RE: TKPROF output

In addition to my previous post: 

I noticed that the number of sql*net waits are the same as the number of 
fetches. So most probably the waits are indeed happening between the fetch 
calls.
Also, the time spend on sending the rows to the application server (sql*net 
message to client) only takes 0.01 second for all the waits, meaning that just 
reducing the number of round trips will not help you.
Something is happening on the application server. Either it is very slow in 
processing the records already send to it or there is a network problem from 
your application server to your db server (as the sending from the db server to 
the application server is happening fast)

regards,

Freek D'Hooge
Uptime
Oracle Database Administrator
email: freek.dhooge@xxxxxxxxx
tel +32(0)3 451 23 82
http://www.uptime.be
disclaimer: www.uptime.be/disclaimer
---
From: oracle-l-bounce@xxxxxxxxxxxxx [oracle-l-bounce@xxxxxxxxxxxxx] On Behalf 
Of D'Hooge Freek [Freek.DHooge@xxxxxxxxx]
Sent: 31 March 2011 08:01
To: lyallbarbour@xxxxxxxxxxxxxxx; oracle-l@xxxxxxxxxxxxx
Subject: RE: TKPROF output

Hi,

The main wait time is "sql*net message from client", meaning that you are 
waiting on response from the aplication server.
Can you check in the raw trace file if these waits are happening between the 
fetches or not?

If they are happening between the fetches, it means that the application server 
is slow to request more rows from the query result.
Reason for this could be some processing that happens on the application server 
or maybe a network problem (like a wrong dns server on the application server).

If the waits are happening after all the fetches are completed, then these 
waits are probably not relevant (unless you are sure that the trace was stopped 
at the moment the application showed the result of the query).
How did you trace this session?


regards,

Freek D'Hooge
Uptime
Oracle Database Administrator
email: freek.dhooge@xxxxxxxxx
tel +32(0)3 451 23 82
http://www.uptime.be
disclaimer: www.uptime.be/disclaimer--
http://www.freelists.org/webpage/oracle-l

Other related posts: