Re: AW: Update: Performance Scalability

2000-07-05 Thread David Sierra Fernandez


Hi Tom, I would like to know exactly the diference between
ejb-location and pooled-location.

...and other thing...but I thing that Orion team is the only one
who can answer me... How does ORion manage the connection pools???
I mean that Orion only let you set the max-number of connections but I
want to know how many connection does it create at the begining, the
number of connection that it creates when the initial ones are busy, if
these extra connections (out of initials) are destroyed after a certain
time whithout using them, an so on

This is very important because imagine that your application has a
lot of clients and inmediatly you start the server you gonna have 200
hundred clients.. in this situation it cuold be great tha the server
creates at least 250 connections at start. Now imagine that you have a hot
peak of affluence and you run out of connection (250 all busy) you will
need more but because the extension of your application, it would be a
great idea to create extra connections but in groups of 25, and destroy or
not destroy them in a timeout depends on your decision. All this with the
security that they will not be over a certain number of connections that
you set and could cause a fall in your database.

Now imagine the opposite situaion, you have a small application
whith perhaps only two clients connected to the DB at the same time, it
would be a lose of resource if your server creates 100 connections at
start, the best could be to create only 2. And if this connections are
busy it would be awfull to create a extra group of 20 connections at the
same time because probably youwill only need one or two extra connections.

Here you can see how you can take advantage of connection pooling.

I'm sure that Orion manages the connection pools eficently but If
the problem was the connection pools and weblogic let you set all this
values, Have you thought that perhaps your configuration in weblogic is
not the best for your applicaction? 

Repect to the Orion team...It would be very interesting to know something
more about how Orion manages connection pooling.


THANKS ALL.


-
David Sierra Fern ndez
E.T.S.I. Telecomunicaci¢n
Universidad de ValladolidAULA CEDETEL
Campus Miguel Delibes   E-Mail: [EMAIL PROTECTED]
47011 Valladolid (SPAIN)
--

 -- Sierr@ --

On Tue, 4 Jul 2000, Jens Stutte wrote:

 Hi Kirk,
 
 i assume he confused (like me some time ago) the different jndi names for
 the database connections. In a data-source you specify different locations,
 which will be treated differently by the server. The "location" is a 'naked'
 standard JDBC connection without pooling or anything. The "ejb-location"
 provides server specific wrapper classes for the connections, which provide
 such things as pooling, transaction handling for EJBs etc. I don't know the
 specific purpose of the "xa-location" and when to use it. Anyway, from
 within EJBs you should always use the "ejb-location".
 
 Regards,
 
 Jens Stutte
 
 PS: Example of a data-source:
 data-source 
   name="BaseDB" 
   class="com.evermind.sql.ConnectionDataSource" 
   location="jdbc/BaseDB" 
   xa-location="jdbc/xa/BaseXADB" 
   ejb-location="jdbc/BaseEJBDB" 
   inactivity-timeout="60" 
   
 url="jdbc:informix-sqli://beatnix.net-media.de:1536/BaseDB:INFORMIXSERVER=on
 _beatnix" 
   connection-driver="com.informix.jdbc.IfxDriver" 
   username="informix" 
   password="topsecret" 
 / 
 
 
  -Ursprüngliche Nachricht-
  Von:Kirk Kalvar [mailto:[EMAIL PROTECTED]]
  Gesendet am:Dienstag, 4. Juli 2000 17:19
  An: Orion-Interest
  Betreff:RE: Update: Performance  Scalability
  
  Tom:
  
  Could you be more specific?  You weren't accessing the db 
  connection pool, but what did you do to fix it?
  
  Thanks in Advance,
  
  Kirk S. Kalvar
  
 -Original Message-
From: [EMAIL PROTECTED] 
[mailto:[EMAIL PROTECTED]] 
Sent: Tuesday, July 04, 2000 9:27 AM
To:   Orion-Interest
Subject:  Update: Performance  Scalability

All,

In a previous message I expressed my concern about 
performance as compared with Weblogic 5.1.  With
 assistance 
from Karl Avedal, I made some changes to my application
 and 
am very pleased to announce that Orion is approximately 2x
 
faster than Weblogic 5.1.

I wasn't using the DB connection pool and it is a pleasure
 to 
work with a product that does conform and support the
 latest 
J2EE specification.  


