Version 8.2.5	SEARCH   INDEX   CONTENTS   |   PDF PAGE   PDF BOOKS   |   HOME   UP   PREV   NEXT

This document describes the Adabas 8.2 enhancements.

Support for Two-Gigabyte Large Pages in z/OS

Adabas 8.2 SP5 introduces support for two-gigabyte (2G) large pages in z/OS environments, if you have specific zaps installed. The required zaps that must be applied for this support are: AI825005, AI825007, AI825011, AN825014, AO825018, AO825019, and AU825009. In addition, z/VSE users will need to apply AD825006 when they apply AI825011 (even though z/VSE does not support large pages).

If your z/OS operating system supports 64-bit virtual storage and 2G large pages (page-fixed 2G pages of real storage), you can request that the Adabas nucleus use 2G large pages. The ADARUN LARGEPAGE parameter has been extended to allow you to indicate whether your Adabas nucleus should use 2G large pages. The default is not to use them.

Two new values for the LARGEPAGE parameter, "G64" and "L64", are introduced in release 8.2 SP5. A value of "L64" is the same as specifying the value "YES"; when either of these values is specified, the Adabas nucleus will attempt to acquire virtual storage above the 2G bar backed by page-fixed 1M large pages. When the "G64" value is specified, the Adabas nucleus will attempt to acquire virtual storage above the 2G bar backed by page-fixed 2G large pages. In all cases, the virtual storage will be backed by large pages only if the operating system supports them and if sufficient large pages are available in the system.

For more information, read LARGEPAGE Parameter: Large Page Support Option.

Adabas Caching Facility Two-Gigabyte Large Page Support

Updates have also been made for 2G support in cache space (for Adabas Caching Facility 8.2 SP3 and later releases):

• Two new ADARUN parameters, CASSOG64 and CDATAG64 have been added to support Associator and Data Storage RABNs in 2G large page cache space .

• The CFILE ADARUN parameter has been updated to include a new valid value "G64" or "G" (representing virtual 64-bit storage backed by 2G large pages) as the type of cache storage to be used for RABN caching.

• The CSTORAGE and CWORKSTORAGE ADARUN parameters have been updated to include a new valid value "G64" that supports 2G large page cache space.

Adabas Parallel Services Two-Gigabyte Large Page Support

If your z/OS operating system supports 64-bit virtual storage and 2G large pages, you can now request that the Adabas Parallel Services 8.2 SP3 (and later Adabas Parallel Services releases) use 2G large pages. ADARUN parameter CLUCACHETYPE has been extended to allow you to indicate whether your Adabas nucleus should use 2G large pages. The default is not to use them.

Adabas Parallel Services 8.2 SP3 introduces this support only if zap AS823003 has been applied to your Adabas Parallel Services 8.2 SP3 installation and only if the following Adabas 8.2 SP5 zaps have been applied for Adabas 8.2 SP5 databases: AI825005, AI825007, AI825011, AN825014, AO825018, and AU825009. In addition, z/VSE users will need to apply AD825006 when they apply AI825011 (even though z/VSE does not support large pages).

Updated LNKLNK8 and LNKLNKR8 Sample Jobs

In Adabas 8.2 SP5, the sample jobs LNKLNKR8 and LNKLNK8 in the ADAvrs.JOBS library have been changed in this release to make linking of ADALNK and ADALNKR jobs more flexible. If you have used these sample jobs in the past as models for your own jobs, you will need to modify your own jobs accordingly or they will not work.

This change involved changing the way in which the independent LNKIND module is linked with the dependent ADALNK8 and ADALNKR8 modules; ADALNK8 and ADALNKR8 are no longer distributed with LNKIND linked within them. Instead, the new LNKLNKR8 and LNKLNK8 jobs perform this linkage by including a specific INCLUDE control statement for the independent LNKIND module.

Note:
Effective with Adabas 8.2 SP5, if you use link routine user exits, Adabas Review, or Adabas System Coordinator, the jobs used to link these components with the batch, reentrant batch or the IMS link routine should be modified to explicitly include the LNKIND module when the link routines are relinked to incorporate user or Software AG product link routine exits.

Updated LNKLNI8 Job for Linking the IMS Adabas Link Routine

In Adabas 8.2 SP5, the sample job, LNKLNI8 in the AIIvrs.MVSSRCE library has been updated with an INCLUDE statement for LNKIND. With Adabas 8.2 SP5 and later releases, the ADALNI8 IMS-dependent Adabas link routine module is no longer distributed with the independent LNKIND module included. The new LNKLNI8 job should be used as a template when relinking the ADALNI link routine module for operation under IMS.

Note:
Effective with Adabas 8.2 SP5, if you use link routine user exits, Adabas Review, or Adabas System Coordinator, the jobs used to link these components with the batch, reentrant batch or the IMS link routine should be modified to explicitly include the LNKIND module when the link routines are relinked to incorporate user or Software AG product link routine exits.

Batch ADALNK and ADALNKR GETMAIN Change

The Adabas 8.2 SP5 batch link routines (ADALNK and ADALNKR) restore the GETMAIN prefetch processing to the method delivered in Adabas 8.2 SP2 and 8.2 SP3. Specifically, the link routine will determine the AMODE of the caller and obtain storage above or below the line accordingly. Thus if the caller is in AMODE=31 a GETMAIN request from the PREFETCH module will obtain storage above the 16MB line. If the caller is in AMODE=24, then the storage will be obtained below the line.

With this change the use of PREFETCH=OLD is discouraged, especially for run-units where the initial Adabas call is done in 31-bit mode but subsequent calls may be done in 24-bit mode. Such a calling sequence may lead to an addressing exception and abnormal termination of the job step. When PREFETCH=YES is employed, it is recommended that the ADAMLF module also be from the 8.2 SP5 library, or if for an earlier version of ADAMLF is used, all current maintenance should be applied to ensure that ADAMLF will always be in 31-bit mode before calling the batch link routine.

PREFSBL ADARUN Parameter Maximum Changed

The maximum value that can be specified for the PREFSBL ADARUN parameter has been changed in release 8.2 SP5. The new maximum is the setting of the PREFTBL ADARUN parameter. In addition, the calculation used to estimate the initial size of PREFSBL has been simplified.

For complete information about the PREFSBL parameter, read PREFSBL Parameter: Prefetch Buffer Length per Call.

Natural zIIP Extension Support

With version 8.2 SP5, Adabas provides support for Software AG's Natural zIIP extensions. For complete information, refer to your Natural documentation.

High Performance FICON for z/Series (zHPF) Support (z/OS Only)

With version 8.2 SP5, Adabas provides support for the use of high performance FICON for z/Series (zHPF) processors under z/OS. zHPF provides a new, streamlined channel protocol and channel program structure, designed to improve I/O performance.

For further details of the software and hardware requirements for using zHPF, review the zHPF documentation provided on the following Software AG web page: http://www.softwareag.com/Corporate/products/bis/platforms/ibm_availability.asp.

Extended Address Volume (EAV) Support

Adabas supports the use of extended address volumes (EAV) under z/OS. For a description of the Software AG requirements for this support, review the documentation at the following link: http://www.softwareag.com/corporate/images/Software_AG_ETS_Product_Support_for_Extended_Address_Volumes__EAV_07222010_tcm16-43365.pdf.

New BS2000 Exit Assembly Job

Release 8.2 SP4 provides a new BS2000 sample exit assembly job called ASMUEX. You can use this job to assemble and link all user exits in BS2000 environments.

Adabas Review Client Reporting Support in the Link Routines

A new ADARUN parameter, RVCLIENT, has been introduced in release 8.2 SP4. Use the RVCLIENT parameter to activate Adabas Review client reporting when you want to run client reports in batch environments. For more information, read RVCLIENT Parameter: Client Reporting Activation.

