proxool-developer — Developer's discussion list

[Proxool-developer] Connection Events

[Kiran T <mah...@ho...>]

Hi Bill,

I have been trying to make use of ConnectionListener events and I somehow 
feel that the events are not comprehensive. Just for an example, I was 
trying to make use of onDeath() to send out emails whenever connection gets 
"killed" but I didn't want to send emails when these connections were 
terminated naturally( ie after 4 hours of being in the pool).

Also, I figured out that the java.sql.Connection parameter passed to the 
onDeath() is not the Proxool Connection but the native java.sql Connection. 
Was there any reason to not pass proxool's connection? I normally put the 
connection in the map while getting the connection and pop out during 
onDeath() call. Because it was not a proxool connection I had to use 
ProxoolFacade.getDelegateConnection() and then use it.

I remember an earlier discussion on splitting onExecute() to 
onExecuteStart() and onExecuteEnd(). Thats a great idea. Would it be 
possible to integrate in 0.9.0?

Great work on Proxool Bill. Keep the energy....

Thanks
Kiran

_________________________________________________________________
Check your PC for viruses with the FREE McAfee online computer scan.  
http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963

[Proxool-developer] Re: GetSQL() in connection

[Bill H. <bi...@lo...>]

Hi Vick,

On Fri, 2003-12-19 at 20:39, Vick Timber wrote:

> I was just curious to find if you have further considered putting a method 
> like getSQL() in the Connection to retrieve the SQL statement out of a 
> connection? That will help  a lot in the debugging process and will 
> certainly make Proxool far superior than the existing pools.

Ah, that's just the sort of comment that will make it appear in a
version soon :) I did do a little work on this, and then realised that
we'd have to cope with connections having multiple statements, and
statements doing multiple executions (I think). Not very hard but needs
a little thought. Scheduled for 0.9.0. 

Bill

[Proxool-developer] GetSQL() in connection

[Vick T. <vik...@ho...>]

Hi Bill,

I was just curious to find if you have further considered putting a method 
like getSQL() in the Connection to retrieve the SQL statement out of a 
connection? That will help  a lot in the debugging process and will 
certainly make Proxool far superior than the existing pools.

I am "looking forward" to the implementation ;-)


Vick

_________________________________________________________________
Make your home warm and cozy this winter with tips from MSN House & Home.  
http://special.msn.com/home/warmhome.armx

[Proxool-developer] [Fwd: Re: Cglib and JDK 1.2]

[Bill H. <bi...@lo...>]

Hi all,

-----Forwarded Message-----
> From: Chris Nokleberg <ch...@si...>
> To: Bill Horsman <bi...@lo...>
> Subject: Re: Cglib and JDK 1.2
> Date: Wed, 17 Dec 2003 17:31:50 -0800

> Hi Bill,
> 
> After looking at it some more I'm not sure we will necessarily want to
> make it compile with JDK 1.2. By using the -target and -bootclasspath
> options to javac we can compile a binary compatible with JDK 1.2 using a
> modern compiler. It will be kind of a pain to test compilation with the
> actual javac from JDK 1.2.

What do we think about this? I think I can live with JDK 1.2 users
having to compile with a modern compiler. The important thing is that it
runs flawlessly in JDK 1.2 environments. Most JDK 1.2 users are
restrained by their production boxes not their development workstations.
And it's not like they *have* to use the source distribution.

I'd like to do some testing with Jar file built with JDK 1.4 running in
a 1.2 environment. Unless anybody out there is already doing that?

Cheers,
Bill

[Proxool-developer] Proxool 0.8.3, Hibernate 2.1.1 and JDK 1.2

[Bill H. <bi...@lo...>]

Hi all,

Hibernate 2.1.1 was released today and it includes Proxool 0.8.3 and the
latest Cglib code (2.0RC2, although I haven't checked that). This means
that Hibernate should be happily using Proxool again.

However, I didn't notice (silly me) that Cglib 2.0RC2 no longer compiles
with JDK 1.2. I have just made a patch to our version of this code (we
repackaged Cglib within Proxool in 0.8.3) so that it now does. This is
tagged in cvs as v-0-8-3-PATCH-JDK12.

