In <[EMAIL PROTECTED]>,
on 10/28/2005
at 12:00 AM, Ted MacNEIL <[EMAIL PROTECTED]> said:
>Possibly, but I don't think so.I was probably over-reacting.We don't
>implement a 4TB system overnight.We plan and understand (sometimes on
>both points)
K3wl. What do you do when someone fails to plan? The issue in dispute
does not and can not arise when the DB2 people plan for the new
release; it only arises when they go off half cocked. Do you just let
them system crash and then do a finger-pointing exercise? Or do you
use tools like IEFUSI to attempt to keep the system up?
>Again, we plan.If a SYSPROG doesn't know better, I don't want him on
>my account.
Sure, but that doesn't mean that you have any say in the matter. From
Mark's message, it doesn't sound as though he has the authority to
remove careless or incompetent DB2 staff.
>We don't suddenly drop DB2 V8 (for example) into the environment,
Who is "we"? The DB2 admin in question did just that.
>Again, it's documented.
What's documented? The DB2 admin set the parameters, not IBM. What IBM
did was to replace a burned-out fuse with a copper penney.
>We don't complain about the amount of resource (within reason).
Crashing the production system does not qualify as "within reason".
>If the business needs it, it needs it.
Do you allow another department to unexpectedly purchase gear out of
your budget, or to you require them to go through channels? Mark
hasn't said that he would refuse to acquire the extra DASD to back up
the DB2 paging; he just wants to keep the system up in the meantime.
>Again, we don't blindly implement, do we?
Yes. Read Mark's message.
>My point was more of don't mess with it.Let it go through.
And crash.
>What I meant was, don't limit DB2, but make sure you know what the
>DBA's are doing.
Who's going to bell the cat? There is no way to know what the DBA's
are doing. IEFUSI used to be a way to keep them honest. Now they have
more liberty to spring surprises on you.
>I also meant that SYSPROG-types should not be applying arbitrary
I failed to see in Mark's message where he wanted to apply an
arbitrary limit. He did, however, want to prevent the system from
crashing.
>limits and impacting the business.
It's the *failure* to impose a limit that impacted the business in
Mark's case.
>DB2 V8 can save a ton of money because of the cost of CPU (hardware
>& software) that has been consumed in V6 & V7 to reduce the impact
>of the 2GB limit and the use of HIPERPOOLs.
And it can burn those savings and more in a NY minute if it's exempt
from resource controls.
>But, plan
IEFUSI used to be a way to enforce planning.
>But, don't impose arbitrary limits.
There are always limits, e.g., the total amount of space for LOCAL
paging. That doesn't mean that they are arbitrary. You're asking for
the moon if you want there to be no limits.
>SYSPROGs don't always 'know better'.
But the sysprog is often the one left holding the bag when the balloon
goes up. Marks' boss isn't going to accept "Ted doesn't want limits"
as an excuse when the system crashes.
Note: I am *not* saying that 64MiB is the appropriate limit, just that
there should be one commensurate with the sizing of local paging.
--
Shmuel (Seymour J.) Metz, SysProg and JOAT
ISO position; see <http://patriot.net/~shmuel/resume/brief.html>
We don't care. We don't have to care, we're Congress.
(S877: The Shut up and Eat Your spam act of 2003)
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
