Re: Database Size Vs System Performance

There is some very small potential.  However I don't think I can stress 
'small' enough.

1 - If you using "supplemental groups" and you increase the number of 
objects this may affect the size of your SAVSYS.  Someone here went nuts 
with supplemental groups and our SAVSYS jumped from 4 minutes to 44.  I 
blasted them all and used a different solution.

2 - If your machine abnormally terminates (like someone flips a UPS lever) 
and you bring it back up some you may get a longer rebuild time on access 
paths.

3 - The time to run RCLSTG SELECT(*DBXREF) may increase.

4 - Any increase in usage because the development staff vastly improves 
their productivity because of safe test data to play with may impact your 
system.



CPF0998-Unattended IPL in progress after abnormal system end.  Time sent . 
. . . . . :   10:52:06
10:53:08-START OF QSPFFACB
(lots of CPF7001's)
11:05:44-END OF QSPFFACB
11:10 - Big series of CPF3124 right on up until
11:11-CPI3225-System access path protection processing has started. 
Followed by a lot of
CPF3123-Access path for member XXXXX recovered or built during IPL.
11:11-CPF3112-Data base IPL recovery complete.   But those CPF3123 kept on 
coming.
11:12:18-CPF0934-IPL Completed.
11:12:23-CPI32A6-Cross-reference notification 70 for request 1, type 2.


select count(*) 
from routines/dskdtl 
where odobtp='*FILE' 
and (odobat like 'PF%' or odobat like 'LF%')
167,958

I believe that our RCLSTG SELECT(*DBXREF) takes about 20 minutes is all. A 
full RCLSTG takes a lot longer:
DSPDTAARA QRCLSTG
Value 
 *...+....1....+.
'1050624   191358
'1050624   222557
'V5R3M0    GDIHQ 


Rob Berendt