I'm waiting to see what happens with Cglib and its progress towards 2.0
final before deciding whether to release Proxool 0.8.4. One of these
options, probably:

1) Cglib release 2.0 but it doesn't include the changes necessary for
JDK 1.2 - we contemplate releasing 0.8.4 with our patched copy of Cglib.

2) Cglib release 2.0 quite soon and it does compile under JDK 1.2 and we
release 0.8.4 soon after

3) Cglib don't release anything for a while and people start complaining
that Proxool doesn't compile under JDK 1.2 and we release 0.8.4 with our
patched copy of Cglib.

Cheers,
Bill

[Proxool-developer] Re: [Proxool-user] Debugging proxool Error

[Bill H. <bi...@lo...>]

Hi Kiran,

On Mon, 2003-12-15 at 22:29, Kiran T wrote:

> I tried to switch over my existing DB pool( called pool123 below) to Proxool 

What were you using before, out of interest?

> 0.8.2 in my web application running JRun3.0 . Now I seem to get the weird 
> error message during OnDeath() method call in the ConnectionListenerIF . It 
> reads

Have you implemented a ConnectionListenerIF? If so, can you post the
code?

> java.sql.SQLException: [Microsoft][SQLServer 2000 Driver for 
> JDBC][SQLServer]sp_cursorclose: The cursor identifier value provided 
> (abcdf5a) is not valid.

This looks very much like something trying to use a ResultSet that
either belongs to another Statement or has been closed.

> For all the connection open, I have close() method but I have not closed 
> Statement/PreparedStatement( assuming proxoll does). 

It does.

> There might be places 
> in the  code where Resultset might not have been closed. 

Technically, this is the delegate driver's responsibility (Your
SQLServer driver, that is). I don't think it can do any harm to close
ResultSets and Statements explicitly in your code.

Cheers,
Bill Horsman

[Proxool-developer] Debugging proxool Error

[Kiran T <mah...@ho...>]

Hi Bill,

I tried to switch over my existing DB pool( called pool123 below) to Proxool 
0.8.2 in my web application running JRun3.0 . Now I seem to get the weird 
error message during OnDeath() method call in the ConnectionListenerIF . It 
reads

java.sql.SQLException: [Microsoft][SQLServer 2000 Driver for 
JDBC][SQLServer]sp_cursorclose: The cursor identifier value provided 
(abcdf5a) is not valid.

"[ERROR] pool123 - -#0004 encountered errors during closure:  
<java.sql.SQLException: [Microsoft][SQLServer 2000 Driver for 
JDBC][SQLServer]sp_cursorclose: The cursor identifier value provided 
(abcdf5a) is not valid.>

After few instances of these errors I then get an error "Exception occurred 
during fill/append.[Microsoft][SQLServer 2000 Driver for 
JDBC][SQLServer]Could not allocate new page for database 'tempdb'. There are 
no more pages available in filegroup DEFAULT. Space can be created by 
dropping objects, adding additional files, or allowing file growth.
Increasing the cache could not solve this problem.

For all the connection open, I have close() method but I have not closed 
Statement/PreparedStatement( assuming proxoll does). There might be places 
in the  code where Resultset might not have been closed. Similarly, I have 
used selectMethod as cursor irrespective of Transaction or non trnasactional 
query.

Any insight is highly appreciated.

Thanks
Kiran

_________________________________________________________________
Don’t worry if your Inbox will max out while you are enjoying the holidays.  
Get MSN Extra Storage!  http://join.msn.com/?PAGE=features/es

[Proxool-developer] Proxool 0.8.3 is Released

[Bill H. <bi...@lo...>]

Proxool 0.8.3 has been released. From the change log:

- Switched to Cglib 2.0 and ASM 1.4. Repackaged them both to
org.logicalcobwebs.clgib and org.logicalcobwebs.asm respectively. This
means that you no longer need to include cglib in your classpath. (Note:
other projects may still require it - the repackaged version is just for
use by Proxool). This removes our dependency on any version of Cglib or
ASM and avoids any potential Jar conflicts. 

- Made house keeper connection test more robust - credit John Hume 

- Removed JDK 1.4 specific code in ShutdownHook 

Excuse the cross posting, a significant proportion of you only belong to
one of the above lists.

Regards,
Bill Horsman

[Proxool-developer] Proxool 0.8.3 is coming

[Bill H. <bi...@lo...>]

Hi all,

This is a summary of some email conversations with Cglib and Hibernate
over the past week or so. Hibernate 2.1 has been released, but with
Cglib 2.0RC1 instead of Cglib 2.0final (which is due this weekend).
Unfortunately, this means that you can't use Proxool with Hibernate 2.1.

Hibernate 2.1 shipped with a patched version of Proxool 0.8.2, but it
still has problems with Cglib 2.0RC1.

The upshot of all this is that we have decided to repackage Cglib inside
Proxool (under org.logicalcobwebs.cglib) and make a new release of
Proxool 0.8.3 this weekend. This will make Proxool completely
independent of other versions of Cglib that might be in the user's
environment. And this means that Proxool 0.8.3 will work with Hibernate
2.0 and 2.1.

Cheers,
Bill Horsman

Re: [Proxool-developer] Connection release latency when using proxool with PostgreSQL

[Jean-Henry B. <ber...@ne...>]

bi...@lo... wrote:

>Hi Jean,
>
>On Thu, 2003-12-11 at 11:25, Jean-Henry Berevoescu wrote:
>
>  
>
>>>What's the URL you are using to ask for connections to the database? Are
>>>you sure you are using Proxool? Is it possible that you are asking the
>>>Postgresql Driver directly?
>>>      
>>>
>  
>  
>
>>I think you are right with this one - but then, why is it working all
>>right with Informix?
>>    
>>
>
>I suspect that you are bypassing Proxool with Informix too.
>
>  
>
>>If I have something like:
>>
>>jdbc:postgresql://host:port/dbName:user=...;password=...
>>
>>what would be the  URL for Proxool? 
>>    
>>
>
>The URL is simply:
>
>  proxool.globexplorer_0
>
>Regards,
>Bill Horsman
>  
>
Bingo! It is working all right now!
I'd have to go back and analyze the Informix case also. I think you are 100%
right: I was bypassing it in that case also.
Looking more into my code I realize that there was a part of connections
I was requesting that were indeed bypassing the Proxool mechanism
and this was the actual problem. I fixed it and now everything seems to
be back on track. Thank you so much!

Best regards,
Jean

Re: [Proxool-developer] Connection release latency when using proxool with PostgreSQL

[Bill H. <bi...@lo...>]

Hi Jean,

On Thu, 2003-12-11 at 11:25, Jean-Henry Berevoescu wrote:

> > What's the URL you are using to ask for connections to the database? Are
> > you sure you are using Proxool? Is it possible that you are asking the
> > Postgresql Driver directly?
  
> I think you are right with this one - but then, why is it working all
> right with Informix?

I suspect that you are bypassing Proxool with Informix too.

> If I have something like:
> 
> jdbc:postgresql://host:port/dbName:user=...;password=...
> 
> what would be the  URL for Proxool? 

The URL is simply:

  proxool.globexplorer_0

Regards,
Bill Horsman

Re: [Proxool-developer] Connection release latency when using proxool with PostgreSQL

[Jean-Henry B. <ber...@ne...>]

bi...@lo... wrote:

>Hi Jean,
>
>On Thu, 2003-12-11 at 11:02, Jean-Henry Berevoescu wrote:
>
>  
>
>>The weird part comes after that: when I do a server request (which
>>usually asks for around 4 connections
>>from proxool) no lines like these appear (which would mean, I guess,
>>that no new connections are
>>created by proxool) 
>>    
>>
>
>What's the URL you are using to ask for connections to the database? Are
>you sure you are using Proxool? Is it possible that you are asking the
>Postgresql Driver directly?
>
>Regards,
>Bill
>
>  
>
I think you are right with this one - but then, why is it working all 
right with
Informix?
If I have something like:

jdbc:postgresql://host:port/dbName:user=...;password=...

what would be the  URL for Proxool? Something like this:

jdbc:proxool://host:port etc?

But I don't have a real host port and everything for proxool. When I 
registered
proxool I did all this by setting its properties.

