Hi,
I have a esql/c program that run this sql (output of onstat -g ses):

Insert into

cs_nyd_disg_ofi(cod_periodo,grupo_datos,numero_cli ente,monto_a_compensar,di
......pond_t_pro_e,di_pond_d_pro_e,nhs)
values(?,?,?,?,?,?,?,?,?,?....,?,?,?,?,?,?,?,?,?,? ,?,?,?)


When I run, the checkpoint rises. Usually this time is 0 seconds, but
when this insert runs, the checkpoints rise to 5 seconds aprox. We
consider this a normal behaviour.

But, when I run onstat -u each second, I notice that the nwrites don't
increase (this increase only after many seconds), but the nreads
increases quickly.

c0000000a032f4f8 --BPX-- 3642148 luz - 0
0 5678 5953233 466518
...
c0000000a032f4f8 --BPX-- 3642148 luz - 0
0 5686 5975871 466518
...
c0000000a032f4f8 --BPX-- 3642148 luz - 0
0 5690 5985582 466518


Is this behaviour normal ? Why the nreads increases if the sql that
shows onstat -g ses is an insert ?

Thanks in advance

Peter

What else is the application doing? Is it selecting data from other tables
before inserting? Is the insert being used as part of a PUT CURSOR?

Art

Art S. Kagel
Advanced DataTools (www.advancedatatools.com)
IIUG Board of Directors (art (AT) iiug (DOT) org)
Blog: http://informix-myview.blogspot.com/

Disclaimer: Please keep in mind that my own opinions are my own opinions and
do not reflect on my employer, Advanced DataTools, the IIUG, nor any other
organization with which I am associated either explicitly, implicitly, or by
inference. Neither do those opinions reflect those of other individuals
affiliated with any entity with which I am affiliated nor those of the
entities themselves.



On Thu, Nov 4, 2010 at 7:03 PM, Peternt <roger_vilca (AT) yahoo (DOT) es> wrote:

As Art says what else is it doing; try to get an sqexplain;

(should be possible using onmode nowadays) or get a sqlidebug...

eq

export SQLIDEBUG=2:/somedirwithloootttsofspace/somefile

../<run your program.>


how does the schema look like
maybe a trigger gets fired,
(which may do a select without a where clause or?? dono)
or it has zillion of indexes to maintain??

Superboer


On 5 nov, 03:41, Art Kagel <art.ka... (AT) gmail (DOT) com> wrote:

One more thing was a bit too trigger happy:

Make sure TBLSPACE_STATS 1 in your onconfig

if you can run this stand alone eq noone else accessing it during a
test then maybe
zero the counters (onstat -z) and run your stuff.

then when done select * from sysmaster:sysptprof -->> check for nreads
etc
so now you can see what the engine is reading during your program.


Superboer




On 5 nov, 09:10, Superboer <superbo... (AT) t-online (DOT) de> wrote:

On Nov 4, 6:03*pm, Peternt <roger_vi... (AT) yahoo (DOT) es> wrote:
Well - it is if there are indexes on the table ;-)

We have to read through the index structure to figure out where to put
in the new index item.