Note:
This ADARUN parameter is valid only in z/OS environments (Adabas Review client reporting is only supported in z/OS environments).

New Hyperdescriptor Exit Sample

With Adabas 8.2 SP4, a new hyperdescriptor exit source sample is supplied in the Adabas source library in member USERHX03. A sample job to assemble and link the hyperdescriptor exit is supplied in member ASMUHX03 of the sample job library in z/OS environments and in member ASMUHX03.X of the sample job library in z/VSE environments. In BS2000 environments, you can use the sample exit assembly job, ASMUEX, to assemble and link all user exits.

For complete information on hyperdescriptor exits, refer to the Adabas User, Hyperdescriptor, Collation Descriptor, and SMF Exits Manual.

ADAM97 Message Reports Server Termination

In prior releases, message ADAM97 was issued in z/OS environments at the beginning of database termination to remind users to restart the initiator that z/OS would stop. With release 8.2 SP4 (and later releases), this message has now been altered because z/OS now restarts the initiators automatically. If your installation uses automated operation scripts that searched for and took action when the original ADAM97 message was issued, those automated scripts may need to be changed.

The new ADAM97 message is an informational message that is issued slightly later during the termination process than the original message. It indicates only that the database is terminating and that commands to that database are no longer accepted.

For complete information on error messages, refer to the Adabas Messages and Codes Manual.

Support for Adabas Review 4.6

Adabas 8.2 SP4 provides support for Adabas Review 4.6. Older versions of Adabas do not support Adabas Review 4.6.

The Adabas link routines for z/OS have been updated to support new features provided by Adabas Review 4.6. The batch/TSO, CICS, Com-plete and IMS links have been updated.

This support includes the addition of the following new keywords for the Adabas link globals table (LGBLSET) macro:

REVHID requires REVIEW=YES

REVHID=hubid

REVREL={45 | 46}

RVCLNT=YES requires REVIEW=YES

RVCLNT={YES|NO}

L3/L6 Commands: Response Code 113 (ADARSP0113) Causes Message ADAN75

Effective with release 8.2 SP4, message ADAN75 is issued whenever an L3 or L6 command gets response code 113 (invalid ISN -- ADARSP0113). The user receiving the response code can be identified from the last eight bytes of the communication ID. For a file loaded with a nonzero MINISN, the ADAN75 message now reports the external ISN (starting at MINISN).

For complete information on error messages, refer to the Adabas Messages and Codes Manual.

Enhanced Error Recovery

More robust error recovery has been implemented in this release, to safeguard against the possibility of users hanging following a nucleus failure. The new routine receives control and issues new message ADAS36 if normal error recovery doesn't complete. It will attempt to post all incomplete commands.

For complete information on error messages, refer to the Adabas Messages and Codes Manual.

ADADBS OPERCOM RALOCKF Behavior Altered

The behavior of the ADADBS OPERCOM RALOCKF function has changed with this release. If an ADADBS OPERCOM RALOCKF function is issued against a file that is not advance-locked, ADADBS now finishes with response code 0 after printing an informational message.

In Adabas Version 8.2 SP3, the following sequence of utilities resulted in user abend 034 (ABENDU034) for the ADADBS OPERCOM RALOCKF function, and ADADBS utility message ERROR-72 was produced. In Adabas versions prior to 8.2 SP3, the following sequence of utilities resulted in no error for the ADADBS OPERCOM RALOCKF function, which completed with condition code 0.

ADADBS OPERCOM ALOCKF=ff
ADAULD UNLOAD FILE=ff 
ADADBS OPERCOM RALOCKF=ff

This 8.2 SP4 enhancement allows the utility to largely return to the behavior in versions prior to Adabas 8.2 SP3. With Adabas 8.2 SP4, the ADADBS OPERCOM RALOCKF function will terminate with response code 0 when it is issued against a file that is not advance-locked, but an informational message that reads "File nnnnn was not advance-locked." (where nnnnn is the file number) is also produced.

License Refresh and Verification Support

You can now refresh the license file currently used by the nucleus. This allows you to overwrite the current license with a new one or force the system to reread and reload its existing license. To support this functionality, Adabas 8.2 SP3 introduces a new operator command: LICREFRESH. For complete information, read LICREFRESH Command.

Improved ADASAV Volume Information on z/OS

Adabas 8.2 SP3 includes an enhancement to ADASAV SAVE that provides more accurate volume information in z/OS environments. In previous releases, due to the asynchronous nature of ADASAV I/O processing, a number of blocks written on one volume might have been reported by ADASAV as written on the preceding volume.

Important:
To enable this improved ADASAV volume information, apply optional Adabas zap AY823135 and IBM APAR OA33465. If you choose to apply AY823135, make sure you also apply IBM APAR OA33465. Installing AY823135 without IBM APAR OA33465 can cause the ADASAV SAVE backup to become corrupted. An ADASAV RESTORE would then fail (with system abend S002) and the database might then be unusable.

Session Autorestart Progress Messages

While performing a session autorestart after a nucleus failure, Adabas 8.2 SP3 now prints progress messages that indicate how much repair and recovery work has already been done and how much remains to be done. By default, these progress messages occur at the end of each phase of autorestart processing (backward repair, forward repair, autobackout). A new ADARUN parameter, INTAUTO, has been provided that allows you to specify the time interval (in seconds) between autorestart progress messages, thus allowing you to request that additional progress messages be printed for each phase of the autorestart process.

For complete information about the INTAUTO ADARUN parameter, read INTAUTO Parameter: Time Interval Between Autorestart Progress Messages.

New Sample COBOL Programs

Adabas 8.2 SP3 includes two new sample COBOL programs, COBADA8 and COBACI8. These programs are provided to illustrate how to make Adabas ACBX direct calls in a batch/TSO environment (COBADA8) and under CICS/TS (COBACI8). Their source modules should be compiled using the Enterprise COBOL compiler Version 3 Release 3 or later.

The COBADA8 member is provided in the ADAvrs.MVSSRCE library and the COBACI8 member is provided in the ACIvrs.MVSSRCE library.

Note:
These programs are provided "as is" and will not be supported by Software AG.

For more information about these sample programs, read ACBX Examples.

Unique Descriptor Pool Minimum Size Increased

Starting with Adabas 8.2 SP3, the minimum setting of the unique descriptor pool will be increased from zero to 5000 bytes. This size is established by the ADARUN LDEUQP parameter.

Note:
This new functionality is also available in Adabas 8.2 SP2 if you apply zap AN822111.

For complete information, read about the LDEUQP parameter in LDEUQP Parameter: Length of Unique Descriptor Pool .

SMGT Response Code Handling

The following changes, made in Adabas 8.2 SP2, affect how SMGT response codes are handled:

• PINRSP will now be loaded automatically at startup if the ADARUN SMGT parameter is set to "YES", similarly to ADAMXY.

• The output from the PIN routines can be directed to DDTRACE1 rather than DDPRINT if ADASMXIT has been updated by the user to specify this and if DDTRACE1 is provided in the nucleus startup JCL.

To activate this support, be sure to apply fix AN822020 and use the ADASMXIT from the ADA822.S001 data set.

Note:
If you are using Adabas 8.2 SP3 or a later version of Adabas, applying this fix is unnecessary as the fix has already been applied to the source for you.

Information about this processing and DDTRACE1 are provided for z/OS sessions in Adabas Session Execution and in Installing The Error Handling And Message Buffering Feature.

