
LOGON to the DBA virtual machine and ATTACH a tape drive as drive 181.
ACCESS the Adabas library disk with a filemode other than “A”. Filemode “C” is recommended.
Issue command TAPE FSF 1.
Issue command TAPE LOAD * * filemode, where filemode is the mode specified in step 4.
Issue INSTPROD to start the online installation procedure (after the files are unloaded from the installation tape, INSTPROD starts INSTADA automatically).
Enter the installation parameters on the online screens that are presented. Section Using INSTPROD (Online Install) contains a detailed description of the INSTADA procedure.
Customize the ADAINPL EXEC file for your environment as preparation for loading the Adabas Online System (if used) into a Natural system file.
Stop the Adabas nucleus by entering ADAEND from the console of the Adabas virtual machine, or from a secondary console.
Execute the ADASAV EXEC, specifying the filename of the z/VM file that contains the ADASAV control statements (i.e., SAVE) and the DBID as parameters.
This completes the installation using the online procedure. If the installation is interrupted, INSTADA can be restarted; all the parameter values already entered are stored.
Existing installations may use the INSTADA procedure to convert an Adabas 6 database to Adabas 7. In this case, the CONVERT parameter should be set to YES.
The EXECs supplied on the Adabas release tape are based on the control statement files and EXECs created by INSTADA. Software AG strongly recommends that new databases be defined and formatted using only the INSTADA EXEC from step 7.
Additional databases can be defined using the INSTADA EXEC. The parameters for each database are entered in the online procedure panels during the initial installation. Section Using INSTPROD (Online Install) describes how INSTADA can be used, and shows the INSTADA screens and menus.
The ADAMAINT EXEC can be used to maintain the Adabas environment, including updating all files created by INSTADA. This EXEC presents the same panels as INSTADA, but does not format the database. This prevents accidental destruction of a database while performing maintenance. The ADAMAINT EXEC allows you to either increase a dataset or add a new extent by making changes to the online screen.
The INCREASE EXEC can be used to run the INCREASE or ADD functions of the Adabas ADADBS utility as well as formatting with ADAFRM. The ADAMAINT EXEC must be run before using the INCREASE EXEC.
The following steps are required to load the release tape and install Adabas without the online procedure. Many of the steps correspond to steps performed automatically by the online procedure, but not necessarily in the same order.
to load the release tape and install Adabas without using the online
procedure:
LOGON to and DISCONNECT from the ID table manager virtual machine.
LOGON to the Adabas nucleus virtual machine. If a database administrator (DBA) virtual machine has been defined, DISCONNECT from Adabas; otherwise, ATTACH a tape drive to Adabas as unit 181 and continue with step 4.
LOGON to the DBA virtual machine and ATTACH a tape drive as drive 181.
ACCESS the Adabas library disk with a filemode other than “A”. Filemode “C” is recommended.
Issue command TAPE FSF 4.
Issue command TAPE LOAD * * fm.
Create a PROFILE EXEC for the DBA virtual machine.
Refer to the section z/VM Operating System Environment for an example of the CP directory entries for the DBA virtual machine.
Include the following functions in the PROFILE EXEC:
MULTI write LINK commands to the Adabas database minidisks (ASSO, DATA, WORK, TEMP, SORT).
ACCESS for the library disk as a read only extension of the A-disk (ACC cuu fm/A).
Issue EXEC SETTXTLB.
Define ADARUN, DATADEF, DISPDD, and RELDD as nucleus extensions. This can be done by executing the NUCXTNTS EXEC with no parameters.
Execute the PROFILE EXEC.
Create a zap file for ADAITM if you are going to modify the default values for the target-ID, VMID, or restart parameters:
target ID of ADAITM at location X’20’. (default X’FFFF’).
VMID of the DBA virtual machine at location X’28’. (default: 8 blanks)
automatic restart parameter at location X’30’ (default: C’Y’).
Create a zap file for ADALDI if you are going to modify the default values for the VMID of the ID table manager.
The VMID of the ID table manager is at location X’34’ (default C’DBIDSERV’).
Apply the zaps if necessary.
Issue the following commands, where nnnnn is the database ID and fm is the filemode. Modify the copied EXECs, which always have a filename containing the five-digit DBID (with leading zeros, if needed), as follows:
COPY ADADEFS EXEC fm ADFnnnnn EXEC fm COPY ADAFRM CONTROL fm FRMnnnnn CONTROL fm COPY ADADEF CONTROL fm DEFnnnnn CONTROL fm COPY RUNDB CONTROL fm RDBnnnnn CONTROL fm COPY RUNDEV CONTROL fm RDVnnnnn CONTROL fm
Copy the ADALOD control statements to load the demo files:
COPY loadname CONTROL fm demoname LODLIB fm
—where loadname is EMPLLOD, VEHILOD or MISCLOD, and demoname is EMPL, VEHI, and MISC, respectively. Change the copied files as required.
Create the file DBnnnnn VOLUMES filemode, where nnnnn is the five-digit database ID.
This file must contain one record per direct access mini-disk (that is, one for the Associator, one for Data Storage, and so on). Each record must contain the file name, file type, disk label, and virtual unit address (in that order) where the disk will be formatted and reserved. The record entries must be separated by blanks.
Copy the ADALOD control statements to load the demo files.
Execute the DBINIT EXEC specifying the DBID as a parameter.
Execute the ADALOD EXEC specifying FILENAME (EMPL,VEHI,MISC) and the DBID as parameters.
LOGON to the ID table manager virtual machine.
Create a PROFILE EXEC for the ID table manager virtual machine.
Refer to the section z/VM Operating System Environment for an example of the CP directory entries for the ID table manager virtual machine.
Include the following functions in the PROFILE EXEC:
multi-read LINK commands to the Adabas library disk (ADAVvvv TXTLIB).
ACCESS for the library disk as a read only extension of the A-disk (ACC cuu fm/A);
EXEC SETTXTLB;
LOAD ADAITM * (START.
Execute the PROFILE EXEC created in step 17.
DISCONNECT from the ID table manager virtual machine.
LOGON to the Adabas virtual machine.
Create a PROFILE EXEC for the Adabas virtual machine.
Refer to the section z/VM Operating System Environment for an example of the CP directory entries for the Adabas virtual machine.
Include the following functions in the PROFILE EXEC:
multi-read LINK commands to the Adabas library disk (ADAVvvv TXTLIB).
ACCESS for the library disk as a read-only extension of the A-disk (ACC cuu fm/A).
EXEC SETTXTLB
* SET STORECLR ENDCMD
EXEC NUCXTNTS (defines ADARUN, DATADEF, DISPDD, and RELDD as nucleus extensions).
Execute the PROFILE EXEC created in step 21.
Start the Adabas nucleus by entering ADANUC, specifying the DBID as a parameter.
