The archive directory contains the information about the save data managed in the archive, i.e.
BS2000 files, job variables or node files
save files
save versions
occupied and free save volumes.
The name of the archive directory is defined with DIRECTORY-NAME when creating the archive by means of the HSMS statement CREATE-ARCHIVE.
By default, HSMS generates a new archive directory when creating an archive; it can be given any name. As an alternative, an existing directory can be reassigned to a new archive, provided it is not (no longer) associated with any other archive. However, directories of a migration or long-term archive can no longer be assigned to a backup archive after the first write request has been issued.
The archive directory need not be assigned to the user ID of its owner. However, the right to create archive directories under a user ID other than one’s own is restricted to the HSMS administrator.
The user ID of the archive directory is the same user ID for which free tapes are reserved with MAREN.
Public archives can be protected against unauthorized access by assigning protection (password, BACL or GUARDS) to the archive directory. If one of these mechanisms is used, it is essential to have write access to the archive directory, even if USER-ACCESS=*ALL-USERS(ACCESS=*READ) has been defined for the archive.
Similarly, private archives can also be opened for a group of users, with authorized access through co-ownership (via SECOS). Then, it doesn’t matter which user access is specified for the archive.
An archive directory for BS2000 files and job variables cannot be used as archive directory for node files and vice versa.
Note for ARCHIVE users
The archive directory is implemented by an ARCHIVE directory file. Existing ARCHIVE directory files can be placed under HSMS management (see section "Special aspects when switching from ARCHIVE to HSMS operation").
Notes on restricting the record length in directory files
Currently the data of a BS2000 file can be distributed to a maximum of approximately 300 volumes as the record length in directory files is limited. This value depends on the save options, e.g. with SAVE-PLAM-INFO = *YES it can decrease to around 294.
If more volumes are required when saving the file, not all the information can be written to the directory concerned. In this case the request is terminated as "COMPLETED WITH ERRORS" and message ARC0176 is issued.
With some user scenarios, gaps can occur in data records of an archive directory. Some examples of such scenarios are the following:
1. using statements //MOVE-SAVE-FILE for backup archives, //REORGANIZE-VERSION-BACKUP for version backup archives
2. with //MODIFY-ARCHIVE ... SAVE-FILE=*DELETE
3. //BACKUP-FILES with mixture of CCOPY=*NO and CCOPY=*FROM-SNAPSET
4. reorganization of migration archives
These gaps can lead to the situation that the maximum number of records for one file and/or job variables (JV) is exceeded. Although there is actually still some free space left in the directory.
To make this free space usable again it is necessary to reorganize the directory. One option is using the program DIRCONV and its statement //REORGANIZE-DIRECTORY (see section "Reorganizing directory and repository files").
From HSMS V12.0D there is also the function of automatic reorganization of the directory: automatic reorganization starts during HSMS backup runs when the amount of records in the directory is close (=250) to the max 256 per a file.
In order to activate the function, it is necessary to set the parameter AUTO-DIR-REORG to YES in SYSPAR.ARCHIVE file.