ADALNK Access to Multiple ID Tables on BS2000

In prior versions of Adabas, routing of data from a client to a database on BS2000 systems could only take place over a single ID table defined by the IDTNAME parameter in the following ways:

• As a parameter in the file attached to the link name DDCARD (ADARUN statements);

• As a parameter in the file attached to the link name DDLNKPAR (ADALNK statements); or

• By modifying the parameter ENVNAME in the macro B2CONFIG and assembling the SSFB2C parameter repository module.

In Version 8.2 SP2, it is now possible to route data to specific database IDs to other ID tables, and optionally, to change the database ID to another in the target ID table.

System Field Support

Support is added in Adabas 8.2 SP2 for system fields. A system field is a field in an Adabas file whose value is automatically set by the Adabas nucleus when records are inserted or updated on the file. Optionally, you can specify that some system field values only be set when records are inserted. The following types of system fields are currently supported:

• Job name: The job name of the user inserting or updating a record.

• ETID: The eight-byte user ID of the user inserting or updating a record. This is the user ID set in the Additions 1 field of an OP (open) command for the user session.

• Session ID: The 28-byte user ID of the user inserting or updating a record.

• Session user: The last eight bytes of the 28-byte session ID or the user inserting or updating a record.

• Time: The date or date and time at which a record is inserted or updated.

Two new field options have been added to this release to be used when defining your system fields with the ADACMP COMPRESS utility function: SY and CR. The SY option should be used to identify the contents that should be stored in the system field; the CR option should be used to indicate that the system field value should only be set when the record is inserted into the Adabas file (it will not be modified when an update operation occurs). For more information about the SY and CR field options, read Field Options.

In addition, the ADALOD LOAD utility function can now specify the maximum number of occurrences for MU system fields allowed in a file using a new SYFMAXUV parameter. For more information, read LOAD: Load a File .

Finally, a new SYFMAXUV parameter is supported by the ADADBS MODFCB utility function that allows you to modify the maximum number of occurrences for MU system fields allowed in a file. For more information, read MODFCB: Modify File Parameters .

For complete information about system fields, read System Fields.

Partial LOB Field Processing

Adabas 8.2 SP2 introduces new partial large object field (LOB field) processing. Your application programs can now read or write parts of LOB values, or LOB segments. In addition, in a series or read or write requests for a LOB field, your applications can control the position of the next LOB segment within the LOB value. This support introduces the following new features:

• New LOB segment notation is provided for you to use in the format buffer of a command to select the specific portion (segment) of the LOB value you want to read or write in the call. For more information about this notation, read Field Selection Criteria .

• A new command option, L, can be specified in Command Option 2 of ACB or ACBX calls for L1/L4 or A1 commands. This command option requests that the current position in the LOB value be tracked in the ISN Lower Limit (ISL) field of the ACBX. In conjunction with this, a new current positioning notation (an asterisk) can now be used in the LOB segment notation of a call to indicate that the call should start at the current position (the position where the last call left off) of the LOB value.

  These features allow you to read or write the LOB value in segments using multiple calls. The current positioning notation (an asterisk) allows you to easily indicate where the next call should start, without changing the format buffer for every call.

  For more information, read A1 Command: Update Record and L1 and L4 Commands: Read / Read and Hold Record .

For complete information about processing partial LOB fields and all of these new features, read Processing LOB Segments.

Shared Hold Status

Adabas 8.2 SP2 introduces a new kind of competitive updating control called shared hold status. You can use shared hold status to lock data records in shared mode, rather than in the existing exclusive hold status. This allows your database users to read the same record in parallel transactions, but ensures that no one can update the record concurrently.

Using shared hold status, your users can protect large object values from concurrent updates without locking out other users who may need to read the same LOB value or other LOB values in the same record. It also allows your users to protect the records they read against concurrent updates for specific periods of time:

• For the duration of the read command;

• When the next record in a sequence is read;

• When the user's transaction ends;

• Indefinitely.

This new functionality introduces four new command options for the BT, ET, HI, L4 - L6, RI, and S4 commands, although all four new command options do not necessarily apply to all of these commands. These new options are specified in the command option 3 field of ACBX only calls for the associated command. Each command option specifies a different shared hold time period.

• Option C puts the record in shared hold status for the duration of the read operation. It ensures that the version of the record being read has been committed by the last updater. This option is available for the L4, L5, L6 and S4 commands. For the S4 command, the shared hold remains in place for the duration of the read operation.

• Option Q puts the record in shared hold status until the next record in the read sequence is read or the read sequence or transaction is terminated, whichever happens first. It ensures that the record being read cannot be updated concurrently until the next record in the sequence is read (or the transaction is terminated). This option is available for the L4 (when command option 2 is set to "N"), L5, L6 and S4 commands.

• Option S puts the record in shared hold status until the end of the transaction. It ensures that the record being read cannot be updated concurrently until the transaction is terminated. This option is available for the HI, L4, L5, L6, RI and S4 commands.

• Option H keeps a record in shared hold status indefinitely (until the next ET or BT command). This option is available for the BT and ET commands. Records in shared hold status at the time of the BT or ET command are kept in shared hold status beyond the end of the transaction until another ET or BT command is issued (without this H option or the prefetch or multifetch options). Any records in exclusive control are also changed to shared hold status beyond the end of the transaction.

If the same record is placed in shared hold status more than once (using the C or S options or the Q option for different read sequences), it stays in shared hold status until all of the specified hold lifetimes have expired.

For more information about these command options and their detailed functioning in each command, read about the individual commands in Commands.

Resource deadlock may occur when users are simultaneously holding records and waiting for records held by other users. Normally, Adabas detects a user deadlock situation and returns a response code to the user who would create the deadlock if placed in wait status. However, the deadlock may not be detected if the users involved are serviced by different nuclei in a cluster. Even in a single nucleus, a deadlock may not be detected if a record involved is being held as a shared resource by more than one user. If a deadlock is not detected, all users involved will wait until the first one of them exceeds the transaction timeout (TT) threshold.

Putting more records in hold status (including shared hold status) may decrease the possible parallelism of transaction processing (and thus, the performance of multiuser applications) and increase the likelihood of deadlocks between transactions (where each transaction holds a record that the other wants).

Using shared hold status affects the ADARUN NH and NISNHQ parameter settings. Each shared hold request with a different command ID (CID), as well as a (shared or exclusive) hold request without a CID, is counted against the NISNHQ and NH limits. This affects application programs that make use of the new Q option for sequential reads. For example, if a program reads records with the Q option and then updates every record, the shared hold operations from the use of the Q option and the exclusive hold operations from the update commands are counted separately. Such a program might need NISNHQ and NH limits set twice as large as when the Q option is not specified.

For complete information about shared hold status, read Shared Hold Status.

SMF Support (z/OS only)

For z/OS environments only, support for IBM System Management Facility (SMF) records has been added to Adabas 8.2 SP2. SMF recording is widely used in IBM and ISV software products. Adabas provides support for SMF records in two ways:

1. Adabas nuclei now generate SMF records to record performance and resource consumption. Existing Adabas statistics are the basis for standard Adabas records and a new user exit permits you to include your own data.

2. Adabas nuclei can now be registered with SMF to record CPU consumption in IBM-defined type 89 records.

Seven new ADARUN parameters are introduced as a result of this support:

• The SMF parameter can be used to enable or disable Adabas SMF recording.

• The SMF89 parameter controls whether the Adabas nucleus registers with z/OS for type 89 SMF records.

• The SMFRECNO parameter can be used to define the Adabas SMF record number for user-defined records.

• The SMFINTERVAL parameter can be used to specify whether interval SMF records should be generated and, if so, how often.

