r.skoru...@bremultibank.com.pl (R.S.) wrote:

Side question: How can I switch On/Off CA reclaim?
Is it some DC parameter or IDCAMS DEF CL one?

At the data set level, you can use RECLAIMCA and NORECLAIMCA on ALTER and DEFINE. At the system level, you can use CA_RECLAIM in IGGCATxx to turn it off at the system level(NONE) or put it under control of what's specified in the data class (DATACLAS).

--
John Eells
z/OS Technical Marketing
IBM Poughkeepsie
ee...@us.ibm.com
r.skoru...@bremultibank.com.pl (R.S.) wrote:
> W dniu 2014-12-11 o 21:53, John Eells pisze:
>> mark.jac...@custserv.com (Mark Jacobs) wrote:
>>> There "might" be some slight performance impact, and also check for any
>>> missing maintenance. That being said, we've had it turned on since zOS
>>> 1.12 and it's never caused a problem.
>>>
>> <snip>
>>
>> I am not a VSAM expert, and I didn't stay in a hotel of any
>> description last night, but...
>>
>> - For the overwhelming majority of things I'm told we expect CA
>> reclaim to improve performance (and keep a lid on space utilization
>> too), sometimes significantly.
>>
>> - There is a rather pathological case for which you should not enable
>> CA reclaim, which as far as I know we have only ever seen for real
>> within IBM. If you insert enough records into a specific keyrange to
>> split a CA (maybe more than once), and then delete them all, and then
>> add records back into that same keyrange, rinse, and then repeat a few
>> times, the overhead of repetitive CA reclaim/split processing won't
>> exactly help performance.  If it develops that you actually do have a
>> data set that's processed that way, turn it off for that data set.
>>
> Side question: How can I switch On/Off CA reclaim?
> Is it some DC parameter or IDCAMS DEF CL one?
>
>

--
John Eells
z/OS Technical Marketing
IBM Poughkeepsie
ee...@us.ibm.com

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Reply via email to