Date Last Reviewed:
03/15/2007
MVS's disk management facility is called Automatic Backup and Recovery, or ABR. What most people notice about ABR is the disappearance of some of their disk data sets. Those missing data sets are safely stored away by ABR's archive system. The less noticable but there-when-you-need-it part of ABR is its backup system. Here's how it all works.
Archiving keeps the disk packs (the ACA volumes) from filling up with unused or infrequently used data sets. Once every weekend ABR scans each data set for its last-referenced date (the last time the data set was used); if that date is 31 or more days ago, the data set will be archived--copied to tape and scratched from disk. Archived data sets are not lost forever; they will remain on the archive tapes for one year and are easily accessible at any time via a RESTORE request.
Anyone who has ever lost a data set by accidentally scratching it or by saving another data set over it can appreciate the value of a backup system. ABR's backup system automatically performs a full backup, copying every disk data set to tape, once every 7 days. Also, a backup copy is made daily of every data set that has been modified since the most recent full backup. All backup copies are written to tape and are kept for two weeks, and like archived data sets, they can be brought back to disk using a RESTORE request.
A RESTORE request brings the specified data set(s) back to disk from either an archive tape or a backup tape. All restore requests are put on a queue (in a waiting line). RESTOREs from archive tapes are processed approximately every half-hour. RESTOREs from backup tapes are processed once a day at about 2 a.m. Because ABR uncatalogs all cataloged data sets when they are archived, you may need to re-catalog restored data sets.
We recommend that restores from archive tapes be made for only the data sets you need to use. If all data sets are brought back to disk, the space-saving effects of ABR will be lost.
Simple and easy-to-use Superwylbur macros are available for public use of ABR functions. These macros will prompt you with a series of questions and set up a batch job using your responses. You can either run this job (it will run FETCH) or save the job in a TEMP (temporary file) for future use.
To use the ABR macro commands, first type:
CALL FROM &PUBLIC.ABR.MACROS:CMDS
Press Enter and you will get the usual Superwylbur prompt of "?". Now, just type in one of the following macro commands:
We recommend that you place the CALL command shown above in your PROFILE to make these ABR commands available automatically each time you log on.
This macro is used to request that a data set be brought back to disk from an archive or backup tape. If a data set has been archived or backed up more than once, the most recent version is restored.
You are required to provide this information:
If you know this information, simply type RESTORE. If you're not sure, use either ARCLIST or BACKLIST to get it. Use ARCLIST if one of your data sets isn't on the disk where you put it; it was probably archived. Use BACKLIST if you accidently scratched or damaged one of your data sets; there may be a backup copy on the backup tapes. Once you have entered the required responses, you may skip the remaining optional responses by pressing the ESCAPE key.
After you respond to the prompts, the macro will display the job it has created and ask if you want to run the job or save it in a temporary working storage data set.
To run the job, type "R". The job will run FETCH so that you can view the results at the terminal. If you want a printed listing of the job, type: PRINT nnnn, where nnnn is the job number.
To place the job in a temporary data set, type a 1- to 8-character data set name. To run the job, you can then issue the command:
RUN T tempname
where "tempname" is the name you supplied for the temporary data set. You will then get a printed listing in your bin.
To restore other than the most recent version of a data set, you will need to answer some of the optional prompts using information obtained from ARCLIST or BACKLIST. Example:
These macros are used to get lists of data sets stored on the archive and backup tapes for a given AccountID. These are two different sets of tapes.
All responses to the ARCLIST macro are optional. For a complete list of all your archived data sets, just press ENTER (or RETURN) in response to the prompts from the macro.
When using the BACKLIST macro, type responses to the prompts as shown below to get a complete list of all your data sets on the backup tapes. Next to each number is the prompt you will see. Below each prompt is the response you should give.
Enter 'y' for instructions:
Type y and press Enter
You won't need to read the instructions every time, but you should read them the first time you use the BACKLIST macro so that you will be able to properly identify which backup version you wish to bring back using the RESTORE macro. If you don't wish to read the instructions, just press Enter.
Press RETURN to list YOUR data sets, or enter a dataset name or group (optional):
Press Enter (or Return)
Enter the volume or group to be listed (optional):
Type: ACA* and press Enter
Enter 'CUR', 'ALL', or a number from 0 to 13 (optional):
Type: ALL
Enter 'Y' to list most recent backup for scratched datasets (optional):
Type: Y
Your AccountID will be added to the beginning of the name unless you put "&&&" at the beginning of the data set name. A group of data sets may be specified by appending an asterisk (*) to the data set name (this option is not available for RESTORE from backup). Here is a list of possible responses when prompted for a data set name:
Response Result
DATA.SET The data set yourAccountID.DATA.SET
will be restored.
&&&AccountID.DATA.SET The data set AccountID.DATA.SET will be
restored.
DATA* All data sets beginning with
yourAccountID.DATA will be restored.
If you do not want to use the Superwylbur macros, you may want to create and run your own batch jobs. Here's the format of a job to get a list of archived data sets and a job for restoring data sets:
Format of an ARCLIST job:
//ARCL JOB ,'your name'
/*JOBPARM R=xxx
//S1 EXEC ABRPRT
//SYSIN DD *
PRINT ARCHIVE,DSG=AccountID.,VOLG=ACA
/*
//This job will print a list of all the data sets archived from all the academic disks for the AccountID specified by DSG=.
Format of a BACKLIST job:
//BACL JOB ,'your name'
/*JOBPARM R=xxx
//S1 EXEC ABRPRT
//SYSIN DD *
PRINT VTOC,XREF,DSG=AccountID.,
OLDBACKUP=ALL,
FORMAT=CRT,
VOLG=ACA
PRINT SCRATCH,XREF,DSG=AccountID.,
FORMAT=CRT,
VOLG=ACA
/*
//This job will print a list of all data sets on the backup tapes for the AccountID specified by DSG=.
Format of a RESTORE job:
//REST JOB ,'your name'
/*JOBPARM R=xxx
//S1 EXEC ABRUT
//SYSIN DD *
RESTORE type,DSN=AccountID.data.set.name,VOL=volume
/*
//This job will restore the most recently archived or backed up copy of the specified data set. Just do the following:
ACF2 rules are in effect for all ABR requests. If you are making requests for only your own data sets, you will not have to concern yourself with setting new ACF2 rules because, as usual, you have access to all data sets beginning with your own AccountID. However, if you need to make requests for another user's data sets, the data set owner must write some special access rules. See "Quick ACF2" for details on how to set up these rules.
We encourage you to become familiar with ABR. You will find that its archive system is almost like having your own tape; a complete listing of all your archived data sets is easily available, and data sets can be brought back to disk quickly. ABR's backup system provides an invaluable safety feature which may help you keep your sanity the next time you accidentally lose an important data set.