Working With Manage Area Retrieve In Z/OS 1.12

t chain. If the focus re-inserts archives inside of the key operation owned by the void CI, they will go there. However, if the focus inserts new archives in other key ranges, VSAM puts them in to already-occupied information CAs, leaving the orphaned CA waving in the self-evident breeze.

The CA is never reclaimed. In fact, if the focus inserts keys in with every time aloft values -- say the key is a timestamp -- there will be a solid parade of void CAs all the way down to the finish of the information set's allocation. As a systems programmer complained, it's wholly probable to finish up with a full KSDS without a singular record in it.

CA retrieve to the rescue CA retrieve solves the complaint by progressing a sequence of index CIs that own void CAs. VSAM hangs the list off of index CI2, whose location may change depending on the number of index levels and the size of the information set.

When VSAM detects an void CA, it removes the index CI indicating to it from the sequence set sequence and adds it to the giveaway chain. Other index CIs will be updated to the sequence as CAs turn empty. Note that higher-level index CIs may be updated to the giveaway sequence as good if all the sequence set CIs they indicate to are free.

As an focus adds new records, VSAM will obtain an void CA from the giveaway sequence and re-insert the index CI in to its place in the sequence set sequence instead of going to the finish of the information component. Not usually does this save space, but it improves consecutive opening and means fewer continual re-orgs.

Of course, liberation is a large concern, as CA reclamation involves a lot of estimate and I/Os during that any number of things can go wrong. According to IBM, if an ABEND occurs in the center of a reclaim, VSAM will do its most appropriate to full the operation. If the most appropriate bid fails, the reclamation algorithm ensures the KSDS will not be broken, nonetheless there may be a few orphaned space along with a few IDC* messages about the cluster's status.

Enabling CA retrieve is a small wily since it may be activated at several not similar levels with not similar defaults. By default, retrieve is off at the network level. At the information category and information set level ,the default is on. Therefore, installations will must be do a few formulation and coordination before activating this feature.

At the network level, varying the IGDSMS* CA_RECLAIM parameter to DATACLASS turns reclamation on. There are two ways to succeed CA retrieve at the information set level. The information category CA Reclaim blame on relates to groups of information sets and does not take outcome until a cluster is deleted and re-defined. The IDCAMS ALTER order can toggle reclamation at the information set level and becomes efficient upon open.

CA retrieve isn't the most appropriate selection for every information set. Data sets whose applications insert and undo archives at pointless will not see any benefit. On the other hand, information sets where a operation of keys are deleted but not subsequently re-inserted will see the greatest boost. Applications that undo archives in slight key ranges, usually to re-insert them later, may obviously see a opening hit. Also, note that CA reclamation is not retroactive. Even when switched on, information sets with existing waif CAs will not remove them until someone re-orgs the information set.

Finally, be certain to have all the compatibility PTFs practical if you're going to blend z/OS releases in the same Sysplex. Failure to do so may result in information honesty problems.

ABOUT THE AUTHOR : For 24 years, Robert Crawford has worked off and on as a CICS systems programmer. He is gifted in debugging and tuning applications and has created in COBOL, Assembler and C++ using VSAM, DLI and DB2.

What did you regard of this feature? Write to SearchDataCenter.com's Matt Stansberry about your information center concerns at mstansberry@techtarget.com .