RE: AW: Update: Performance Scalability

2000-07-05 Thread Tom Wnuk

That's a good question but I don't know the answer either.  I use
ejb-location in my code.  I'm going to guess that 'pooled-location' can be
used via jsp pages or other Java objects that are not participating in Ejb
transactions but maybe someone can confirm that.

Unfortunately, the one thing lacking with Orion is documentation.  There is
someone who is starting a documentation site at:
http://www.orionsupport.com/

It still has a long way to go and I'm sure it will get there.  I know
there's a number of tools that come with Orion but I have never seen any
docs on this.

As for connections, I would like to see 'startup-connections' and 'soft-max'
added to the connection pools.  Obviously, 'startup-connections' would be
the number at startup and 'soft-max' would be the number of connections
always maintained in the pool.  Then 'max-connections' would be the absolute
maximum that the pool would never exceed.  I liked your idea about
increasing connections as needed but depending on the database there may be
a maximum available from the database server.

Tom

-Original Message-
From: David Sierra Fernandez [mailto:[EMAIL PROTECTED]]
Sent: Wednesday, July 05, 2000 1:28 AM
To: [EMAIL PROTECTED]
Cc: Orion-Interest
Subject: Re: AW: Update: Performance  Scalability



Hi Tom, I would like to know exactly the diference between
ejb-location and pooled-location.

...and other thing...but I thing that Orion team is the only one
who can answer me... How does ORion manage the connection pools???
I mean that Orion only let you set the max-number of connections but I
want to know how many connection does it create at the begining, the
number of connection that it creates when the initial ones are busy, if
these extra connections (out of initials) are destroyed after a certain
time whithout using them, an so on

This is very important because imagine that your application has a
lot of clients and inmediatly you start the server you gonna have 200
hundred clients.. in this situation it cuold be great tha the server
creates at least 250 connections at start. Now imagine that you have a hot
peak of affluence and you run out of connection (250 all busy) you will
need more but because the extension of your application, it would be a
great idea to create extra connections but in groups of 25, and destroy or
not destroy them in a timeout depends on your decision. All this with the
security that they will not be over a certain number of connections that
you set and could cause a fall in your database.

Now imagine the opposite situaion, you have a small application
whith perhaps only two clients connected to the DB at the same time, it
would be a lose of resource if your server creates 100 connections at
start, the best could be to create only 2. And if this connections are
busy it would be awfull to create a extra group of 20 connections at the
same time because probably youwill only need one or two extra connections.

Here you can see how you can take advantage of connection pooling.

I'm sure that Orion manages the connection pools eficently but If
the problem was the connection pools and weblogic let you set all this
values, Have you thought that perhaps your configuration in weblogic is
not the best for your applicaction?

Repect to the Orion team...It would be very interesting to know something
more about how Orion manages connection pooling.


THANKS ALL.


-
David Sierra Fern ndez
E.T.S.I. Telecomunicaci¢n
Universidad de ValladolidAULA CEDETEL
Campus Miguel Delibes   E-Mail: [EMAIL PROTECTED]
47011 Valladolid (SPAIN)
--

 -- Sierr@ --

On Tue, 4 Jul 2000, Jens Stutte wrote:

 Hi Kirk,

 i assume he confused (like me some time ago) the different jndi names for
 the database connections. In a data-source you specify different
locations,
 which will be treated differently by the server. The "location" is a
'naked'
 standard JDBC connection without pooling or anything. The "ejb-location"
 provides server specific wrapper classes for the connections, which
provide
 such things as pooling, transaction handling for EJBs etc. I don't know
the
 specific purpose of the "xa-location" and when to use it. Anyway, from
 within EJBs you should always use the "ejb-location".

 Regards,

 Jens Stutte

 PS: Example of a data-source:
 data-source
   name="BaseDB"
   class="com.evermind.sql.ConnectionDataSource"
   location="jdbc/BaseDB"
   xa-location="jdbc/xa/BaseXADB"
   ejb-location="jdbc/BaseEJBDB"
   inactivity-timeout="60"


url="jdbc:informix-sqli://beatnix.net-media.de:1536/BaseDB:INFORMIXSERVER=on
 _beatnix"
   connection-driver="com.informix.jd