If a DBA virtual machine has been defined, DISCONNECT from the nucleus virtual machine before continuing; otherwise, continue with step 25.
LOGON to the database administrator (DBA) virtual machine.
Execute the ADAREP EXEC, specifying as parameters the DBID and the filename of the z/VM file containing the ADAREP control statements (i.e. REPCPLST).
Stop the Adabas nucleus by entering ADAEND from the console of the Adabas virtual machine, or from a secondary console.
Execute the ADASAV EXEC specifying the filename of the z/VM file that contains the ADASAV control statements (i.e. SAVE) and the DBID as parameters.
For systems using the Adabas Online System, perform the following additional step:
LOGON to the DBA virtual machine.
This completes the installation without the online procedure.
In z/VM, the following entities are required or recommended to run Adabas.
Adabas nucleus (required)
ID table manager (required)
database administrator (recommended)
user programs (recommended)
Each of the above Adabas entities operates as a task within its own z/VM environment, known as a virtual machine. Each virtual machine comprises system resources that emulate virtual storage, virtual DASD (minidisk) space, and I/O capabilities such as an operator console and printer.
The actual system resource allocated to each Adabas virtual machine is either predefined in a CP directory or dynamically defined in the PROFILE EXEC that is invoked when each Adabas virtual machine begins operating. The following sections describe each of the Adabas virtual machines and how its resources are initially defined when installing Adabas.
Following this description is specific information concerning individual virtual machine requirements and operations in a z/VM environment.
The Adabas nucleus executes in its own virtual machine, which normally runs disconnected. The nucleus virtual machine requires various z/VM minidisks. The typical required virtual storage size for the Adabas nucleus machine is 6 MB.
An A-Disk of at least 500 4-kilobyte blocks is required, and if no DBA virtual machine exists, either a library minidisk of at least 1500 4-kilobyte blocks is needed, or a read only LINK to another nucleus virtual machine containing the library minidisk. Refer to the section Adabas Library Requirements.
Typically, the database resides on z/VM minidisks defined in the CP directory of the Adabas nucleus virtual machine. These minidisks must have multi-write passwords in the CP directory. See the section Disk Space Requirements for the Database.
To authorize the nucleus’ use of z/VM IUCV, place the following statement in the CP directory:
IUCV ALLOW (PRIORITY)
—where “PRIORITY” is optional.
If access to the nucleus machine is to be restricted, the IUCV statement should be in the CP directory of the user machines. In addition, the CP directory must have the MAXCONN parameter of the OPTION statement set high enough to accommodate one path to each z/VM or guest operating system’s user machine and two paths to the ID table manager virtual machine. For example:
OPTION MAXCONN 5
—would be sufficient to support three Adabas user machines and the ID table manager machine.
Before the Adabas nucleus is started, the following commands must also be issued:
ACCESS cuu fm/A EXEC SETTXTLB SET STORECLR ENDCMD —where cuu is the virtual unit address of the library minidisk. fm is the z/VM filemode.
If used, the online installation procedure creates these statements in the nucleus user machine’s PROFILE EXEC file.
The Adabas nucleus machine requires four nucleus extensions with the following attributes:
ADARUN USER SERVICE DATADEF SYSTEM DISPDD SYSTEM RELDD SYSTEM
The NUCXTNTS EXEC may be used to load the four nucleus extensions. If the online installation procedure is used, it adds a statement in the PROFILE EXEC to invoke the NUCXTNTS EXEC when the nucleus virtual machine starts.
If a new version of one of the nucleus extensions is to be activated the following commands must be entered:
ERASE extname MODULE A NUCXDROP extname NUCXTNTS
—where extname is the nucleus extension name. If NUCXTNTS is invoked without the two previous commands, the old version of the program remains active.
The DBA virtual machine can be defined to z/VM as a secondary console for the Adabas nucleus machine. This allows the DBA to control the nucleus machine, and issue Adabas operator commands using the SEND command from a terminal. The secondary console support is defined by specifying the CONSOLE statement in the CP directory of the Adabas nucleus machine as follows:
CONSOLE 009 devtype T dbavmid —where devtype is the DBA console device type. dbavmid is the DBA virtual machine ID.
The following is an example of the entries in the CP directory for the Adabas nucleus virtual machine running in ESA mode:
USER ADA00001 ADA00001 6MB 64M G ACCOUNT xx xxxxx IPL CMS PARM AUTOCR MACHINE ESA OPTION MAXCONN 100 IUCV ALLOW PRIORITY CONSOLE 009 3215 T SAGDBA SPOOL 00C 2540 READER * SPOOL 00D 2540 PUNCH A SPOOL 00E 1403 A
The LINK and MDISK statements are as follows:
LINK MAINT 190 190 RR LINK MAINT 19E 19E RR . . MDISK 191 3380 067 005 vvvvvv MR rpassword wpassword mpassword MDISK 200 3380 304 016 vvvvvv MW rpassword wpassword mpassword MDISK 300 3380 400 051 vvvvvv MW rpassword wpassword mpassword MDISK 400 3380 451 021 vvvvvv MW rpassword wpassword mpassword MDISK 410 3380 472 011 vvvvvv MW rpassword wpassword mpassword MDISK 420 3380 483 006 vvvvvv MW rpassword wpassword mpassword MDISK 430 3380 489 002 vvvvvv MW rpassword wpassword mpassword MDISK 431 3380 491 002 vvvvvv MW rpassword wpassword mpassword MDISK 440 3380 493 002 vvvvvv MW rpassword wpassword mpassword MDISK 441 3380 495 002 vvvvvv MW rpassword wpassword mpassword
The ECMODE OPTION statement, though not required, is recommended because it enables the use of certain VM assists.
The ID table manager executes in its own virtual machine, which normally runs disconnected. For information about the function of the ID table manager, refer to the section Functions of the ID Table Manager.
The virtual storage size required for the ID table manager is four megabytes. The A-disk must be at least 100 4-kilobyte blocks and have a read-only link to the Adabas library disk (where the ADAVvrs TXTLIB and all CONTROL files reside).
By default, the ID table manager allows 20 ID table entries. A zap can be used to extend this limit:
NAME ADAITM ADAITM VER 0022 00FF dafault 20 REP 0022 00xx new default
The ID table manager must be authorized for IUCV communication. The following entries should be placed in the CP directory:
IUCV ALLOW (PRIORITY) IUCV ANY (PRIORITY)
—where “PRIORITY” is optional, but recommended. In addition, the OPTION statement’s MAXCONN parameter in the CP directory must be set high enough to accommodate an IUCV path to each user virtual machine, either z/VM or a guest operating system, and two paths to each Adabas nucleus machine.
The PROFILE EXEC of the ID table manager virtual machine should contain the following commands:
LINK SAGDBA cuu cuu RR rpassword ACC cuu fm/A EXEC SETTXTLB LOAD ADAITM (RESET ADAITM START * ...