Thanks,
Jean

Re: [Proxool-developer] Connection release latency when using proxool with PostgreSQL

[Bill H. <bi...@lo...>]

Hi Jean,

On Thu, 2003-12-11 at 11:02, Jean-Henry Berevoescu wrote:

> The weird part comes after that: when I do a server request (which
> usually asks for around 4 connections
> from proxool) no lines like these appear (which would mean, I guess,
> that no new connections are
> created by proxool) 

What's the URL you are using to ask for connections to the database? Are
you sure you are using Proxool? Is it possible that you are asking the
Postgresql Driver directly?

Regards,
Bill

Re: [Proxool-developer] Connection release latency when using proxool with PostgreSQL

[Jean-Henry B. <ber...@ne...>]

I redid the tests in verbose mode and I notice that at init time I get 
the usual lines like:

4220 [Prototyper] DEBUG org.logicalcobwebs.proxool.globexplorer_0  - 
000000 (00/24/00) - Connection #24 (719f1f) created to achieve minimum 
of 30 = AVAILABLE -> 
jdbc:postgresql://pogo:5434/gex_runtime:user=postgres;password=post.gres 
(185ad79) by thread Prototyper

There are 30 of these (as I said the min number of conns is 30) - 
nothing unusual here.

The weird part comes after that: when I do a server request (which 
usually asks for around 4 connections
from proxool) no lines like these appear (which would mean, I guess, 
that no new connections are
created by proxool) but doing netstat I notice that the number of 
connections is N+4 now, and the
number of active ones remains N - the others are in TIME_WAIT mode and 
are cleaned when the
sweeper kicks in.
Then the question would be: what's with these additional connections I see?
(in time, if I do a stress test, the number of connections raises to 
hundreds or even thousands, but the
pattern remains: the active ones are as many as I expect and the bulk 
are connections in TIME_WAIT
state)

Thanks,
Jean


Bill Horsman wrote:

>Hi Jean,
>
>On Thu, 2003-12-11 at 00:16, Jean-Henry Berevoescu wrote:
>
>  
>
>>The traces are not showing anything unusual, but I noticed another
>>thing:
>>- I start with minumumConnectionCount set to, say, 30. Proxool is
>>correctly
>>building 30, but the first time I need a connection it builds another
>>one, and
>>another one and so on, every time a need a connection from the pool.
>>So it looks like the reusability is not working.
>>Again, it was working all right with Informix, but not with
>>PostgreSQL.
>>    
>>
>
>I'd like to see more of the log to find out what's going on there. It
>should show those initial 30 connections being built. And then it will
>show what happens when you ask for your first connection to use. We can
>then see whether it is reusing one of those initial 30 or making a new
>one.
>
>  
>
>>987 [Prototyper] DEBUG org.logicalcobwebs.proxool.globexplorer_0  -
>>org.postgresql.Driver does not support getHoldability. Proxool doesn't
>>mind.
>>988 [Prototyper] DEBUG org.logicalcobwebs.proxool.globexplorer_0  -
>>org.postgresql.Driver does not support setHoldability. Proxool doesn't
>>mind.
>>    
>>
>[snip]
>  
>
>>What exactly "Holdability" means? I noticed I have it with Informix,
>>but not here.
>>    
>>
>
>It's part of the JDBC API, but not all Drivers support it. Like the log
>says, Proxool doesn't mind either way. You'll find that your Informix
>driver does support it whilst your Postgresql one doesn't.
>
>  
>
>>Also, how one can programatically set the connection to be readOnly? I
>>always use it as readOnly in what I try to do here and I noticed
>>proxool is setting it to default = false.
>>    
>>
>
>Proxool is simply remembering the readOnly state of new connections. The
>default state is false (that is, writable). That is how the connections
>are supplied from the Postgresql Driver. Proxool resets all these states
>to their original values when you return a connection to the pool.
>Otherwise, if you changed something it would remain at that new value
>when you asked for another connection and that might not be what you
>want. So: every time you ask Proxool for a connection it comes in the
>sames state as when it arrived, fresh from the Postgresql Driver.
>
>Regards,
>Bill Horsman
>
>
>
>-------------------------------------------------------
>This SF.net email is sponsored by: SF.net Giveback Program.
>Does SourceForge.net help you be more productive?  Does it
>help you create better code?  SHARE THE LOVE, and help us help
>YOU!  Click Here: http://sourceforge.net/donate/
>_______________________________________________
>Proxool-developer mailing list
>Pro...@li...
>https://lists.sourceforge.net/lists/listinfo/proxool-developer
>
>  
>