• The SMFSUBSYS parameter can be used to associate the SMF records with an IBM or user-defined subsystem.

• The SMFDETAIL parameter can be used to specify the detail sections of Interval and Termination records that should be included (if any).

• The UEXSMF parameter can be used to identify the name of the SMF user exit module that should be used.

For complete information about these ADARUN parameters, read ADARUN Parameter Directory .

This enhancement also introduces six new operator commands: SMFRECNO, SMFINTERVAL, SMFSUBSYS, SMFDETAIL, SMFDETAILADD, and SMFDETAILDEL. These commands can be used to alter the comparably named ADARUN parameter settings in a running Adabas session. The SMFDETAILADD and SMFDETAILDEL commands allow you to add and remove specific detail sections of Interval and Termination records from the list of records that should be recorded. These commands can be issued only as operator commands; they cannot be issued in an ADADBS OPERCOM utility function run. For complete information about these commands, read Operator Commands.

Finally, this enhancement allows you to provide a user exit that adds a user-specified detail section to the Adabas SMF record. It is a separate load module whose name is provided in the UEXSMF ADARUN parameter. For complete information, read SMF User Exit.

For complete information about Adabas SMF records, read Adabas SMF Records.

Command and Protection Log Alert Messages

Adabas 8.2 SP2 introduces the ability to optionally have Adabas generate alert messages indicating the status of the command log (CLOG) and the protection log (PLOG). The alert messages are written when Adabas detects that a PLOG or CLOG data set has not been copied within a specified time interval. Thereafter, an alert message is written at regular intervals as long as there is a PLOG or CLOG data set waiting to be copied. These alert messages are issued until the PLOG or CLOG data set is copied or until no logs remain to be copied. Your user exits 2 or 12 can be set up so that the messages appear providing a comprehensive trace of activity so you can resolve and problems with the PLOG or CLOG.

This feature helps you prevent the potential problem of Adabas overwriting an older CLOG or PLOG data set without waiting for direction from the user. Ordinarily, when a PLOG or CLOG data set fills up, the Adabas nucleus examines the return code from user exit 2 or user exit 12 to determine whether it should wait before overwriting an older PLOG or CLOG data set with new information. If there is no user exit 2 or 12, Adabas does not wait, but simply overwrites the older log. In prior releases of Adabas, no alert messages were issued before this happened. This new feature provides you with these alert messages. If this LOGWARN parameter is set to a value other than zero, the statuses of the CLOG and PLOG data sets are analyzed at the frequency specified by the parameter, warning about their status and providing you with the information you need to prevent any overwriting of older logs.

The production of CLOG and PLOG alert messages is optional and is determined by the setting of a new LOGWARN ADARUN parameter. A value of zero (0), the default, indicates that no CLOG and PLOG status checks should occur and that no alert messages should be produced. Any other valid LOGWARN value represents the interval (in seconds) at which the PLOG and CLOG status is checked and resulting alert messages are produced. When a value other than zero (0) is specified for LOGWARN, a valid user exit 2 or user exit 12 must also be specified. For complete information about the LOGWARN parameter, read LOGWARN : CLOG and PLOG Alert Message Interval.

In addition, a new LOGWARN command has been added that allows you to alter the setting of the LOGWARN ADARUN parameter for a given Adabas session. This command can be run from the operator console or via the ADADBS OPERCOM utility function. For complete information, read LOGWARN Command.

Printing Multifetch Statistics to a File

Effective with Adabas 8.2 SP2, multifetch statistics can be printed when a CL command is issued if both of the following conditions are met:

• The new PREFSTDD ADARUN parameter is specified in the startup JCL. This new parameter identifies the job control statement label (DD name) that defines the location to which the multifetch statistics should be written.

• The nucleus startup JCL includes a valid job control statement with the label name identified by the PREFSTDD ADARUN parameter.

For complete information, read PREFSTDD: Multifetch Statistics DD Name.

LORECX DSECT Updates

In Adabas 8.2 SP2, the fields LOX1STYP and LORSTYPE in the LORECX DSECT (distributed in the Adabas source library) have been altered so that the CLOGs reflect accurate information describing which search algorithm was used.

Logically Deleting and Undeleting Descriptors

Adabas 8.2 SP2 allows you to logically delete and undelete a descriptor from a database file. Logically deleting a descriptor from a file removes the descriptor from the file, but retains the field data in the database. A logically deleted descriptor cannot be used as a search descriptor. This functionality is provided using two new ADADBS utility functions: DELDE (logically delete a descriptor) and UNDELDE (undelete a logically deleted descriptor). For more information, read DELDE: Logically Deleting a Descriptor and UNDELDE: Undeleting a Logically Deleted Descriptor.

Logically Deleting and Undeleting File Fields

Adabas 8.2 SP1 allows you to logically delete and undelete a field from a database file. Logically deleting a field from a file removes the field from the FDT, but retains the field data in the database. Logically undeleting a field that was previously logically deleted reinstates the field in the FDT. This functionality is provided using two new ADADBS utility functions: DELFN (delete a field) and UNDELFN (undelete a field). For more information, read DELFN: Logically Delete Files and UNDELFN: Logically Undelete Files.

Maximum Length of Fixed Point Fields Increased

The maximum length of fixed point fields (F format fields) was increased to eight bytes in Adabas 8.2 SP2. Fields defined with format F can now be two, four, or eight bytes long.

For more information, read Field Length and Data Format.

New Date-Time Edit Masks

New date-time edits masks are introduced in Adabas 8.2 SP1. For a complete reference of these date-time edit masks, read Date-Time Edit Mask Reference.

• A new DT field option is introduced that allows you to specify the edit mask for a field definition. For more information, read Date-Time Edit Mask Option DT.

  This DT option is supported by ADACMP utility FNDEF field definition statements. For more information, read DT: Date-Time Edit Mask Field Option.

• Date-time edit masks are also supported in the format and search buffers of direct calls. You can add date-time edit mask to fields with formats of B, F, P, or U in a loaded file. For more information, read Date-Time Edit Mask Processing in Format and Search Buffers.

Date-Time Edit Mask Enhancements

The following date-time edit mask enhancements have been made in Adabas 8.2 SP2:

• The new DT option, introduced in 8.2 SP1, is now supported by the ADADBS NEWFIELD function. In ADADBS NEWFIELD functions, the DT option is supported in FNDEF statements, allowing you to specify a date-time edit mask for a new field.

  For more information, read NEWFIELD: Add New Field .

• Changes have been made to the ADADBS CHANGE utility function so you can now change the time zone and the date-time edit mask. For complete information, read CHANGE: Change Standard Length or Format of a Field .

• Changes have been made to the ADADBS ONLINVERT utility function so you can now define a date-time field ( with edit mask and time zone) as a descriptor field. For complete information, read ONLINVERT: Start Online Invert Process .

Time Zone Support

Adabas 8.2 SP2 supports variations in time zones. The following updates have been made:

• A new library, the Adabas time zone library, is provided in your installation.

  A new TZINFO member is provided in your Adabas source library. It lists the time zone support in Adabas and references members of the new time zone library containing the actual code used for time zone conversions. For more information about the Adabas time zone library, refer to the Adabas installation information for the appropriate system platform. For information about the TZINFO library member, read Supported Time Zones.

  Adabas uses the time zone data taken from the public-domain tz database, (also called the zoneinfo or Olson database) to create its own time zone library.

• A new TZ parameter can now be specified in the ADACMP COMPRESS and DECOMPRESS utility functions to specify the local time zone. As records are compressed and loaded into the file, the date-time data is converted from the specified time zone to UTC time (Coordinated Universal Time, also known as Greenwich Mean Time); likewise as records are decompressed and read from a file, the date-time data is converted from UTC time to the specified time zone. Date-time field data is always stored in UTC time in an Adabas file. For more information, read COMPRESS: Compress an Adabas File and DECOMPRESS: Decompress an Adabas File .

