Translating with DrWatson… this can take a few seconds the first time.
This is a genuine, complex translation. DrWatson protects commands, error codes, and log output while naturally translating the surrounding text. It’s translated once and saved.
Lester Policarpio's Read Ahead Utilization was 63.79%, with RA_THRESHOLD 30 and RA_PAGES 32, and asked how to reach near 100%. Eric Vercelletto said read ahead depends on access patterns and matters little for OLTP. Art Kagel said v11.50 changed read ahead so RAU is effectively worthless, and recommended his ratios.shr_ak package. No tuning change was identified.
Auto-generated by Claude from the posts below — may be imperfect; read the full thread.
Good Day,
I saw a post regarding the computation of BR, BTR and RAU with below scripts
select 'BR (Bufferwaits Ratio): [< 7%] ---> ' || trunc(100 * (select value from
sysprofile where name='buffwts')/((select value from sysprofile where
name='bufwrites')+(select value from sysmaster:sysprofile where
name='pagreads')),2) || ' %' from systables where tabid = 1 ;
select 'BTR (Buffer Turnover Ratio): [< 10] ---> '||
round(((( pagreads + bufwrites )
/nbuffs ) / (
select (ROUND (((
sh_curtime - sh_pfclrtime)/60)/60) )
from sysshmvals ) ),1)
from sysmaster:sysbufpool;
select 'RAU (Read Ahead Utilization): [near 100%] ---> ' || trunc(100 *
(select value
from sysmaster:sysprofile where name='rapgs_used')/((select value from
sysmaster:sysprofile where name='btradata')+(select value from
sysmaster:sysprofile where
name='btraidx')+(select value from sysmaster:sysprofile where name='dpra')),2)
|| ' %' from sysmaster:systables where tabid = 1 ;
the result for our production server is good except for RAU, does anyone know
what to do so we can achieve a near 100% result?
BR (Bufferwaits Ratio): [< 7%] ---> 1.55 %
BTR (Buffer Turnover Ratio): [< 10] ---> 1.6
RAU (Read Ahead Utilization): [near 100%] ---> 63.79 %
Thanks,
Hi,
our current onconfig value for RA_THRESHOLD and RA_PAGES are 30 and 32
respectively. Do we need to increase/decrease both values? how can we measure
the amount of increase/decrease we need to make? or is it a trial and error
thing?
Lester,
Read ahead is used only when relevant and totally depends on how your
applications access the data.
You expect to have a lot of read ahead in cases access your data sequentially,
i.e more when running Decision Support Queries.
Is this the case?
If you app is mostly OLTP type of activity, you probably don't want to be
worried with a bad RAU
The problem is that in v11.50 IBM changed how readahead works including
additional structures being read ahead that were not being processed that way
before. Unfortunately few of these new RA pages are being reported anywhere so
the RAU calculation is effectively worthless <sigh>.
Also, note that I have replaced the BTR calculation with the more accurate
BTR3 calculation which is a bit more involved. The best thing is to download
my ratios.shr_ak which includes a set of stored procedures that calculate the
various metrics and a shell script to produce a formatted report.
Your privacy choices
We use strictly necessary cookies to make this site work. With your
consent we’d also use optional cookies for analytics and marketing. You can accept all,
reject all, or choose. Read our Cookie Policy.