Re: [Proxool-developer] Connection release latency when using proxool with PostgreSQL

[Jean-Henry B. <ber...@ne...>]

bi...@lo... wrote:

>Hi Jean,
>
>On Thu, 2003-12-11 at 00:16, Jean-Henry Berevoescu wrote:
>
>  
>
>>The traces are not showing anything unusual, but I noticed another
>>thing:
>>- I start with minumumConnectionCount set to, say, 30. Proxool is
>>correctly
>>building 30, but the first time I need a connection it builds another
>>one, and
>>another one and so on, every time a need a connection from the pool.
>>So it looks like the reusability is not working.
>>Again, it was working all right with Informix, but not with
>>PostgreSQL.
>>    
>>
>
>I'd like to see more of the log to find out what's going on there. It
>should show those initial 30 connections being built. And then it will
>show what happens when you ask for your first connection to use. We can
>then see whether it is reusing one of those initial 30 or making a new
>one.
>  
>
I have set the verbose and trace properties to "true", but I didn't see 
anything
reported by proxool when the client-side is asking for a connection from 
the pool.
Is there a way to set the verbosity/debug level somehow?

>>987 [Prototyper] DEBUG org.logicalcobwebs.proxool.globexplorer_0  -
>>org.postgresql.Driver does not support getHoldability. Proxool doesn't
>>mind.
>>988 [Prototyper] DEBUG org.logicalcobwebs.proxool.globexplorer_0  -
>>org.postgresql.Driver does not support setHoldability. Proxool doesn't
>>mind.
>>    
>>
>[snip]
>  
>
>>What exactly "Holdability" means? I noticed I have it with Informix,
>>but not here.
>>    
>>
>
>It's part of the JDBC API, but not all Drivers support it. Like the log
>says, Proxool doesn't mind either way. You'll find that your Informix
>driver does support it whilst your Postgresql one doesn't.
>  
>
Indeed - the Informix driver has it. I understand it has something to do 
with the
number of rows in a ResultSet - or something in this line.

>>Also, how one can programatically set the connection to be readOnly? I
>>always use it as readOnly in what I try to do here and I noticed
>>proxool is setting it to default = false.
>>    
>>
>
>Proxool is simply remembering the readOnly state of new connections. The
>default state is false (that is, writable). That is how the connections
>are supplied from the Postgresql Driver. Proxool resets all these states
>to their original values when you return a connection to the pool.
>Otherwise, if you changed something it would remain at that new value
>when you asked for another connection and that might not be what you
>want. So: every time you ask Proxool for a connection it comes in the
>sames state as when it arrived, fresh from the Postgresql Driver.
>  
>
I will keep setting the connection to readOnly when I ask for one. No 
problem.

Thanks,
Jean


>Regards,
>Bill Horsman
>
>
>
>-------------------------------------------------------
>This SF.net email is sponsored by: SF.net Giveback Program.
>Does SourceForge.net help you be more productive?  Does it
>help you create better code?  SHARE THE LOVE, and help us help
>YOU!  Click Here: http://sourceforge.net/donate/
>_______________________________________________
>Proxool-developer mailing list
>Pro...@li...
>https://lists.sourceforge.net/lists/listinfo/proxool-developer
>  
>

Re: [Proxool-developer] Connection release latency when using proxool with PostgreSQL

[Bill H. <bi...@lo...>]

Hi Jean,

On Thu, 2003-12-11 at 00:16, Jean-Henry Berevoescu wrote:

> The traces are not showing anything unusual, but I noticed another
> thing:
> - I start with minumumConnectionCount set to, say, 30. Proxool is
> correctly
> building 30, but the first time I need a connection it builds another
> one, and
> another one and so on, every time a need a connection from the pool.
> So it looks like the reusability is not working.
> Again, it was working all right with Informix, but not with
> PostgreSQL.

I'd like to see more of the log to find out what's going on there. It
should show those initial 30 connections being built. And then it will
show what happens when you ask for your first connection to use. We can
then see whether it is reusing one of those initial 30 or making a new
one.