  In addition, review important information about the daylight savings time enhancements provided in this release. If a time zone uses daylight savings time, you must be sure to store and retrieve the daylight savings indicator with your date-time data or there will be no way to distinguish date-time values in the hour before the time is switched back to standard time.

• A new TZ (time zone) record buffer option is available for the OP command. You can use this option to set the time zone of an individual user session. If no time zone is specified for a user session, response 55 (ADARSP055) subcode 30 is issued when a field with the TZ option is accessed or updated. For more information about using the TZ record buffer option in an OP command, read about the OP command Record Buffer.

• A time zone field option, TZ, can now be specified to identify date-time fields that should be presented in the user's local time and stored in UTC time, thus allowing for differences in time zones. For more information, read Time Zone Option TZ.

• Changes have been made to the ADADBS CHANGE utility so you can now change the time zone and the date-time edit mask for a field. For complete information, read CHANGE: Change Standard Length or Format of a Field .

• Changes have been made to the ADADBS NEWFIELD utility so you can now specify the time zone for a new field. For complete information, read NEWFIELD: Add New Field .

• Changes have been made to the ADADBS ONLINVERT utility function so you can now define a date-time field ( with edit mask and time zone) as a descriptor field. For complete information, read ONLINVERT: Start Online Invert Process .

Note:
The default number of time zones that can be used in a nucleus session is limited to 8 and the time zone pool size default to 32K. However, Software AG allows you to increase these numbers (refer to the appropriate zap in member ZAPOPT of the distributed source library) so that all time zones can be loaded.

Daylight Savings Time Support

Adabas 8.2 SP2 provides daylight savings time support. If the current time zone uses daylight savings time, Adabas provides the following support:

• A new DST parameter can now be specified in the ADACMP COMPRESS and DECOMPRESS utility functions to indicate that date-time data includes a daylight savings time indicator. If a time zone uses daylight savings time, you must be sure to store and retrieve the daylight savings indicator with your date-time data or there will be no way to distinguish date-time values in the hour before the time is switched back to standard time.

  The DST parameter requires that the TZ parameter be set in the same ADACMP run.

  For more information, read COMPRESS: Compress an Adabas File and DECOMPRESS: Decompress an Adabas File .

• A new daylight savings time indicator, D, can be specified in format and search buffers to indicate that a daylight savings adjustment is required for a date-time field with the TZ option specified. For complete information, read Daylight Savings Indicator (D).

RI Commands: Change Possibly Affecting Your Application Programs

Adabas 8.2 SP2 on mainframes no longer releases records unconditionally. If your application issues an RI command for a record that has been updated in the current transaction, Adabas will now return response code 113 (ADARSP113); if ISN=0 was specified, response code 2 (ADARSP002) is returned. This change in behavior may affect how your application programs are coded.

This change was necessary to accommodate Adabas support for shared hold status and partial large object (LOB) field processing. It prevents problems that might occur if updated records are released from hold, the transaction is subsequently backed out, and there are competitive updates to the same records. Also, this change ensures that Adabas on mainframe and open systems behave in the same manner regarding RI commands.

Application programs that issue RI commands may be affected, if they perform updates and if they try to release updated records from hold status. Such application programs should be adjusted to either not issue RI commands for updated records or suppress the resulting response code 113 (ADARSP113). In addition, if your installation has Natural installed, you must apply Natural zap NA76045 for downward compatibility, which is available using Empower.

For complete information about the RI command, read RI Command: Release Record .

Terminology and Name Changes

The following terminology and name changes have occurred in Adabas 8.2 SP2:

• UEXITB, also known as user exit B (the before or pre-command user exit), has been renamed LUEXIT1 or link user exit 1 in z/OS environments.

• UEXITA, also known as user exit A (the after or post-command user exit), has been renamed LUEXIT2 or link user exit 2 in z/OSenvironments.

Lowercase Field Name Enhancement and FDT Changes

Adabas 8.2 SP1 introduces the ability to create field names that contain lowercase alphabetic characters. With this change, the maximum number of field definitions for a file has increased from 926 to 3214. Thus, fields WM, wm, wM, and Wm are now all valid field names and now represent four unique fields.

Note:
Lowercase fields will not display correctly (they will be converted to uppercase) if you use the ADARUN parameter settings MSGCONSL=UPPER, MSGDRUCK=UPPER, or MSGPRINT=UPPER.

This support has also introduced changes in the FDT limitations. In past releases of Adabas, the FDT was stored in up to four Associator blocks (with a maximum possible record buffer length set to the length of four Associator blocks). With Adabas 8.2 SP1, this FDT limitation has been dropped.

Expanded File Chain Enhancements

With Adabas 8.2 SP1 you can dynamically insert and remove files from an expanded file change using a new ADADBS function, EXPFILE. For more information on this utility function, read EXPFILE: Insert or Remove Files in an Expanded File Chain.

Putting Files in Read-Only State

Adabas 8.2 SP1 allows you place specific database files in read-only state while you maintain them. Your database users will be able to read data from the read-only files while your system administrators perform maintenance on those files. In addition, other files in the database, not in read-only status, can be read and updated as normal.

This support is provided in the form of the following Adabas utility enhancements:

• A new READONLY parameter has been added to the ADASAV RESTONL FILES, ADASAV RESTONL FILES, ADASAV RESTONL GCB, ADASAV RESTORE FILES, ADASAV RESTORE FMOVE, ADASAV RESTORE GCB,ADAORD STORE, and ADALOD LOAD utility functions. READONLY can be used to change the read-only status of a database file.

• A new FILEREADONLY parameter has been added to the ADADBS MODFCB utility function. FILEREADONLY can be used to indicate whether an Adabas database file should be placed in readonly status (where it can be updated only by Adabas utilities) or in normal status (where it can be updated in any normal manner).

  Notes:

  1. If the FILEREADONLY parameter is specified, only the FILE parameter (which is required) and the optional PASSWORD and TEST parameters can be specified in the same ADADBS MODFCB run.
  2. The ADADBS DSREUSE, ISNREUSE, NEWFIELD, RELEASE DE, and RENAME utility functions do not need to lock the file for their use; these functions can perform their processing in parallel with active users. This means that you do not need to set a file in read-only status (FILEREADONLY=YES) if you will be performing these other ADADBS functions only.

For complete information, review the descriptions of the READONLY and the FILEREADONLY parameter as they are used by each these utility functions in Utilities.

Database Conversion Improvements

You now no longer need to run multiple ADACNV utility runs to convert or revert an Adabas database from one version to another. Adabas 8.2 SP1 includes an updated ADACNV utility that will convert or revert a database in a single step.

For complete information, read ADACNV Utility: Database Conversion .

Unique Descriptor (UQ) Option Removal

Adabas 8.2 SP1 introduces the ability to remove the UQ (unique descriptor) option from a descriptor field definition using the ADADBS CHANGE utility function. A new option, NOUQ, is introduced for the OPTION parameter. When you run the utility specifying OPTION=NOUQ for a descriptor field, the UQ setting is removed from the descriptor field.

For general information about unique descriptor fields, read Descriptor Options DE, UQ, and XI . For specific information about the new NOUQ option in the ADADBS CHANGE utility function, read CHANGE: Change Standard Length or Format of a Field.

Work Area Monitoring

Adabas 8.2 SP1 allows you to monitor the use of parts 1, 2, and 3 of the Work data set. This information allows you to proactively adjust the sizes of these areas and ensure 24-by-7 access to your databases. It also allows you to reduce the risk of receiving response code 9 (ADARSP009), subcode 15 errors while using your Adabas database.

