FEW VLF RETRIEVES

[Monika Amiss]

Dear group, 

at our shop we use the LLA-freeze-option for all our LLA-Dataset, 
mostly Lnklst-Datasets but also some non-Lnklst-DS (non-PDS/E). 
We defined our CSVLLA-Class big enough:  
 
CLASS NAME(CSVLLA)   
  EMAJ(LLA)  
  MAXVIRT(131072)  
  
but the amount of storage is only used up to 40% (SMF41(1)). Now, when i issue 
a 'D LLA,STATS' i see many DASD FETCHES, i.r.:   

LIBRARY: SYS1.SERBLINK 
  MEMBERS:195  
  MEMBERS FETCHED: 25  MEMBERS IN VLF:  0  
  DASD FETCHES:   501  VLF RETRIEVES:   6  
  
Are there some requirements for the modules to use VLF? 
During some testings I made a pgm, which loaded a module (non-RENT) 
about 4000 times. If this module is in VLF at starting time, it always
will be loaded from VLF-cache (VLF RETRIEVES get high) . If the module 
is not in cache at start time of my pgm, 4000 DASD calles will be issued.
Strange behaviour. Does somebody has an idea, or a hint, where to find 
some usefull IBM-documentations about VLF. We have z/OS 1.13, but with 
z/OS 1.12 we had the same behaviours. CSVLLIX1 and CSVLLIX2 or on.
 
 Any hint appreciated
 With best regards
 Monika  
  
  

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

Re: FEW VLF RETRIEVES

[Peter Relson]

The general approach for VLF is that a module is eligible for caching only 
after being fetched 10 times.

The functionality of caching is asynchronous. It will be kicked off, via 
POST, after the 10th fetch. It is possible that your test program made it 
through its LOADs before the VLF processing got around to caching the 
module.

Peter Relson
z/OS Core Technology Design

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