> 987 [Prototyper] DEBUG org.logicalcobwebs.proxool.globexplorer_0  -
> org.postgresql.Driver does not support getHoldability. Proxool doesn't
> mind.
> 988 [Prototyper] DEBUG org.logicalcobwebs.proxool.globexplorer_0  -
> org.postgresql.Driver does not support setHoldability. Proxool doesn't
> mind.
[snip]
> What exactly "Holdability" means? I noticed I have it with Informix,
> but not here.

It's part of the JDBC API, but not all Drivers support it. Like the log
says, Proxool doesn't mind either way. You'll find that your Informix
driver does support it whilst your Postgresql one doesn't.

> Also, how one can programatically set the connection to be readOnly? I
> always use it as readOnly in what I try to do here and I noticed
> proxool is setting it to default = false.

Proxool is simply remembering the readOnly state of new connections. The
default state is false (that is, writable). That is how the connections
are supplied from the Postgresql Driver. Proxool resets all these states
to their original values when you return a connection to the pool.
Otherwise, if you changed something it would remain at that new value
when you asked for another connection and that might not be what you
want. So: every time you ask Proxool for a connection it comes in the
sames state as when it arrived, fresh from the Postgresql Driver.

Regards,
Bill Horsman

Re: [Proxool-developer] Connection release latency when using proxool with PostgreSQL

[Jean-Henry B. <ber...@ne...>]

The traces are not showing anything unusual, but I noticed another thing:
- I start with minumumConnectionCount set to, say, 30. Proxool is correctly
building 30, but the first time I need a connection it builds another 
one, and
another one and so on, every time a need a connection from the pool.
So it looks like the reusability is not working.
Again, it was working all right with Informix, but not with PostgreSQL.

The trace shows something like this when I start:
==========================================================================
0 [main] INFO org.logicalcobwebs.proxool.ProxoolFacade  - Proxool 0.8.2 
(05-Nov-2003 10:48)
39 [main] DEBUG org.logicalcobwebs.proxool.globexplorer_0  - Recognised 
proxool property: proxool.driver=org.postgresql.Driver
39 [main] DEBUG org.logicalcobwebs.proxool.globexplorer_0  - Recognised 
proxool property: 
proxool.url=jdbc:postgresql://pogo:5434/gex_runtime:user=postgres;password=post.gres
41 [main] DEBUG org.logicalcobwebs.proxool.globexplorer_0  - Recognised 
proxool property: proxool.minimum-connection-count=30
42 [main] DEBUG org.logicalcobwebs.proxool.globexplorer_0  - Recognised 
proxool property: proxool.maximum-active-time=4000
42 [main] DEBUG org.logicalcobwebs.proxool.globexplorer_0  - Delegating 
property to driver: user=postgres
43 [main] DEBUG org.logicalcobwebs.proxool.globexplorer_0  - Recognised 
proxool property: proxool.maximum-connection-lifetime=60000000
43 [main] DEBUG org.logicalcobwebs.proxool.globexplorer_0  - Recognised 
proxool property: proxool.house-keeping-sleep-time=600000
43 [main] DEBUG org.logicalcobwebs.proxool.globexplorer_0  - Delegating 
property to driver: password=********
44 [main] DEBUG org.logicalcobwebs.proxool.globexplorer_0  - Recognised 
proxool property: proxool.verbose=true
44 [main] DEBUG org.logicalcobwebs.proxool.globexplorer_0  - Recognised 
proxool property: proxool.trace=true
45 [main] DEBUG org.logicalcobwebs.proxool.globexplorer_0  - Recognised 
proxool property: proxool.house-keeping-test-sql=select count(*) from 
sid_content_four;
45 [main] DEBUG org.logicalcobwebs.proxool.globexplorer_0  - Recognised 
proxool property: proxool.maximum-connection-count=1000
207 [main] DEBUG org.logicalcobwebs.proxool.ShutdownHook  - Registered 
shutdownHook
226 [main] DEBUG org.logicalcobwebs.proxool.HouseKeeperController  - 
Registering 'globexplorer_0' house keeper
244 [main] DEBUG org.logicalcobwebs.proxool.HouseKeeperController  - 
Starting a house keeper thread
247 [HouseKeeper] INFO org.logicalcobwebs.proxool.globexplorer_0  - 
Proxool statistics legend: "s - r  (a/t/o)" > s=served, r=refused (only 
shown if non-zero), a=active, t=total, o=offline (being tested)
256 [HouseKeeper] DEBUG org.logicalcobwebs.proxool.globexplorer_0  - 
000000 (00/00/00) - House keeping triggerSweep done
671 [Prototyper] DEBUG org.logicalcobwebs.proxool.globexplorer_0  - 
Remembering default value: isReadOnly() = false
987 [Prototyper] DEBUG org.logicalcobwebs.proxool.globexplorer_0  - 
org.postgresql.Driver does not support getHoldability. Proxool doesn't mind.
988 [Prototyper] DEBUG org.logicalcobwebs.proxool.globexplorer_0  - 
org.postgresql.Driver does not support setHoldability. Proxool doesn't mind.
993 [Prototyper] DEBUG org.logicalcobwebs.proxool.globexplorer_0  - 
Remembering default value: getTransactionIsolation() = 2
995 [Prototyper] DEBUG org.logicalcobwebs.proxool.globexplorer_0  - 
Remembering default value: getTypeMap() = null
996 [Prototyper] DEBUG org.logicalcobwebs.proxool.globexplorer_0  - 
Remembering default value: getCatalog() = gex_runtime
==========================================================================

