| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
Home -> Community -> Usenet -> c.d.o.server -> Re: bad performance with sqlplus on solaris
if you have a sql script that you are using to test, then i'd suggest
modifying it as follows
at the start issue either
alter session set events '10046 trace name context forever, level 8';
or
exec dbms_support.start_trace(waits => true);
whichever you prefer and then at the end add either
alter session set events '10046 trace name context off';
or
exec dbms_support.start_trace;
respectively.
Run the resulting trace file through TKProf and see what you are waiting for. My guess would be that you will see an increase in
sql*net message to/from client and/or
sql*net more data tp/from client
in the solaris version. If so this would strongly indicate the network.
-- Niall Litchfield Oracle DBA Audit Commission UK "computer person" <fake_address_at_nothing.com> wrote in message news:WeeJb.80766$INs.56038_at_twister01.bloor.is.net.cable.rogers.com...Received on Fri Jan 02 2004 - 08:19:20 CST
> "Daniel Morgan" <damorgan_at_x.washington.edu> wrote in message
> news:1073021402.541797_at_yasure...
> > computer person wrote:
> >
> > > Hi,
> > >
> > > We are migrating our systems from aix4.3.3 to solaris8. During
migration
> we
> > > noticed that the sqlplus queries take a lot longer with the Solaris
> client
> > > then with the aix sqlplus. This was proved by running the same query
> against
> > > the solaris db server from a solaris client and aix and tabulating the
> > > results.
> > >
> > > We found 1 thing that slowed it down and adjusted the SQL accordingly.
> The
> > > LINESIZE was set real high even though the result set was only 250
bytes
> max
> > > (columns). This yielded very good results but we are looking for some
> more
> > > improvment, if possible.
> > >
> > > Any stories of people that use SQLPLUS on solaris and have gone
through
> some
> > > tuning with the client and got good results? Looks like our db server
is
> > > working well with exception of this.
> > >
> > > Thanks
> >
> > Based on what you've written my guess is that you are ignoring the
> > obvious and trying to fix the trivial.
> >
> > What got migrated from AIX to Solaris? Server or Client or both?
> > What is the hardware difference? RAM? CPU? DASD?
> > What differences in init parameters?
> > Which version and edition of Oracle?
> > Which optimizer?
> > If CBO were statistics run with DBMS_STATS after migration?
> >
> > Unless you are being billed by the byte a bit of information would be
> > helpful. Cryptic information begats cryptic responses.
> >
> > --
> > Daniel Morgan
> > http://www.outreach.washington.edu/ext/certificates/oad/oad_crs.asp
> > http://www.outreach.washington.edu/ext/certificates/aoa/aoa_crs.asp
> > damorgan_at_x.washington.edu
> > (replace 'x' with a 'u' to reply)
> >
>
> That would have been a great response if I had not qualified that I ran
the
> same test from 2 clients (aix and sun) to the same db server. Thanks
anyway. > >
![]() |
![]() |