This section covers the following topics:
The Adabas Bridge for DL/I uses several internal fields to reflect the
DL/I hierarchy. These fields are the Z-field (Z0, Z1, Z2 etc.) and
the secondary index fields (their Adabas names start usually with
‘X’). The Z-fields are stored with each dependent record, the
secondary index fields with each secondary index source record.
Some of the internal fields contain Adabas ISNs; with ADL 2.2 the
Z0 and Z1 fields additionally contained a DBID and
file number. The layout of ADL internal fields has been changed with ADL 2.3 to
be able to handle 4-byte ISNs and 2-byte DBIDs and file numbers. The new layout
(see the section Performance
Considerations in the ADL Interfaces
documentation for more information) is not compatible with the ADL 2.2 layout,
i.e. ADL 2.3 does not support the ADL 2.2 layout and vice versa. Therefore all
data which have been converted with ADL 2.2 or below must be migrated to the
new ADL 2.3 layout before you can work with ADL 2.3.
Although ADL 2.3 is able to run with ADL 2.2 directory entries, physical DBD definitions must be reconverted to generate the new ADACMP and DAZUEX06 cards. It is not required to re-convert PSB definitions or logical DBD definitions.
For the migration, all ADL files (i.e. files converted from DL/I) must be unloaded and reloaded with Adabas utilities. At the reload the Adabas User Exit 6 DAZUEXMI maps the old layout to the new layout.
The following steps must be performed for the migration of the data to the new layout:
After these steps have been performed for all ADL files, you can run your applications against ADL 2.3.
For emergency, ADL offers a backward migration from ADL 2.3 to ADL 2.2. See section Backward Migration for details.
The following sections describe the migration steps in more details.
Copy the ADL directory file with standard Adabas utilities. The new file will be used as ADL 2.3 directory.
Note:
Although this step is not required, it allows to convert the DBD
definitions to the new directory (step 3) while ADL 2.2 is still
running.
Save the ADL 2.2 User Exit 6 extension (if available) and the old ADACMP cards. The old User Exit 6 extension simplifies the data migration (step 6) because you do not need to specify the SEQ parameter. The old ADACMP cards will be checked later against the new ADACMP card to identify manual changes (step 5)
The installation of ADL 2.3 is described in Upgrading to New ADL Releases in the section Miscellaneous.
Load the new error messages into the new ADL 2.3 directory file.
If you have used logical file numbers with ADL 2.2, you can delete the DAZLDT entries from the ADL 2.3 parameter module.
The CBC utility is described in details in the ADL Conversion documentation, section ADL Conversion Utilities for DBDs and PSBs. Consider the following:
There is no need to re-assemble the DBD definition (step 1 of the CBC utility).
The DBID and FNR
parameters of the GENDBD/GENSEG functions reflect the physical file, i.e. the
Adabas file where the data is stored (with ADL 2.2 it reflected the logical
file).
If the DBID of the data file is the same as the DBID of the ADL
directory, omit the DBID parameter for the GENDBD
function. This eases the creation of mirror databases.
Specify as LOGID the previous value of FNR
for every GENDBD/GENSEG function where the FNR parameter
was specified.
Specify all other parameters as with ADL 2.2. Especially if you have
used the BACKW parameter, it must use the same value as
before.
The member with the ADACMP cards is named Wfffff, where fffff is the file number. Use several libraries to save ADACMP cards of several DBIDs.
The member with the User Exit 6 extension is named Ifffff, where fffff is the file number.
Assemble the Adabas User Exit 6 extension and link-edit this with
DAZUEXMI. The Adabas User Exit 6 extension is the output of step 3
of the control block conversion.
If available, use the extension generated by ADL 2.2, otherwise the extension generated by ADL 2.3.
If you had adjusted the Adabas FDT (additional user fields, super descriptors, etc.) with previous ADL versions, you must apply the changes to the ADACMP cards accordingly.
Unload each Adabas file used to store the converted data using the standard Adabas utility ADACMP DECOMPRESS. Specify the option ISN.
The data of each file is loaded individually using the standard Adabas utilities ADACMP COMPRESS and ADALOD.
The sequential file produced by the ADACMP DECOMPRESS is taken as input for the ADACMP COMPRESS step, as are the ADACMP statements generated by the CBC utility.