This support is available using the DRES operator command. This support is also available via the ADADBS OPERCOM function. The information is also provided in Adabas Online System and Adabas Manager screens.

Note:
The maximum pool value of Work part 1 is derived from the LP parameter. It corresponds to the maximum number of blocks a transaction can spend on Work Part 1 before Adabas decides to back it out.

For more information about this new information, read DRES Command.

PLOG and CLOG Data Set Enhancements

Adabas 8.2 SP1 introduces several enhancements to its handling of protection log (PLOG) and command log (CLOG) data sets:

• A new NPLOGBUFFERS ADARUN parameter has been introduced that you can use to increase the number of PLOG protection I/O buffers, allowing multiple PLOG blocks to be written in a single write I/O to the PLOG and improving the performance of applications that were formerly constrained by writes to the PLOG. For more information, read NPLOGBUFFERS : PLOG Protection I/O Buffer Control .

• The ADARES CLOG and ADARES PLOG utility functions have been expanded to check the PPT at the end of the CLOG or PLOG run to determine whether any additional CLOG or PLOG data sets need to be copied. If so, the ADARES utility invokes user exit 2 or 12, as appropriate, for each uncopied CLOG or PLOG data set it detects. For complete information, read CLCOPY: Copy Dual Command Log and PLCOPY: Copy Protection Log to Sequential Data Set.

• The ADADBS utility introduces four new functions (ADDCLOG, ADDPLOG, DELCLOG, and DELPLOG) that you can use to dynamically add and delete CLOG and PLOG data sets without terminating your current nucleus session. Running these utility functions is invalid when Adabas is running with dual CLOGs or PLOGs.

  For complete information, read ADADBS Utility: Database Services .

• You can now log the extended I/O list to the CLOG data sets, for both CLOGLAYOUT=5 and CLOGLAYOUT=8. A new ADARUN parameter, LOGVOLIO, is provided so you can specify this functionality for an Adabas nucleus.

  You can also use two new operator commands to dynamically control this logging.

  1. Use the new LOGVOLIO operator command on the console (or in an ADADBS OPERCOM utility function run) to dynamically start logging of the extended I/O list. For more information, read LOGVOLIO Command .

  2. Use the new NOLOGVOLIO operator command on the console (or in an ADADBS OPERCOM utility function run) to dynamically stop logging the extended I/O list. For more information, read NOLOGVOLIO Command.

• A new ADADBS utility function, REACTLOG, now allows you to reactivate command logging in an active nucleus where it had been disabled previously as a result of an I/O error. The cause of the I/O error needs to be corrected before running this utility function or command logging will simply fail again and will not be reactivated.

  For complete information, read REACTLOG: Reactivating Command Logging.

• A new SRLOG ADARUN parameter allows you to indicate how spanned records should be logged to the protection logs. Entire records (ALL), modified segments of spanned records (UPD), or entire records up to and including the modified segments (PART) can be logged. The default is UPD. For more information, read SRLOG : Controlling Spanned Record Logging .

Spanned Record Logging

Adabas 8.2 SP1 allows you to write complete spanned records to the PLOGs. In the past, spanned records were only partly written to the PLOGs, making them impossible to decompress and use.

• A new SRLOG ADARUN parameter allows you to indicate how spanned records should be logged to the protection logs. Entire records (ALL), modified segments of spanned records (UPD), or entire records up to and including the modified segments (PART) can be logged. The default is UPD. For more information, read SRLOG : Controlling Spanned Record Logging .

  The SRLOG parameter replaces the PLOGSEGSDS ADARUN parameter; the PLOGSEGSDS parameter has been dropped in this release.

• The ADASEL utility has been enhanced with three new SET GLOBALS parameters to support the decompression of spanned records from the PLOG: SPANREC, MAXLOGRECLEN, and LWP. Each of these parameters is described in SET GLOBALS Parameter.

  In addition, a SPANREC parameter has been added to the ADASEL OUTPUT instruction and can be used in an ADASEL OUTPUT instruction to indicate that new spanned record SELH and SELC output headers should be used for all EXPAn output if there is the possibility that the output decompressed spanned records will exceed the physical record length limitation. DSECTs for the SELH and SELC headers can be found in the Adabas source library. For complete information, read OUTPUT Instruction.

• The ADACDC utility has been enhanced with two new parameters: SPANREC and MAXLOGRECLEN. Each of these parameters is described in Running the Utility .

  The SPANREC parameter indicates that new spanned record CDCH and CDCN output headers should be used for all CDCOUT output if there is the possibility that the output decompressed spanned records will exceed the physical record length limitation. DSECTs for the CDCH and CDCN headers can be found in the Adabas source library.

New Adabas Event Log Support

Adabas 8.2 SP1 introduces a new Adabas event log that can be used to log database events. In this release, only response code 145 (ADARSP145) events are logged to the Adabas event log. Each entry in the log is currently 128 bytes long, although this may change in future releases.

The size of the Adabas event log for a database is controlled by the INFOBUFFERSIZE ADARUN parameter. In addition, the entries in the Adabas event log can be displayed using a new ADADBS DEVENTLOG utility function.

Response Code 145 (ADARSP145 - Hold Queue Entry) Event Enhancements

Several enhancements have been made in Adabas 8.2 SP1 to support response code 145 (ADARSP145) events. Response code 145 (ADARSP145) occurs when a hold queue entry is not available for an N1or N2 command or when a command is issued that attempts to put a record on hold but cannot because the record is held by some other user. In this release of Adabas, events triggering response code 145 (ADARSP145) are now logged in the:

• Adabas event log, if an Adabas event log has been allocated. The Adabas event log size and allocation is controlled by the INFOBUFFERSIZE ADARUN parameter for a database.

• command log, if CLOGLAYOUT=8 is specified. A new record type (x'000D') is now produced in a CLOGLAYOUT=8 command log to accommodate response code 145 (ADARSP145) events.

In addition, a special L2 command call can now be issued if you receive a response code 145(ADARSP145). When you execute the L2 command with ISN=0, FNR=-4(X'FFFC'), and an appropriate format buffer, you can obtain information about the user currently holding the ISN when the response code was issued. For complete information on this call, read Special L2 Command Call for Response 145.

Multiple CICS Task-Related User Exit (TRUE) Support

Adabas 8.2 SP1 supports the installation of multiple CICS task-related user exits (TRUEs) and Adabas application stubs from a single execution of the ADACIC0 installation program. Multiple TRUEs allow your site to tailor different Adabas CICS execution options in the same CICS region with a centralized installation procedure and software. For complete information, read Installing Adabas with CICS under Adabas 8, in the Adabas z/OS Installation Guide.

CICS Installation Changes

Additional CICS installation changes have occurred in Adabas 8.2 SP1. We no longer deliver ADACIRQ with Adabas on z/OS and z/VSE systems. The CICS input transient data queue is no longer required and the DDLINK input is no longer required. These have been replaced by the Adabas CICS installation options table (member ACIOPT) and the Adabas CICS names module (member ACINAMES). For more information, read Installing Adabas with CICS under Adabas 8, in the Adabas z/OS Installation Guide.

Performance Enhancements

Various updates have been made in Adabas 8.2 SP1 to ensure it continues to operate at peak performance:

• The performance (CPU consumption) of opening new Adabas user sessions, when there are already thousands or tens of thousands of concurrent users, has been improved.