—where cuu is the virtual unit address of the library minidisk and fm is the z/VM filemode.
Software AG recommends starting the ID table manager machine automatically at system startup time using the AUTOLOG command in the PROFILE EXEC of the AUTOLOG1 virtual machine. If set up to do so, the ID table manager virtual machine’s PROFILE EXEC executes AUTOLOG to start the Adabas virtual machines.
The ID table manager accepts parameters for the following values:
DBA virtual machine ID, keyword DBAVMID;
ID table manager target ID, keyword NODEID;
ID table manager node name, keyword NODENAME;
Entire Net-work IUCV resource ID, keyword NETRESID.
These keyword parameters are specified in the z/VM START command. For example, to execute the ID table manager with a node ID of 1001 and a node name of CMSNODE, enter the following START command:
START * NODEID 1001 NODENAME CMSNODE
Except for the DBAVMID parameter, these parameters are only relevant to operation with Entire Net-Work.
The following is an example of the entries in the CP directory for the ID table manager virtual machine running in ESA mode:
USER DBIDSERV DBIDSERV 6M 8M G ACCOUNT xxxxxxxx OPTION MAXCONN 100 IUCV ALLOW PRIORITY IUCV ANY PRIORITY IPL CMS ESA PARM AUTOCR CONSOLE 009 3215 T SAGDBA SPOOL 00C 2540 READER * SPOOL 00D 2540 PUNCH A SPOOL 00E 1403 A LINK MAINT 190 190 RR LINK MAINT 19E 19E RR MDISK 191 3380 2741 vvvvvv MR rpassword
Under z/VM, commonly addressable and modifiable storage areas are not available. This means that there can be no centrally located routing table as in other operating systems.
Instead, this information is maintained and distributed by the ID table manager program, ADAITM, which executes in its own virtual machine. It is a required virtual machine and operates continuously in disconnected or background mode.
ADAITM also provides all information needed for communication between z/VM and any guest operating systems running under VM.
The Adabas nucleus (ADAMPM) and ADALNK establish an IUCV communications path to the virtual machine running ADAITM during the startup procedure. The VMID of the ID table manager virtual machine is at a fixed location in ADALDI, a program used by ADAMPM and ADALNK for communications. Both this VMID and the target ID of the ID table manager can be set at installation time and should not be changed. The default VMID is DBIDSERV and the default target ID is 65535.
The DBA virtual machine ID, ID table manager target ID, and node name can all be specified in the ID table manager PROFILE EXEC as parameters:
LOAD ADAITM ( RESET ADAITM’ START * DBAVMID vm-id NODEID node-id NODENAME node-name NETRESID res-id —where vm-id is the ID of the DBA virtual machine node-id is the ID table manager target ID node-name is the ID table manager node name res-id is the Entire Net-Work IUCV resource ID
If desired, the DBA virtual machine ID and ID table manager target ID (node ID) can also be zapped.
With the exception of the DBAVMID parameter, the parameters above are only relevant to operation with Entire Net-Work.
The following zap modifies the default DBID and the database administrator machine VMID (described later) in the ID table manager:
NAME ADAITM ADAITM VER 0020 FFFF VER 0028 4040,4040,4040,4040 REP 0020 FFFF REP 0028 E2C1,C7C4,C2C1,4040
This zap illustrates how to modify VMID of the ID table manager in ADALDI:
NAME ADALDI ADALDI VER 0034 C4C2,C9C4,E2C5,D9E5 REP 0034 C4C2,C9C4,E2C5,D9E5
If an ADAITM program error occurs and the restart option is active, all information concerning the Adabas z/VM environment is written to the z/VM file ADAITM RESTART on the ID table manager’s A-Disk and a message is sent to both the CP operator and (if one has been defined) to the database administrator virtual machine consoles. ADAITM then severs all IUCV paths, places a command in the z/VM program stack to reload and restart itself, and stops.
If the restart option is not active, no ADAITM RESTART file is written; the ID table manager logs off, and terminates all active nuclei abnormally.
During the time that ADAITM is not available, existing local communication paths may still be used; however, no new paths can be initiated and no new Adabas nuclei can be started. All remote communications are terminated.
After restarting, ADAITM reconnects to all Adabas z/VM nuclei and to any Adabas users in z/VM machines who are still active. If another error occurs during error recovery or restart, ADAITM sends a message to both the CP operator and the database administrator virtual machine consoles, and then logs off the ADAITM virtual machine. This causes all local Adabas nuclei to abend. Stopping the ADAITM virtual machine with FORCE also abends all local nuclei.
ADAITM accepts operator commands for administration purposes. These operator commands comprise two general categories: listing commands and trace commands. Listing commands (LISTxxxx) display console lists of the requested information, usually as AITMnn messages. Trace commands provide chronological information such as nucleus/user initialization and termination.
The commands are
| Command | Description |
|---|---|
| DISPON | Displays events on the z/VM console as they occur |
| DISPOFF | Stops event display on the z/VM console |
| LISTLINK | Displays all active links to directly addressable network nodes (for Entire Net-Work systems only) |
| LISTLOG | Displays the contents of the logging area on the console |
| LISTMSG | Displays the contents of the NETITM MSGS file, which contains messages related to Entire Net-Work communication |
| LISTNODE | Displays all the active network nodes on the console |
| LISTTARG | Lists active targets (active nuclei, etc.) on the console |
| LISTUSER | Lists all user virtual machines on the console |
| LOGON | Records events in a logging area |
| LOGOFF | Stops logging of events |
| RESET | Clears the event logging area |
To give the database administrator (DBA) maximum control of the Adabas environment, a separate DBA virtual machine should be allocated. This allows the DBA to
perform multi-user utility operations on multiple databases;
maintain a single Adabas library disk, ADAVvrs TXTLIB; and
issue operator commands for nuclei running in disconnected machines (when the DBA virtual machine is defined as a secondary console for the nucleus machine).
The DBA virtual machine requires a minimum of 4 megabytes of virtual storage. Depending on the Adabas utilities to be run in the DBA virtual machine and the parameters specified, more storage may be required. The standard virtual storage size is 4 megabytes.
The A-Disk must be at least 750 4-KB blocks (or the equivalent). If this virtual machine is being used, the Adabas library disk must be defined in the CP directory for the DBA virtual machine. See the section Adabas Library Requirements.
Multiple write links must be defined from the DBA virtual machine to each database minidisk that the DBA machine supports. If the online installation procedure is used and LINK passwords are specified, LINK statements are automatically created in the DBA virtual machine’s PROFILE EXEC.
The DBA virtual machine must be authorized for IUCV communication. The following entry should be placed in the CP directory:
IUCV ALLOW PRIORITY
The MAXCONN parameter of the OPTION statement in the CP directory must be set high enough to accommodate an IUCV path to the ID table manager and paths to each Adabas nucleus machine.