Each Adabas file used must be loaded with the option USERISN. This applies to both the ADACMP and the ADALOD steps (for ADACMP it is already generated by the CBC utility).
The ADACMP COMPRESS step uses Adabas User Exit 6. This exit consists of two parts which were linked together in step 4:
Fixed Part
The fixed part consists of the DAZUEXMI module, which
maps the ADL 2.2 layout to the ADL 2.3 layout.
User Exit 6 Extension
The User Exit 6 extension is generated by the CBC
utility and contains information on the structure of the DBD being migrated,
and the default record layouts of the Adabas file(s) used to store the data.
You can use the User Exit 6 Extension generated by ADL 2.2 as well as the
extension generated by ADL 2.3.
Adabas User Exit 6 needs a control statement to indicate which Adabas file should be migrated. The syntax of this control statement is as follows:
FNR=nnnnn,MODE=MIGR,SEQ=sss,DBID=nnn
| Parameter | Description | Possible values | Default |
|---|---|---|---|
| FNR | Specifies the file number to be processed. The value must be
the same as the FNR specified at the CBC utility run, i.e. the
logical file number if the extension was generated with ADL 2.2 or the physical
file number if it was generated with ADL 2.3.
|
1 - 65534 | |
| MODE | Specifies whether the data should be migrated from ADL 2.2 to ADL 2.3 or vice versa. |
MIGR -The data is migrated from ADL 2.2 to ADL 2.3 BACK - The data is migrated back from ADL 2.3 to ADL 2.2 |
MIGR |
| SEQ | Specifies the ADL 2.2 processing sequence of the file. This parameter must only be specified if the extension was generated with ADL 2.3. |
SEG - Sequence is "segment". This was the default for the ADL 2.2 conversion. PAR - Sequence is "parent". This is the setting for all DBDs converted with ADL 2.1 or before. |
SEG |
| DBID | Specifies the logical DBID to be inserted to the ADL internal
field. The value must be the same as the DBID specified at the ADL 2.2
CBC utility run. This parameter is mandatory for the backward
migration (MODE=BACK), otherwise it is not required.
|
1 - 255 |
For emergency, ADL offers a backward migration from ADL 2.3 to ADL 2.2. The following steps must be performed for a backward migration:
The ADL 2.2 installation is described in the ADL 2.2 documentation. If the old ADL load library is still available, you can omit this step.
Use the original ADL 2.2 directory (see step 1 of the migration).
Alternatively you can convert the DBD definition with ADL 2.2. The
DBID and FNR parameters of the
ADL 2.2 GENDBD/GENSEG functions reflect the logical file which is described in
details in the ADL 2.2 documentation.
If the logical file numbers differ from the physical file numbers, DAZLDT entries must be added to the ADL parameter module.
Unload each Adabas file used to store the converted data using the
standard Adabas utility ADACMP DECOMPRESS. Specify the option
ISN.
The data of each file is loaded individually using the standard Adabas
utilities ADACMP COMPRESS and ADALOD. The sequential
file produced by the ADACMP DECOMPRESS is taken as input for the
ADACMP COMPRESS step. Use the ADACMP statements
generated by the ADL 2.2 CBC utility run. Each Adabas file used
must be loaded with the option USERISN. This applies to both the
ADACMP and the ADALOD steps.
The ADACMP COMPRESS step uses Adabas User Exit 6. For the
backward migration use the same Adabas User Exit 6 which was used for the
(forward-) migration of the file to ADL 2.3.
The parameters of the Adabas User Exit 6 are described in details in
the section above. For the backward migration you must specify
MODE=BACK. Additionally to the parameters used for the migration,
you must specify the DBID parameter, reflecting the
logical DBID used at the ADL 2.2 CBC utility run.
The JCL requirements for the migration and backward migration are the same as for the initial load. Refer to z/OS JCL Requirements or z/VSE JCS Requirements in the section ADL Data Conversion Utilities of the ADL Conversion documentation.
z/OS CICS application programs which have been linked with the ADL language interface DAZLICI2, must be re-linked with the new language interface DAZLICI3. This is especially true for all precompiled EXEC DLI programs. See the ADL Interfaces documentation, section CICS Installation and Operation for further information.
