Jesse 1 Robinson wrote:

>Whether 'tis nobler in the mind to suffer    
>The slings and arrows of secondary extents,     
>Or to take arms against a sea of B37s,   
>And by opposing end them?   

Good one. +1 for you. For me: c all 'B37' 'ICH408I'  ;-)


>Gil may have been right about this being the October popic. For ServerPac data 
>sets, I encourage the *judicious* use of secondary extents. That's largely 
>because we (OK, almost) never update live data sets. We install maintenance in 
>a service-only container--oh snaps, I just made that up--and migrate it to a 
>live environment with an IPL. A swollen data set can be inconspicuously 
>compressed before it's (re)introduced to the wild. OTOH dealing with a data 
>set that has incrementally outgrown its original primary-only allocation can 
>be a major PITA even in a service-only container. 

Agreed. This is why there is the ability to do 'rolling maintenance'. You leave 
out live things out and setup new IPL/other volsers with bigger than big 
datasets.

Of course, can you do that or not. YMMV.


>The catch is that You Need to Know What You're Doing. I know, this requirement 
>offends manage-by-airline-rag execs who see knowledge and experience as 
>obstacles to hiring practices. So be it.

Those 'manage-by-airline-rag' execs also are extrememely offended (because of 
no brain cells) by this:

1. No applying fixes on 'live' things.
2. Downtime and approvals and double check-outs/verifications are needed.
3. z/OS does not need 'three finger salute' (CTRL-ALT-DEL for reboot).
4. Nothing to click on. no GUI. (yes, I know about the new products with GUI 
interfaces, but ... )
5. ... etc ...


>...  after a couple of years' worth of new function before the next ServerPac, 
>anticipating the long-term growth of every library is pretty much a crapshoot. 

This is why you see in Program Directories the required space for each 
dataset/OMVS files needed for each product you ordered.


>Defining secondary extents that may or may not be needed in the current 
>release is a fairly cheap maneuver to head off an inconvenient space crunch.  

We discourage updates of live modules, but 'they' won't listen. Only when 
something crash, then 'they' rememeber... :-(


>As we all agree, install fixes on live systems infrequently and with great 
>care, but IBM has invested tons in mechanisms that make continuous 
>availability possible if not trivial. Just come clean that it's this or 
>guaranteed IPL. I've made a handful of gambles over the years. Won more than 
>I've lost, but each one is a cliff hanger.    

You're a true risk taker! ;-D


Seymour J Metz wrote:

>I stated that I *prefer* to never update I live linklist library, not that I 
>wouldn't or haven't done it in an emergency situation. What I do to avoid an 
>outage is not what is sound for routine maintenance.

Good. Unless you are in a catch-22 situation, but I believe you also *prefer* 
not to be there. ;-)

This thread is a very interesting one. Thanks!

Groete / Greetings
Elardus Engelbrecht

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to