To execute Adabas utilities, the DBA virtual machine must have the following statements in its PROFILE EXEC:
ACCESS cuu fm/A EXEC SETTXTLB SET STORECLR ENDCMD —where cuu is the virtual unit address of the library minidisk fm is the filemode of the library minidisk
If used, the online installation procedure creates these statements in the PROFILE EXEC.
The database administrator virtual machine requires four nucleus extensions with the following attributes:
ADARUN USER SERVICE DATADEF SYSTEM DISPDD SYSTEM RELDD SYSTEM
The NUCXTNTS EXEC may be used to load the four nucleus extensions. If the online installation procedure is used, it adds a statement to the PROFILE EXEC to invoke the NUCXTNTS EXEC when the DBA virtual machine starts.
If a new version of one of the nucleus extensions is to be activated, enter the following commands with NUCXTNTS:
ERASE extname MODULE A NUCXDROP extname NUCXTNTS
—where extname is the nucleus extension name.
If NUCXTNTS is invoked without ERASE and NUCXDROP, the old version of the program remains active.
The DBA virtual machine’s CP directory requires an OPTION statement. The following is an example of the entries in the CP directory for the DBA virtual machine:
USER SAGDBA SAGDBA 6M 8M G ACCOUNT xxxxxxx OPTION MAXCONN 10 IUCV ANY IPL CMS PARM AUTOCR
Each user virtual machine requires 1536 KB of virtual storage if Natural is to be used without a DCSS. If Natural is installed as a DCSS, 768 KB of virtual storage is sufficient.
Each user virtual machine must LINK and ACCESS to either the Adabas library disk (that is, where ADAVvrs TXTLIB resides), or to a sublibrary disk. Application programs executing in SINGLE user mode must use the Adabas library disk ADAVvrs. A sublibrary disk is created when users are to be isolated from the Adabas library disk. For more user information on the sublibrary disk refer to the section Creating a User Sublibrary.
The user machine only needs to be authorized to use IUCV when the ID table manager and nucleus virtual machines do not have IUCV ALLOW statements in their respective CP directory entries. If neither the ID table manager nor the Adabas nucleus virtual machine has the IUCV ALLOW statements, issue the following statements in the user CP directory:
IUCV idtmvmid IUCV nucvmid —where idtmvmid is the ID table manager ID nucvmid is the Adabas nucleus virtual machine ID
Refer to the section IUCV Security Options. The MAXCONN parameter of the OPTION statement in the CP directory must be set high enough to accommodate an IUCV path to the ID table manager and one path to each Adabas nucleus machine.
To execute ADALNK, the user virtual machine must have the following statements in its PROFILE EXEC:
ACCESS cuu fm/A GLOBAL TXTLIB libname —where cuu is the virtual unit address of the library or sublibrary minidisk fm is the filemode of the library or sublibrary minidisk libname is the TXTLIB filename
A user virtual machine requires four nucleus extensions with the following attributes:
ADARUN SYSTEM SERVICE DATADEF SYSTEM DISPDD SYSTEM RELDD SYSTEM
The NUCXTNTS EXEC must be issued with the parameter USER to load the four nucleus extensions.
To activate a new version of a nucleus extension, enter the following commands:
ERASE extname MODULE A NUCXDROP extname NUCXTNTS USER
—where extname is the nucleus extension name.
If NUCXTNTS is invoked without ERASE and NUCXDROP, the old version of the NUCXTNTS program remains active.
The DBA virtual machine’s CP directory requires an OPTION statement as shown in the following example:
USER ADAUSER1 ADAUSER1 6M 8M G ACCOUNT xxxxxxx IPL CMS PARM AUTOCR MACHINE ESA OPTION MAXCONN 10 CONSOLE 009 3215 SPOOL 00C 2540 READER * SPOOL 00D 2540 PUNCH A SPOOL 00E 1403 A LINK MAINT 190 190 RR LINK MAINT 19E 19E RR MDISK 191 3380 381 007 vvvvvv MR rpassword
ADARUN control statements can be used to control user programs by issuing the appropriate DATADEF statement for DDCARD and optionally for DDPRINT. The DATADEF statement assures that all ADALNK errors are both displayed on the virtual console and recorded in the DDPRINT file.
The Adabas communications environment can be reset and all storage allocated to that environment in a user virtual machine can be released by issuing the following z/VM command:
NUCXDROP ADARUN
Because the initialization and termination of the communication environment uses a lot of system resource, the NUCXDROP command is recommended only when no further use of Adabas communication from the user virtual machine is planned for the session.
The Adabas library disk must be at least 2700 4-KB blocks (or the equivalent). Each added database after the first adds a requirement of 100 4-KB blocks to the library disk. The library minidisk must be defined with a multiread or ALL password in either the DBA virtual machine’s CP directory, or in the nucleus virtual machine’s CP directory if no DBA machine has been defined. All virtual machines accessing the database must have read-only LINKs to the library minidisk.
Installations wishing to isolate multiuser application programs from the Adabas library disk can define a sublibrary on a commonly accessible minidisk, as described in the following section Creating a User Sublibrary. Application programs running in SINGLE user mode must use the Adabas library disk.
A sublibrary disk is created to isolate users from the Adabas library disk. The sublibrary disk must be at least 200 4-KB blocks, be located on a commonly accessible minidisk, and have a multiple-read password. The sublibrary TXTLIB must contain the following members:
$SAGIOS ADALNK NETPARS NIUDEP ADAIOR ADARUN NETRQM NIUEXT ADAIOS ADAUSER (ADABAS) NETTQM ADAILD NETBPM NETTRC ADALDI NETIUCV NETTRT
To copy these members from the Adabas library disk, issue the following commands for each member:
FILEDEF IN DISK ADAVvvv TXTLIB fm (MEMBER membname FILEDEF OUT DISK membname TEXT fmwork MOVEFILE IN OUT —where fm is the filemode of the Adabas library disk fmwork is the filemode of a work minidisk membname is the name of the member to be copied to the sublibrary.
Note:
ADAUSER has the member name ADABAS.
To create the new TXTLIB sublibrary, enter the following command:
TXTLIB GEN sublib membname1,membname2,....membnamex —where sublib is the name of the sublibrary membname1 — membnamex are the names of members taken from the ADAVvrs TXTLIB.
After TXTLIB has been created, the text files can be deleted from the work minidisk. It is important to remember that a sublibrary needs separate maintenance; this means that zaps and new SM levels must be applied to both the Adabas library and each sublibrary. The NUCXTNTS, DEFNUCX, and SETXTLB EXEC routines must also be copied to the new sublibrary.
The following additional modules must be generated and then copied to the TXTLIB sublibrary minidisk:
ADAIOR DATADEF ADALNK DISPDD ADARUN RELDD
The z/VM module files having the same names should also be copied. You must also ensure that the sublibrary minidisk is updated whenever the main library minidisk is changed.
