Re: Odd problem with cache...

From: Steve Howard <stevedhoward_at_gmail.com>
Date: 5 Sep 2006 10:46:32 -0700
Message-ID: <1157478392.131020.206190@i42g2000cwa.googlegroups.com>

Alfons Åberg wrote:
> At my job we have an odd problem with and 8.1.7 DB.
>
> We've a Java JDBC quuery, which uses bind variables. One the first execute
> the performance is slow - succeding fast. Now, starting a new browser
> session and executing the exact same piece of code - the above pattern
> happens again. Should the query be cached at this point?
>
> I should mentioned that this behaviour occours at both our development
> systems and our production environment. Both with very little user
> activity - hence the cache shouldn't be flushed in between.
>
> Do Oracle keep a kind of session cache or what?
>
> Regards, Alfman

Hi Alfman,

Trace the sessions and it will show you what they are doing. Google for...

oracle 10046 trace
tkprof

You can even embed this in the sessions by something like the following before you execute your statement...

conn.createStatement().executeUpdate("alter session set events '10046 trace name context forever, level 12'");

Regards,

Steve Received on Tue Sep 05 2006 - 12:46:32 CDT