ACF2 Database space monitoring

Products

• ACF2 - z/OS

Issue/Introduction

All the ACF2 VSAM databases should be monitored for space usage. Note that, even with regular monitoring, if the  ACF2 VSAM databases have not been reorganized recently then the databases may not need more space (reallocation),  only reorganization. 

Environment

Release : 16.0

Cause

 

 

Resolution

All the ACF2 databases should be monitored using an IDCAMS LISTCAT with the ALL parameter. Note that  the IDCAMS LISTCAT information is based on the last time the VSAM file was closed (that is when the file statistics are updated) which would only occur if the system is IPL'd or if ACF2 is stopped and restarted. 

For each database, check the free space in the IDCAMS report: 

•    If %CI OR %CA IS 25% or higher, or you have many CA/CI splits, then you should do a reorganization. 

•    Check that the  HI-U-RBA is not too close to HI-A-RBA. If HI-U-RBA should become equal to the HI-A-RBA, VSAM can't do any more CI or CA splits, and cannot expand – the database is essentially full.  There may be space available in the database but it is not where it is needed.  

If reorganization or reallocation is needed, the procedure in Expanding Space Allocation of ACF2 databases can be used for either reorganization or reallocation. The only difference would be in the step 5 allocation of  the same space or increased space. 

 

Additional Information

The following are the considerations with regard to scheduling either reorg or reallocation during a quiet time. 

When ACF2 is stopped and restarted with the alternate database, during the short period of time that ACF2 is down, NEW jobs or logons, or currently running jobs/tasks/users that attempt OPEN of a datataset not previously validated will be delayed with these (and other) messages:

ACF90CCC jobname, ACF2 NOT ACTIVE: REPLY 'U' TO CONTINUE, 'C' TO CANCEL,OR 'W' TO WAIT AND RE-START ACF2

ACF91CCC jobname, ACF2 NOT ACTIVE: REPLY 'U' TO CONTINUE, 'C' TO CANCEL,OR 'W' TO WAIT AND RE-START ACF2

To minimize these messages, the initiators on each system that shares the ACF2 databases should be drained. It is not neccessary to wait for current jobs to complete. These will continue to run until a new ACF2 validation request occurs.

When ACF2 is stopped, ACF2 displays numerous messages to the operator and makes its presence known. Every request to start a new job or to access a data set needs operator verification to permit access. Reply U to above ACF messages to allow the access without ACF2 validation, or W to wait. If no reply is made then when ACF2 comes back up these messages will go away. If  "W" is not replied after ACF2 is up, then the address space may still wait. To continue processing issue any ACF2 modify command (valid or invalid) to resolve (post) the wait:

F ACF2,SHOWSYS <=== valid command

F ACF2,GO <=== or any other invalid command

None of the TSO users or any online systems need to be cancelled. Naturally, the less activity occurs while ACF2 is down, the fewer (if any) messages will appear.