Each user machine utilizing the sublibrary minidisk must have the appropriate LINK, ACCESS and GLOBAL TXTLIB statements active. See the section User Virtual Machine.
The database can be on either DASD having MVS or VSE VTOCs, or on VM minidisks. For MVS or VSE disks, there must be multi-write LINKs or ATTACH commands to those disks in the PROFILE or CP directory of the nucleus virtual machine. For VM minidisks, the nucleus virtual machine’s CP directory must contain the MDISK commands, and multi-write passwords must be defined.
The Adabas database requires at least five VM minidisks, as follows:
| File | Suggested Unit Address | Required Space in Blocks | 3380 Cylinders | 3350 Cylinders |
|---|---|---|---|---|
| ASSO | 200 | 8640 (2k) | 30 + 2 | 35 + 1 |
| DATA | 300 | 10800 (4k) | 70 + 2 | 85 + 1 |
| WORK | 400 | 1650 (4k) | 10 + 2 | 15 + 1 |
| TEMP | 410 | 2250 (4k) | 15 + 1 | 20 + 1 |
| SORT | 420 | 2250 (4k) | 15 + 1 | 20 + 1 |
Optional minidisks and their recommended sizes and requirements are:
| File | Suggested Unit Address | Required Space in Blocks | 3380 Cylinders | 3350 Cylinders |
|---|---|---|---|---|
| CLOG1 | 430 | 300 (4k) | 1 + 1 | 2 + 1 |
| CLOG2 | 431 | 300 (4k) | 1 + 1 | 2 + 1 |
| PLOG1 | 440 | 300 (4k) | 1 + 1 | 2 + 1 |
| PLOG2 | 441 | 300 (4k) | 1 + 1 | 2 + 1 |
All sizes specified for the minidisks must be rounded up to the next full multiple of cylinders. For FBA devices, see the section FBA Devices.
The first cylinder or FBA pseudo-cylinder of each database minidisk is reserved for the z/VM directory. Therefore, one cylinder must be added to each minidisk allocated in the CP directory. The last column of each of the tables above accounts for the directory cylinder by adding one to the normal cylinder/pseudo-cylinder count.
The z/VM RESERVE command must be specified for each of the database minidisks. If the online installation procedure is being used, this RESERVE operation is performed when the installation procedure is executed.
Any Adabas virtual machine that executes multiuser utilities must define multi-write LINKs to the database minidisks.
Adabas z/VM supports direct access database files on z/VM-formatted disks or on CKD and FBA disks with an MVS/VSE VTOC. The disks can be minidisks or real disks. The files may be contained on one or more real or virtual volumes.
z/VM-formatted disks must not be attached with the CP ATTACH command, but must be defined in the directory of the virtual machine or linked to it using the CP LINK command. Disks with database files need not be accessed using the z/VM ACCESS command. If a file spans multiple volumes, all of the devices must be of the same type and must have the same format; that is, either all devices must have an MVS or VSE VTOC, or all must be z/VM-formatted.
Sequential files are supported as
terminal files (input/output)
card reader files (input only)
card punch files (output only)
printer files (output only)
tape files (input/output)
z/VM-formatted disk files (input/output)
MVS- or VSE-formatted disk files (input only)
The file format for terminal files must be fixed length with a maximum record length of 80. When an input file is assigned to terminal, the user is prompted with the name of the file being read. The end of a terminal input file is signaled by pressing ENTER without any input.
Sequential files on other unit record devices must also be fixed length. The maximum record length is the length supported by z/VM on the device.
All record formats and lengths are supported for tape files. If the tape file was created under an operating system which does not create HDR 2 records, the user must supply file format and length information. Tape files must reside on standard labeled volumes. The z/VM TAPE command may be used to initialize a tape volume and set its density. To avoid confusion during tape handling, it is highly recommended that each tape be labeled with a unique 6-byte name. Both multi-file volumes and multi-volume files are fully supported.
Sequential files on z/VM-formatted disks may be either fixed or variable length. Blocking is done by the z/VM file service routines. Unlike DASD volumes used for direct access files, disks containing sequential files must be ACCESSed. The z/VM file services do not differentiate between nonexistent files and empty files. To prevent open errors in Adabas utilities when a non-existent input file is used, a one-byte record containing a dollar sign ($) is written to sequential files on z/VM formatted disks at close time if no records have been written.
All record formats and lengths supported by MVS or VSE may be used on sequential MVS/VSE DASD files. If there is no record information in the label records (VSE), record format and length must be specified by the user. MVS/VSE DASDs may not be used for sequential output files.
Adabas supports concatenation of sequential files if the record format and the logical record lengths of the concatenated files are the same. Concatenation of files on different storage media is also possible.
Sequential files to be read backward can be on either tape or z/VM-formatted disks. Exceptions to this are tape files created using the improved data recording capability (IDRC) on certain cassette units. When this facility is in use on certain other models of tape units, the read backward channel command is not supported. This point should be considered when planning recovery strategies for rebuilding the database after a possible failure.
Where supported, sequential files to be read backward may also be concatenated. If they are concatenated, they must be specified in the same order as though they were being processed normally.
The tape units to be used must be attached before file assignment by DATADEF unless they are to be dynamically allocated at open time by a tape management system. In this last case, a generic name can be specified in the DATADEF statement (see the section DATADEF File Assignments).
Tape units can be attached to any virtual address, and need not have any logical name. During the open process, Adabas reads the first block of the tape to verify that the correct volume is mounted. If the correct volume is not mounted or is not correctly initialized, the message ADAI48 is issued and the tape is unloaded. Message ADAI40 is then issued to the operator, and message ADAI41 is sent to the CP operator.
When a tape volume is successfully mounted, the message ADAI42 is displayed on the z/VM virtual console. This message also indicates whether the IDRC is in effect. If a cassette created with IDRC is used on a unit which does not support IDRC, an I/O error occurs.
If the volume parameters were supplied to DATADEF, the required volume serial number is included in both messages; if an output file is to be created and no volume serial number has been specified, the constant SCRTCH is displayed. This means that any unused standard label work (commonly called “scratch”) tape should be mounted. The two messages are repeated once each minute until the mount request has been satisfied. To reset the mount request, the z/VM operator must press the ENTER key twice. This causes open processing for the file to be terminated abnormally, and is equivalent to having defined the file as ‘dummy’.
You can control tape file management with a program that uses the optional user exit CMSUX1. If a program with the name CMSUX1 is loaded with the z/VM LOAD command before invoking a utility or nucleus that requires a tape file, the CMSUX1 program is called
when the tape file is opened;
whenever a tape volume must be mounted; and
when the file is closed.
If a tape is rejected after a mount request because of a missing VOL1 record or incorrect volume/serial number, the CMSUX1 program is also reinvoked before a mount request is repeated.
