Public disks are combined to volume sets or SF pubsets. One or more volume sets constitute a pubset.
Several pubsets can be installed simultaneously in a system run, thus there are several catalog IDs (CATIDs).
ARCHIVE supports SM pubsets in the same way as SF pubsets. Specific file attributes of SM pubsets are saved and restored. With RESTORE, ARCHIVE automatically selects the pubset that is most suitable and does not require conversion.
In case of a (later) transition to HSMS, it is recommended for SM pubsets that the directories used are restricted to one SM pubset each.
In order to support several catalogs, ARCHIVE offers the CATID operand in the PARAM statement. This operand is used in save and restoration runs:
PARAM CATID=NO|YES
This operand can be used to define whether or not ARCHIVE is to save the catalog ID of the saved files and job variables. Its meaning varies depending on whether it is used in a save or restoration run.
The following applies to save runs and restoration runs
Copying of save data must be carried out in the same mode as the original save run. For this reason, either CATID=YES or CATID=NO must be specified when copying save data.
Saves that were created in an ARCHIVE version < 2.4A, are created with CATID=NO
CONTINUE with a mode other than that of the original save is rejected with the message ARC0277.
Save runs
In save runs with CATID=NO (default value), the catalog ID is ignored. This means that:
the catalog ID of the saved file is not in the report
no catalog ID is used for DMS macros; the standard system regulations apply
space characters are written into the save file in place of the catalog ID.
the catalog ID is not entered in the FILES records of the directory file the attempt to enter a catalog ID in an ARCHIVE statement is rejected
extending an existing save version (CONTINUE operand) with a different mode to that of the original save operation is illegal.
If CATID=YES is specified for a save run, the following applies:
the catalog ID of the saved file is specified in the report before the user ID in the page header
files are only sought in the catalogs with the specified catalog ID
if a file/job variable name is specified without a catalog ID, the standard catalog ID of the relevant user ID is added to the name.
The following changes in the pubset structure during an ARCHIVE run lead to wrong or incomplete results:
if a pubset which is being processed in the current run is exported
if the standard catalog ID in the JOIN file of the source pubset is modified for the LOGON user ID or another user ID that is affected by the current run.
The same conditions must be satisfied for restarted ARCHIVE processes as applied before the abnormal termination:
the standard catalog IDs must not have been modified
no pubset may be exported or imported between the commencement of the interrupted run and the RESTART run.
Restoration runs
The statement PARAM CATID=NO/YES specifies in which mode the save to be read in was created.
For runs with CATID=NO (default value), the following applies:
the save run to be read in must have been created with
CATID=NOthe files or job variables are written back to the default pubset of the relevant user ID
however, they can also be written back to any other pubset specified by means of the RENAME operand
apart from in the RENAME operand, no catalog IDs can be specified in the ARCHIVE statement
the report does not contain any catalog IDs, unless one was specified in the RENAME operand, in which case it is always entered in the header line before the user ID.
For restoration runs with CATID=YES the following applies:
the save run to be read in must have been created with
CATID=YESby default, the files or job variables are written to the pubset on which they resided at the time of saving (source pubset)
using the RENAME operand, the files or job variables can be written to any default pubset; in the case of catalog entries, only the catalog ID may be renamed
this catalog ID is contained in the header line in the report; when RENAME is specified, the catalog ID of the source pubset is contained in the RENAME message.
If a directory file is being used, the catalog ID is transferred to the key of the FILES records in the case of runs with CATID=YES. The directory file contains an identifier indicating whether a catalog ID is present in the FILES records.
During a save run with CATID=YES it is possible to use a directory file created with CATID=NO, provided this directory file is converted before the run with the aid of DIRCONV (see the “HSMS” manual [9]). Following conversion, it is only possible to work with this directory file with CATID=YES.
As a rule, no pubset with the original catalog ID exists for IMPORT runs. When this is the case, the files and job variables must be renamed.
If file names or job variable names are specified without a catalog ID, the standard ID of the user ID in question is taken as the catalog ID. It must be borne in mind that the catalog IDs may be different at the time of the save run and the restoration run.
If the user ID no longer exists, no standard catalog ID can be determined. In this case the catalog ID of the source pubset is added to the catalog ID (for runs with CATID=YES).