• The performance of large object (LOB) updating has been improved by eliminating redundant serialization. All nuclei now reduce their use of the unique descriptor pool for LOB updates and customers using LOBs from Adabas 8.1 may be able to reduce the unique descriptor pool size (ADARUN parameter LDEUQP). Adabas Cluster Services and Adabas Parallel Services configurations no longer acquire external unique descriptor locks for LOB updates.

• Design changes have been made to the nucleus so that the blocks of a new address converter (AC) extent are now formatted asynchronously rather than synchronously. This allows a database to continue processing without waiting for all of the new AC extent blocks to be formatted.

• The new V64BIT ADARUN parameter may be used to improve the performance of your Adabas nucleus sessions. V64BIT allows the nucleus to use virtual storage above the 2 gigabyte bar, freeing virtual storage below the bar for other uses.

• For systems running in 64-bit mode (ADARUN parameter V64BIT is set to YES), the new ADARUN parameter LARGEPAGE allows the nucleus to use large pages, which may particularly improve the performance of long-running, memory-access intensive applications.

• The new NPLOGBUFFERS ADARUN parameter can be used to increase the number of PLOG protection I/O buffers, allowing multiple PLOG blocks to be written in a single write I/O to the PLOG and improving the performance of applications that were formerly constrained by writes to the PLOG.

• The new NWORK1BUFFERS ADARUN parameter can be used to increase the number of Work part 1 protection I/O buffers, allowing multiple Work part 1 blocks to be written in a single write I/O to the Work data set and improving the performance of applications that were formerly constrained by writes to Work part 1.

• An internal Software AG Vector Table (SAGVT) is introduced in this release for z/OS systems. This is an internal z/OS CSA structure anchored in a CVT-based field assigned by IBM to Software AG. The vector table is used internally by Adabas to identify the location of an internal SVC table Adabas now uses to store and maintain SVC numbers and their corresponding IDT addresses.

  In prior releases, this information was placed in the internal SSCT entry created every time an SVC is installed. To locate the IDT in former releases, Adabas was required to sequentially scan all of the SSCTs. With Adabas 8.2, when an SVC is installed, its IDT information is stored in the SVC table identified by the vector table. To locate the IDT associated with the SVC, Adabas need only scan the internal SVC table, thus eliminating the more time-intensive SSCT scan.

  For information on the Adabas SVC, read Initializing the Adabas Communication Environment in z/OS .

• Two new LGBLSET parameters have been added, UBTYPE and UBSTIME, that can be used to:

  • obtain a single user block element per CICS task (instead of a user block pool) and

  • set a time interval to clean up locked user blocks.

  These new parameters help eliminate abends that can occur when CICS transactions lock the user blocks but do not release them. For more information about the LGBLSET macro, read Modifying Source Member Defaults (LGBLSET Macro) in Version 8 .

Nucleus 64-Bit Processing Support

Adabas 8.2 SP1 introduces support for 64-bit processing. If your z/OS operating system supports 64-bit virtual storage, you can now request that an Adabas nucleus use this virtual storage above the 2 gigabyte bar. A new ADARUN parameter, V64BIT, allows you to indicate whether your Adabas nucleus should use virtual storage above the 2 gigabyte bar. For more information, read V64BIT : 64-Bit Virtual Storage Option.

Adabas Buffer Description (ABD) Structure Changes

Updates have been made in Adabas 8.2 SP1 to support 64-bit and ALET-qualified buffers in Adabas calls. The structure of Adabas buffer descriptions (ABDs) has changed to accommodate this support. Two reserved fields (ABDXRSV4 and ABDXRSV5) have been replaced with new DSECT fields ABDXALET and ABDXADRG respectively.

• ABDXALET is now used to store the ALET for the buffer (if ABDXLOC=C'D').

• ABDXADRG is now used to store the 64-bit indirect address pointer (if ABDXLOC=C'I' or C'D').

For complete information, read Adabas Buffer Descriptions (ABDs) .

Large Page Support

Adabas 8.2 SP1 introduces large page support for z/OS systems. If your z/OS operating system supports 64-bit virtual storage and large pages (1 MB pages of real storage), you can request that the Adabas nucleus use large pages. A new ADARUN parameter, LARGEPAGE, allows you to indicate whether your Adabas nucleus should use large pages or not. The default is not to use large pages. For more information, read LARGEPAGE : Large Page Support Option.

z/OS SVC Dump Enhancements

The z/OS SVC dump support in Adabas 8.2 SP1 has been enhanced to support the changes made by IBM to IEATDUMP in a recent z/OS release. z/OS users running non-authorized should be able to use the transaction dump IEATDUMP service in a similar way to authorized users with SVCDUMP.

For more information, read Using the z/OS SVC Dump Facility.

z/OS SVC Authorization Enhancements

Adabas 8.2 SP1 introduces support for the System Authorization Facility (SAF) in providing authorization for the Adabas 8 SVC in z/OS environments. SAF can now be used in some circumstances to ensure that the Adabas nucleus, as well as other MPM servers (such as Entire Net-Work and the Natural Global Buffer Pool) are authorized to prevent inappropriate use of critical ADASVC functions. There are still some circumstances where APF authorization is required instead. For complete information, read Authorization Requirements.

z/OS SVC Prelinked to SAF

Adabas 8.2 SP1 no longer requires that you link the Adabas SVC (ADASVC) with SAF (SVCSAF) during installation. In Adabas 8.2, the SVC is prelinked to SAF and the SVCSAF WAL library member is no longer supplied.

Reporting on Replication-Related Records in Work Part 1

Adabas 8.2 SP1 provides ADAWRK utility support for replication-related records written to Work part 1. Prior versions of the ADAWRK utility did not provide this support. You can now use the ADAWRK utility to report on replication-related records so you can determine how to handle the data when the database autostart fails and the database will not come up.

For complete information on replication support in the ADAWRK utility, read ADAWRK Utility: Work Area Recovery Reports

Printing Adabas Maintenance and SVC Status Information

A new utility, ADAZIN, is introduced in Adabas 8.2 SP1. ADAZIN can be used to print maintenance information about Adabas load modules and status information about the Adabas SVC on the system in which ADAZIN is run. Names of target load modules and SVC numbers can be specified to the utility to limit the range of the printed report.

Replication of Security Definitions

Adabas 8.2 SP1 allows you to replicate security definitions if you have Event Replicator for Adabas 3.2 or later installed. For more information, refer to your Event Replicator for Adabas documentation.

Address Converter Extent Handling Enhancement

With Adabas 8.2 SP1, contiguous address converter (AC) blocks are added to the last existing AC extent, if possible, rather than creating a new AC extent. However, if contiguous blocks are not available, a new AC extent is created.

This enhancement was made in order to reduce the number extents defined in the File Control Block (FCB). At this time, it affects only your Adabas nuclei. The Adabas utilities do not support this update at this time.

Adabas Delta Save Facility Update: ADASAV MERGE Limitation Lifted

Adabas 8.2 SP1 includes internal enhancements that allow Adabas Delta Save Facility to handle overlapping delta save numbers in ADASAV MERGE processing. In prior releases, the merge of two delta save data sets with overlapping delta save numbers would terminate abnormally with the message:

DSF065 Input Full and Delta Save tapes not fitting together

This enhancement allows data sets with DSIDs that share a delta save number to be merged. For example, the save tape with DSID 7/1-5/timestamp1 can now be merged successfully with a save tape with DSID 7/5-9/timestamp2. The fact that overlapping delta save number 5 exists on both tapes and in both DSIDs can now be handled successfully during the merge.

For complete information about the ADASAV MERGE utility function, refer to your Adabas Delta Save Facility documentation.