When the CMSUX1 user exit receives control, general register 0 contains one of the following function codes:
1 for open 2 for tape mount 3 for close
General register 1 contains the address of a parameter list, which is passed to the CMSUX1 user exit program. The parameters are
address of the data definition block (DDB) (non-modifiable);
address of the six-byte tape volume serial number which will be requested (modifiable);
address of a fullword containing:
an input/output flag byte (X‘00’ for input files, or X‘80’ for output files);
a reserved byte;
a two-byte hexadecimal tape unit address in the format X‘0cuu’ (ESA: X‘ccuu’). If the DATADEF statement specified TAPx in the UNIT parameter, this field contains the corresponding “18x” value according to standard z/VM tape unit conventions.
address of a fullword containing the number of times the mount has been attempted (non-modifiable).
If the volume serial number has been modified, the new volume serial number is requested. If the file is an input file, however, this modification is not reflected in the DDB.
Upon return to the system, one of the following return codes is expected in general register 15:
0 proceed with the mount processing 4 proceed with the mount processing without issuing any operator messages 8 abort the mount processing
If return code 8 is encountered, the mount operation terminates as if an I/O error had occurred.
A sequential system file with the name of DUMP is available to facilitate dumping. This file may be assigned to printer, a z/VM file or DUMMY. If the file is DUMMY a CP dump will not be taken when an error occurs.
All assignments for files which are accessed by an Adabas nucleus or utility must be done using DATADEF, which replaces the z/VM FILEDEF command in the Adabas environment. DATADEF accepts parameters as either a tokenized or extended parameter list; the extended parameter list takes precedence.
The file assignments established by the DATADEF statements can be listed using the DISPDD program. The file name of a specific DATADEF statement can be entered as a DISPDD parameter; if done, only the information for that file is displayed. If no parameter is specified, information for all assigned files is displayed.
The program RELDD can be used to clear active DATADEF entries. RELDD accepts a list of file names to be released or—if no list is specified—clears all active DATADEF entries.
The DATADEF statement creates a data definition block (DDB), which remains in system storage until it is
overwritten by another DATADEF statement with the same name;
cleared by RELDD; or
cleared by a z/VM IPL.
The parameters for DATADEF consist of one positional parameter and one or more keyword parameters separated by commas. An equal sign (‘=’) must be used between a keyword and the parameter value. Depending on how DATADEF is invoked, spaces may be required surrounding equal signs, commas, and parentheses.
The DATADEF parameters are described in the following table:
| Parameter Keyword | Required/ Optional | Maximum Length | Specifies . . . |
|---|---|---|---|
| positional | Required | 8 | the file name (DD) names as specified in Adabas Operations or Adabas Utilities. |
| BLKSIZE | See note 2 | 5 | the length of the physical blocks in the file. If RECFM=FB, BLKSIZE must be an integral multiple of LRECL; if RECFM=V or RECFM=VB, BLKSIZE must be at least equal to LRECL + 4. |
| BUFNO | Optional | 3 (1 - 255) | the number of buffers to be allocated for a sequential file on tape. Default: 5 |
| COMPRESS | Optional | 3 (YES - NO) | whether or not an output file on tape should make use of IDRC available on certain cassette units. If IDRC is not supported, this parameter is set to NO, the default. |
| CONCAT | Optional | 3 (1 - 255) | a concatenation sequence number for the file. This results in the DDB being concatenated to another existing DDB with the same file name. The sequence numbers must be specified in ascending order with no numbers left out. The first file to be concatenated has the number 1. If specified for a file to be read backward, the sequence numbers are to be given in the normal sequential order. A DATADEF statement without CONCAT frees any existing root DDB and any DDBs concatenated to it. |
| DISP | Optional for output only | 3 | whether a sequential file is to be created (NEW), extended (MOD) or overwritten (OLD). If NEW is specified and the file exists a return code is issued. If OLD is specified, the file’s existence is not checked. Default: OLD |
| DSN | Required if not dummy | 44 | the dataset or SFS directory name. |
| DUMMY | Optional | - | that the file does not exist. DUMMY is a keyword without a value. It may not be specified with any other parameter except file name. |
| EXTEND | Optional | - | that an existing DDB is to be extended. EXTEND is a keyword without a value. It may only be specified with the file name, VOL, and UNIT parameters. |
| FILESEQ | Tape only, optional | 3 | the sequence number of the file on a multi-file tape. The default value (1) is required if tapes are to be read backward. |
| FNAME | Required if DSN is a SFS | 8 | the file name of an SFS member. |
| FTYPE | Required if DSN is a SFS | 8 | the file type of an SFS member. |
| LRECL | See note 2 | 5 | the length of the physical blocks in the file. If RECFM=FB, BLKSIZE must be an integral multiple of LRECL; if RECFM=V or RECFM=VB, BLKSIZE must be equal to or more than LRECL + 4. |
| MODE | See note 1 | 2 | the z/VM filemode. |
| RECFM | See note 2 | 2 | the format of the records in the file (F, FB, V, VB, U). |
| UNLOAD | Tape only, optional | 3 (YES - NO) | whether or not the tape is rewound and unloaded. If NO is specified, the tape is rewound at close but is not unloaded. Default: YES. |
| UNIT | See notes 1 and 3 | See note 1 | a list of virtual addresses (cuu, or ccuu for ESA) of the unit or units containing the file, or one of the logical device abbreviations: TRM, PUN, RDR, PRT, or SFS. If a unit address list is given, it must be enclosed in parentheses and entries must be separated by commas. |
| VOL | See note 1 | See note 1 | a list of the serial numbers (each at most 6 characters) of the volumes containing the file; if multiple volumes are specified, they must be separated by commas and enclosed in parentheses. |
Notes:
A MODE parameter is required for sequential z/VM DASD files. For DASD volumes containing database files, either a VOL, UNIT, or MODE parameter is required. If the database file spans multiple volumes, VOL or UNIT must be specified. Specifying MODE=* for an non-existent file results in a return code of 32.
If both VOL and UNIT are specified, the number of volumes and unit addresses must be equal and each volume in the VOL list must be mounted on the unit specified by the corresponding entry in the UNIT list (the first VOL entry must be mounted on the first UNIT entry, and so on).
For tape files, a UNIT must be specified and a real unit attached prior to DATADEF execution, unless the tape unit is dynamically allocated at open time; in this case, TAPx can be specified according to standard z/VM conventions, where TAP1 specifies virtual unit 181, TAP2 specifies unit 182, and so on. Only one tape unit address is allowed in the UNIT parameter.
The VOL parameter is required for input tape files, but is optional for output tape files. If a tape file spanning multiple volumes is to be read backwards, specify the volumes in the normal sequential order.
When creating a multi-volume tape file, Adabas z/VM maintains a list of the file volumes. To refer to that volume list in a later DATADEF, specify VOL=*filename where filename is the name of the multi-volume file.
