Location , Webserver 2.0

Location , Webserver 2.0

Post by Stian Hundhamme » Tue, 10 Jun 1997 04:00:00



I am trying to speed up my web-application, but having problem wherever
am using a owa_util.redirect_url.

After I've done a call to another site, the next request to my web
request broker takes awful long time (about 10 - 15 seconds). The same
thing happens when am doing an
<listner>/OWA/htp.print?cbuf=Location:+http//www...... , so there is
nothing wrong with my owa_util package. All the other packages are also
valid.

Do anyone had this problem? It looks like the time it takes to do the
request, is of the same amout of time as when I'm doing the first
request after I've started my web-listener, and there is no error
message in either owa.err or <listener>.err

Where do I start to look for errors/parameters to fix and so on?

I'm using Oracle 7.3 and WebServer 2.0 on the same machine, an Silicon
Graphics Indy. The requestbroker is set up with SID-communication (and
not SQL Net)

So - can anybody help me?

Stian

--


  telephone:                                            
  h : +47 22 46 20 96                                    
  w : +47 22 87 18 81                                    

  web : http://techteam.origo.no/~stian/

 
 
 

Location , Webserver 2.0

Post by Stian Hundhamme » Wed, 11 Jun 1997 04:00:00



> I am trying to speed up my web-application, but having problem wherever
> am using a owa_util.redirect_url.

> After I've done a call to another site, the next request to my web
> request broker takes awful long time (about 10 - 15 seconds). The same
> thing happens when am doing an
> <listner>/OWA/htp.print?cbuf=Location:+http//www...... , so there is
> nothing wrong with my owa_util package. All the other packages are also
> valid.

> Do anyone had this problem? It looks like the time it takes to do the
> request, is of the same amout of time as when I'm doing the first
> request after I've started my web-listener, and there is no error
> message in either owa.err or <listener>.err

> Where do I start to look for errors/parameters to fix and so on?

> I'm using Oracle 7.3 and WebServer 2.0 on the same machine, an Silicon
> Graphics Indy. The requestbroker is set up with SID-communication (and
> not SQL Net)

> So - can anybody help me?

Here is the answer to my own question :

The minimum number of WRB Execution Engines (WRBXs) running was set to
0. This ment that for every time someone jumpt out my application with
an Location in the HTML-header, one WRB process died, and it took about
10 seconds to start a new one. This problem is now solved by setting the
minimum WRB's to a higher number of processes.

BUT : Even if I sets the minimun processes to 20, they still die one
after another when an owa_util.redirect_url occurs. This means that
after a while, I'm down to 0 processes anyway. In the docomentation it
says : Min

        The minimum number of WRB Execution Engines (WRBXs) running this WRB
cartridge that must be running at all times.

Why do they stop running? Can somebody help me with this?

Thanks -

Stian
-----------------------------------------------


  telephone:                                            
  h : +47 22 46 20 96                                    
  w : +47 22 87 18 81                                    

  web : http://techteam.origo.no/~stian/

 
 
 

1. Problem for WebServer 2.0 on N.T.

Hi Guys,

        I am having real trouble with Oracle WebServers.  I have developed a
web-site for a Property Agency in HongKong.  The Oracle v7.2 workgroup
database is sitting on an extremely powerful NT server, with Oracle
WebServer v2.0 on it as well.  For now, the web-site is having approx. 3000
hits a day.  The NT is a Compaq Proliant with four 200 Mhz pentium pro cpu,
and 500 Mbytes of ram, which I think is way enough to handle only 3000 hits
per day.  The site also has a T1 connection to the HK Internet Exchange.
My problem is, the web listener seems to crash the NT server once its got
about 2000-3000 hits.  Sometimes it came out with an NT memory error
message saying memory couldn't be read or written.  Even if its not
crashed, the server seems to response extremely slow, and this happens
gradually.  That means, once I reboot the server, it goes really fast
again, and after about several thousand hits, it becomes slower and slower.
 It seems like this is the problem with the listener, because when I try to
restart the listener from the service menu under control panel, it would
restart and then it would usually crash, which forces me the reboot the
machine.  I tried to contact the Oracle supporting team but what I got from
them is that they don't know much about NT version.  What I could think of
the cause is that it's actually a bug for the webserver that it wouldn't
free up memory after every access.  

Anyway, does anyone have this experience on NT version or if anyone could
help me out?

Also, because of this, I am think of switching the web-site to UNIX
platform.  I have been recently looking at the SGI Origin 200 server.  Does
anyone know if its suitable for this purpose, cause I thought Oracle in
UNIX version is much more stable.

Thanks in advance, guys!

2. INFORMIX DBA-UNIX-CTR-FLA

3. Java Cartdrige for WebServer 2.0

4. datagrid and sql

5. Oracle Webserver 2.0 - http logs rotating

6. Microsoft.Jet.OLEDB.4.0

7. Webserver 2.0 on NT

8. Authorization problem in Oracle Webserver 2.0 (and 1.0)

9. Authorization information in Oracle Webserver 2.0