New and Changed Utility Functions

Adabas 8.2 introduces the following utility enhancements:

DSF065 Input Full and Delta Save tapes not fitting together

New and Changed Operator Commands

Starting with Adabas 8.2, display-type commands from the operator console (such as DSTAT and DTH) will be executed directly if all Adabas threads or Adabas internal command queue elements are in use by other commands. This enhancement allows you to better determine the status of Adabas should a problem occur that impedes normal command processing. Note that this direct execution of display-type commands applies only to commands from the operator console, but not to commands issued by the ADADBS OPERCOM utility function, Adabas Online System, or Adabas Manager. Also, Adabas might be unable to respond to operator commands (issued from the operator console) in certain error situations, such as if Adabas is looping, in a hard wait, or when it is unable to obtain the CPU processing cycles it needs.

The following table lists new and changed operator commands in Adabas 8.2.

For more information about any Adabas operator command, read Operator Commands.

ADARUN Parameter Changes

This section describes the changes made to ADARUN parameters in this release. For more information about any ADARUN parameter, read Adabas Initialization (ADARUN Statement) .

The following table summarizes the ADARUN parameter changes in Adabas 8.2.

Direct Call Command Enhancements

The following direct call command enhancements were made in Adabas 8.2:

• Statistics for the internally used V* and Y* commands are now included on the Adabas Online System Command Usage screen as well as in Adabas shutdown statistics. These commands are used internally by Adabas and Adabas add-on products and should not be used in direct calls in your applications. Should you use them, errors will result.

• A new daylight savings time indicator, D, can be specified in format and search buffers to indicate that a daylight savings adjustment is required for a date-time field with the TZ option specified. For complete information, read Daylight Savings Indicator (D).

• Date-time edit masks are supported in the format and search buffers of direct calls. You can add date-time edit mask to fields with formats of B, F, P, or U in a loaded file. For more information, read Date-Time Edit Mask Processing in Format and Search Buffers.

• The following table lists new and changed direct call commands in Adabas 8.2.

  Command	New or Changed	Enhancement Description
  A1	Changed	A new Command Option 2 value, L, can now be specified in the Adabas (ACBX or ACB) control block to indicate that the position within a large object (LOB) field be tracked in the ISN lower limit field. This option is useful when multiple A1 commands are being issued in a series for a LOB field. This option can only be used if a LOB field is specified in the format buffer using LOB segment notation with asterisk notation in the bytenum specification. For more information, read Processing LOB Segments
  BT	Changed	A new Command Option 3 value, H, can now be specified in the ACBX control block to indicate that a record should be kept in shared hold status indefinitely. For more information about shared hold status, read Shared Hold Status.
  CL	Changed	The following updates have been made: In ACBX calls only, user CPU consumption values are now returned in eight bytes (instead of four bytes) and are stored in the ACBXISQG field. The unit of measurement is 1/4096 of a microsecond (instead of 1.048576 seconds). You can now request multifetch statistics when a CL command is issued if the PREFSTDD ADARUN parameter is specified in the startup JCL and if the JCL includes a valid job control statement with the label name specified by the PREFSTDD ADARUN parameter.
  ET	Changed	A new Command Option 3 value, H, can now be specified in the ACBX control block to indicate that a record should be kept in shared hold status indefinitely. For more information about shared hold status, read Shared Hold Status.
  HI	Changed	A new Command Option 3 value, S, can now be specified in the ACBX control block to indicate that a record should be kept in shared hold status until the end of the transaction. For more information about shared hold status, read Shared Hold Status.
  L1	Changed	A new Command Option 2 value, L, can now be specified in the Adabas (ACBX or ACB) control block to indicate that the position within a large object (LOB) field be tracked in the ISN lower limit field. This option is useful when multiple L1 commands are being issued in a series for a LOB field. This option can only be used if a LOB field is specified in the format buffer using LOB segment notation with asterisk notation in the bytenum specification. For more information, read Processing LOB Segments
  L4	Changed	Several new Command Option 3 values, C, Q and S, can now be specified in the ACBX control block to indicate that a record should be kept in shared hold status for varying time periods, depending on the option. Option C puts the record in shared hold status for the duration of the read operation; option Q puts the record in shared hold status until the next record in the read sequence is read or the read sequence or transaction is terminated; option S puts the record in shared hold status until the end of the transaction. For more information about shared hold status, read Shared Hold Status. A new Command Option 2 value, L, can now be specified in the Adabas (ACBX or ACB) control block to indicate that the position within a large object (LOB) field be tracked in the ISN lower limit field. This option is useful when multiple L1 commands are being issued in a series for a LOB field. This option can only be used if a LOB field is specified in the format buffer using LOB segment notation with asterisk notation in the bytenum specification. For more information, read Processing LOB Segments
  L5	Changed	Several new Command Option 3 values, C, Q and S, can now be specified in the ACBX control block to indicate that a record should be kept in shared hold status for varying time periods, depending on the option. Option C puts the record in shared hold status for the duration of the read operation; option Q puts the record in shared hold status until the next record in the read sequence is read or the read sequence or transaction is terminated; option S puts the record in shared hold status until the end of the transaction. For more information about shared hold status, read Shared Hold Status.
  L6	Changed	Several new Command Option 3 values, C, Q and S, can now be specified in the ACBX control block to indicate that a record should be kept in shared hold status for varying time periods, depending on the option. Option C puts the record in shared hold status for the duration of the read operation; option Q puts the record in shared hold status until the next record in the read sequence is read or the read sequence or transaction is terminated; option S puts the record in shared hold status until the end of the transaction. For more information about shared hold status, read Shared Hold Status.
  LF	Changed	A new Command Option 2 setting of "X" is supported for this command. This setting produces a record buffer containing extended field information, including collation descriptor, subfield, superfield, subdescriptor, superdescriptor, hyperdescriptor, and phonetic descriptor information with extended date and time formats.
  OP	Changed	A new TZ (time zone) record buffer option is available for the OP command. You can use this option to set the time zone of an individual user session. If no time zone is specified for a user session, response 55 (ADARSP055) subcode 30 is issued when a field with the TZ option is accessed or updated. For more information about using the TZ record buffer option in an OP command, read about the OP command Record Buffer.
  RI	Changed	Adabas on mainframes no longer releases records unconditionally. If your application issues an RI command for a record that has been updated in the current transaction, Adabas will now return response code 113 (ADARSP113) or, if ISN=0 was specified, response code 2 (ADARSP002) is returned. This change in behavior may affect how your application programs are coded. For more information, read RI Commands: Change Possibly Affecting Your Application Programs. In addition a new Command Option 3 value, S, can now be specified in the ACBX control block to indicate that a record should be kept in shared hold status until the end of the transaction. For more information about shared hold status, read Shared Hold Status.
  S4	Changed	Several new Command Option 3 values, C, Q and S, can now be specified in the ACBX control block to indicate that a record should be kept in shared hold status for varying time periods, depending on the option. Option C puts the record in shared hold status for the duration of the read operation; option Q puts the record in shared hold status until the next record in the read sequence is read or the read sequence or transaction is terminated; option S puts the record in shared hold status until the end of the transaction. For more information about shared hold status, read Shared Hold Status.

For more information about direct call commands, read Commands.

New LGBLSET Macro Parameters

The following new LGBLSET parameters were added in Adabas 8.2.

For more information about the LGBLSET macro, read Modifying Source Member Defaults (LGBLSET Macro) in Version 8 .

New and Changed Sample Jobs

The following sample jobs were added or altered in Adabas 8.2.

www.softwareag.com

Copyright © 1971-2013 Software AG, Darmstadt, Germany.

SEARCH   INDEX   CONTENTS   |   PDF PAGE   PDF BOOKS   |   HOME   UP   PREV   NEXT