The parameters RECFM, LRECL, and BLKSIZE are required only for tape input files without a HDR2 label and for VSE sequential DASD files. If RECFM has been specified, the corresponding BLKSIZE and LRECL parameters are also required.
If UNIT=SFS, the DSN parameter is used as the SFS directory name and the FNAME and FYTPE parameters are used as the file name and file type, respectively. For example:
DATADEF DDCARD,DSN=SFSPOOL:USERID.DTR1,UNIT=SFS,FNAME=DB52,FYPTE=RUN1 or DATADEF DDCARD,DSN=.DTR1,UNIT=SFS,FNAME=DB52,FYPTE=RUN1 or DATADEF DDSAVE1,DSN=.DTR1,UNIT=SFS,FNAME=SSF,FYPTE=RUN001
The following error codes may be returned by DATADEF:
| Response Code | Description |
|---|---|
| 16 | No parameter list was supplied |
| 20 | Invalid keyword |
| 24 | No file name specified |
| 28 | Error in DSN or DUMMY specification: neither a DSN nor DUMMY was specified; or conflicting parameter specification for a dummy file |
| 32 | Error in VOL, UNIT or MODE parameter |
| 36 | Incorrect length for VOL,UNIT or MODE parameter |
| 40 | Insufficient virtual storage |
| 44 | Internal error issuing a CP command |
| 48 | Invalid cuu address (internal error) |
| 52 | Volume or unit not available or non-VTOC volume has been attached |
| 56 | Database file resides on volumes with mixed formats (VTOC, non-VTOC) |
| 60 | Database file resides on volumes of different device types |
| 64 | More than one unit specified for a tape file |
| 68 | Invalid file sequence number |
| 72 | Invalid RECFM parameter |
| 76 | Invalid BLKSIZE parameter |
| 80 | Invalid LRECL parameter |
| 84 | Invalid DISP parameter |
| 88 | Invalid concatenation count |
| 92 | Invalid DDB extension |
These codes can be returned by invoking the CODES EXEC with the error number as a parameter.
Adabas operator commands can be entered directly from the virtual machine console of either the Adabas nucleus virtual machine, or from another virtual machine console authorized as a secondary console.
In addition, certain z/VM commands can also be entered as Adabas operator commands. This feature allows the database administrator to perform certain display or query functions while the Adabas nucleus is active. To enter a z/VM command, enter the command with the prefix CMS, followed by at least one blank before the command.
Caution:
If issued while Adabas is active, certain z/VM commands can have
adverse effects on the nucleus, causing abnormal operation and even the
erroneous data. Note also that issuing a z/VM command causes the nucleus to
stop operation temporarily while the command is being processed, making Adabas
unavailable during that time. The use of this z/VM command facility is solely
the responsibility of the user; Software AG cannot accept responsibility for
damage or loss that may occur when using this facility
Default values for the ADARUN parameters device type and database ID can be zapped into ADARUN at the offsets X’5D8’ and X’8E8’, respectively. For convenience, a file containing sample zap control statements to modify ADARUN has been supplied on the release tape. The file is named ADARUN ZAP. After making the necessary changes to this file, the user can enter the following commands to modify ADARUN:
ACC cuu fm ZAP TXTLIB ADAVvvv (INPUT ADARUN ACC cuu fm/A —where cuu is the virtual unit address of the Adabas library minidisk fm is the z/VM filemode
New SM levels may be installed by performing the following steps:
to install new SM levels:
Attach a tape unit to the database administrator virtual machine and mount a scratch tape.
Back up the existing library disk:
TAPE DUMP * * fm (fm is the filemode of the Adabas library disk) TAPE RUN
Mount the system maintenance tape.
Enter the following commands:
ACC cuu fm (cuu is virtual unit addr, fm is the filemode of the Adabas library disk) SMADA
If you have either created a user sublibrary, or if you have the ADARUN, DATADEF, RELDD or DISPDD modules on a commonly accessible minidisk, you must refresh the respective sublibrary and/or minidisk.
In the z/VM environment, a zap should be applied using the ZAP command, to verify and replace data. The zap input control statements are in MVS ZAP format. The ZAP command will apply the zap to a member of a z/VM TXTLIB.
The format of the ZAP command is:
ZAP TXTLIB libname (INPUT filename PRINT
—where libname is the Adabas TXTLIB (ADAVvrs) filename is a z/VM file with a filetype of ZAP that contains the input control statements.
The following are examples of ZAP control statements.
NAME membername csectname VER disp data REP disp data END LOG fixnum ZAPLOG text * (COMMENT)
Note:
In the verify (VER...) and replace (REP...) statements, commas are
acceptable data separators. However, commas with spaces or spaces alone are not
acceptable data separators and can result in errors.
Refer to the IBM z/VM Operator’s Guide for information about the ZAP service program.
The inter-user communications vehicle (IUCV) is a communication facility that allows one virtual machine to communicate with another virtual machine. Installations control the use of IUCV through the virtual machine directory entries.
In the Adabas environment, security on IUCV communications can be implemented by supplying different IUCV control statements. The following are examples of IUCV security options where:
| IUCV ALLOW | is a general authorization indicating that any other virtual machine may establish a communications path with this virtual machine. No further authorization is required in the virtual machine initiating the communication. |
| IUCV ANY | is a general authorization indicating that a communications path can be established between this virtual machine and any other virtual machine. |
| IUCV userid | is the one- to eight-character user identification of the virtual machine with which this virtual machine is authorized to communicate. |
| idtmvmid | is the VMID of the ID table manager. |
| nucvmid | is the VMID of the Adabas nucleus. |
ID Table Manager: IUCV ALLOW (PRIORITY) IUCV ANY (PRIORITY) Adabas Nucleus: IUCV ALLOW (PRIORITY) DB Administrator: IUCV ALLOW (PRIORITY) User: No IUCV CP directory entry required
ID Table Manager: IUCV ALLOW IUCV ANY Adabas Nucleus: No IUCV CP directory entry required DB Administrator: IUCV ALLOW IUCV nucvmid User: IUCV nucvmid
ID Table Manager: IUCV ANY IUCV ALLOW Adabas Nucleus: IUCV idtmvmid DB Administrator: IUCV idtmvmid IUCV nucvmid User: IUCV idtmvmid IUCV nucvmid
Since link routines are dynamically loaded in most environments, it should only be necessary to replace the existing Adabas TXTLIB or TEXT files with the new Adabas TXTLIB, ADAVvrs. Programs that are either GENMODed or in a discontiguous shared segment (DCSS) must be regenerated.
One or two user exits may be linked with ADALNK:.
UEXITB receives control before a command is passed to a target.
Note:
Special commands emanating from utilities and from Adabas Online
System are marked as physical calls. These calls must be bypassed in user
exits. These calls have X‘04’ in the first byte (TYPE field) of the command’s
Adabas control block (ACB). UEXITB must check this byte and return if it is set
to X‘04’. Be sure to reset R15 to zero on return.
UEXITA receives control after a command has been completely processed.
The user exits must be specified in the LOAD command to be active.
At entry to the exit(s), the registers contain the following:
| Register | Contents |
|---|---|
| 1 | Address of the UB.
If the flag bit UBFINUB is reset, the contents of the halfword at Adabas + X’86’ have been moved to UBLUINFO. If those contents are greater than zero, the two bytes starting at UBINFO (UB+X’40’) have been set to zero. If UBFINUB is set, no changes can be made to the UB or ACB (except for ACBRSP). |
| 13 | Address of an 18-word save area (for non-CICS Adalink exits) |
| 14 | Return address |
| 15 | Entry point address: UEXITB or UEXITA |
Any registers except register 15 that are modified by the user exits must be saved and restored; the address of a save area for this purpose is in register 13.
If at return from UEXITB register 15 contains a value other than zero (0), the command is not sent to the target but is returned to the caller. The user exit should have set ACBRSP to a non-zero value to indicate to the calling program that it has suppressed the command: response code 216 is reserved for this purpose.
The UEXITB exit may set the UB field UBLUINFO to any lesser value, including zero; an abend occurs if the user exit sets UBLUINFO to a greater value. The UBLUINFO length cannot be changed when any other exit is used; for example, Adabas Review.
The user information received by a UEXITA exit may have been modified; this modification may include decreasing its length, possibly to zero, by any of the ADANUC user exits.
ADALNK can return the following non-zero response codes in ACBRSP:
| Response Code | Description |
|---|---|
| 216 | UEXITB suppressed the command |
| 218 | No UB available |
The following two EQUates, described at the beginning of the source, can be modified before ADALNK is assembled. Other Adalinks allow this information to be zapped.
| Equate | Description |
|---|---|
| LOGID | The default logical ID, ranging in value from 1 to 65535. The default is 1. |
| LNUINFO | The length of the user information to be passed to Adalink user exits, ranging in value from 0 to 32767. The default is 0. |
The module LNKUES provides Universal Encoding Support (UES). This module must be linked into the standard batch ADALNK. LNKUES converts data in the Adabas buffers and byte-swaps, if necessary, depending on the data architecture of the caller.
Prior to Version 7, Entire Net-Work converted all data for mainframe Adabas. When Entire Net-Work Version 5.5 and above detects that it is connected to a target database that converts data, it passes the data through without converting it.
LNKUES is called only on ADALNK request (X’1C’) and reply (X’20’) calls if the first byte of the communication ID contains X’01’ and the second byte does not have the EBCDIC (X’04’) bit set.
For requests, LNKUES receives control before UEXITB.
For replies, LNKUES receives control after UEXITA.
By default, two translation tables are linked into LNKUES/ADALNK:
ASC2EBC: ASCII to EBCDIC translation; and
EBC2ASC: EBCDIC to ASCII translation.
Note:
It should only be necessary to modify these translation tables in
the rare case that some country-specific character other than "A-Z a-z 0-9"
must be used in the Additions 1 (user ID) or Additions 3 field of the control
block.
If you prefer to use the same translation tables that are used in Entire Net-Work:
in ASC2EBC and EBC2ASC, change the COPY statements from UES2ASC and UES2EBC to NW2ASC and NW2EBC, respectively.
re-assemble the translation tables and relink LNKUES/ADALNK.
Both the Adabas and Entire Net-Work translation table pairs are provided in the section Translation Tables. You may want to modify the translation tables or create your own translation table pair. Be sure to (re)assemble the translation tables and (re)link LNKUES/ADALNK.
//LINK EXEC PGM=IEWL,REGION=0M // PARM=’XREF,REUS,LIST,LET,NCAL,SIZE=(1024K,256K)’ //SYSPRINT DD SYSOUT=* //SYSLMOD DD DISP=SHR,DSN=USER.LOAD(ADALNK) //ADALIB DD DISP=SHR,DSN=ADABAS.LOAD //SYSLIN DD * MODE AMODE(31) RMODE(ANY) ENTRY ADABAS INCLUDE ADALIB(ADALNK) INCLUDE ADALIB(LNKUES) INCLUDE ADALIB(ASC2EBC) INCLUDE ADALIB(EBC2ASC) NAME ADALNK(R) /* //
The (re)linked ADALNK must be made available to Entire Net-Work. If you are calling Adabas version 7 and you do not have the correct LNKUES/ADALNK module, Adabas produces unexpected results: response code 022, 253, etc.
Perform the following steps to create the ADALNK module:
to create the ADALNK module:
assemble ADALNK and UEXITA/UEXITB
To assemble ADALNK, enter the following z/VM commands:
ACC cuu1 filemode/A ACC cuu2 S GLOBAL MACLIB ADAV7vv MVSXA HCPGPI DMSGPI DMSOM ASMAHL ADALNK —where cuu1 is the virtual unit address of the Adabas library minidisk cuu2 is the virtual unit address of the system minidisk containing the MVSXA MACLIB filemode is the filemode for the Adabas library minidisk
To assemble UEXITA/UEXITB, enter the same commands as above:
ACC cuu1 filemode/A ACC cuu2 S GLOBAL MACLIB ADAV7vv MVSXA HCPGPI DMSGPI DMSOM ASMAHL uexitname —where cuu1 is the virtual unit address of the Adabas library minidisk cuu2 is the virtual unit address of the system minidisk containing the MVSXA MACLIB filemode is the filemode for the Adabas library minidisk uexitname is the UEXITA/B filename (with a filetype of TEXT)
load ADALNK and the user exits, by entering the following z/VM commands:
LOAD ADALNK (RESET ADALNK RLD INCLUDE uexitname (RESET ADALNK RLD
—where uexitname is the filename (with filetype of TEXT) of UEXITA or UEXITB. To load both UEXITA and UEXITB, specify the INCLUDE command again for the second user exit.
generate a module, by entering the z/VM command:
GENMOD
ADAUSER is a program that links the user to Adabas. It is specific to an operating system and is independent of release level and mode.
ADAUSER operates in the following way:
ADAUSER is the only program that contains the entry point ADABAS. ADAUSER is dynamically loaded from the Adabas library, ADAVvvv TXTLIB, when a user program is loaded.
On the first call to Adabas, ADAUSER invokes ADARUN, which is installed as a SYSTEM nucleus extension at LOGON time by the NUCXTNTS EXEC, and is release-independent. Subsequent Adabas calls bypass ADARUN.
When invoked, ADARUN processes its control statements. If the ADARUN PROGRAM parameter has the (default) value USER, ADARUN loads ADALNK if the ADARUN MODE parameter specifies MULTI (the default), or ADANUC if the ADARUN MODE parameter specifies SINGLE. Therefore, ADAUSER is mode-independent.