What exactly "Holdability" means? I noticed I have it with Informix, but 
not here.
Also, how one can programatically set the connection to be readOnly? I 
always use it as readOnly in what I try to do here and I noticed
proxool is setting it to default = false.
All the other properties are set in my server with something like:

info.setProperty("proxool.maximum-connection-count", connCount);

Thanks and best regards,
Jean



Jean-Henry Berevoescu wrote:

>
>
> bi...@lo... wrote:
>
>>Hi Jean,
>>
>>On Wed, 2003-12-10 at 11:52, Jean-Henry Berevoescu wrote:
>>
>>  
>>
>>>There is a problem in this new setup: the unused connections are released
>>>at a slower pace
>>>    
>>>
>>
>>Can you clarify what you mean by "released"? The life cycle is that a
>>connection is built, [served, returned,]* destroyed. Which bit is taking
>>a long time?
>>  
>>
> It is al acting in a weird way. What I do is build-use-returnToPool 
> but what I see
> is a lot of connections being created and a lot of them lingering in 
> TIME_WAIT
> state. It is not very clear why.
>
>>>The real problem is that I am expecting to reuse connections a lot 
>>>    
>>>
>>
>>By "reuse" do you mean serve the same connection lots of times? Serving
>>(and returning) a connection should be very quick.
>>  
>>
> Yes - the pattern of use is as I said before: create-use-returnToPool
>
>>Have you tried switching trace on? That will give you the performance of
>>your database calls. Could it be that the database is performing slowly
>>and that is what is causing many connections to be built?
>>
> I switched the tracing on for a short test, but I will do it again and 
> will take a
> closer look.
>
> Thanks,
> Jean
>
>>Regards,
>>Bill Horsman
>>
>>
>>
>>-------------------------------------------------------
>>This SF.net email is sponsored by: SF.net Giveback Program.
>>Does SourceForge.net help you be more productive?  Does it
>>help you create better code?  SHARE THE LOVE, and help us help
>>YOU!  Click Here: http://sourceforge.net/donate/
>>_______________________________________________
>>Proxool-developer mailing list
>>Pro...@li...
>>https://lists.sourceforge.net/lists/listinfo/proxool-developer
>>  
>>

Re: [Proxool-developer] Connection release latency when using proxool with PostgreSQL

[Jean-Henry B. <ber...@ne...>]

bi...@lo... wrote:

>Hi Jean,
>
>On Wed, 2003-12-10 at 11:52, Jean-Henry Berevoescu wrote:
>
>  
>
>>There is a problem in this new setup: the unused connections are released
>>at a slower pace
>>    
>>
>
>Can you clarify what you mean by "released"? The life cycle is that a
>connection is built, [served, returned,]* destroyed. Which bit is taking
>a long time?
>  
>
It is al acting in a weird way. What I do is build-use-returnToPool but 
what I see
is a lot of connections being created and a lot of them lingering in 
TIME_WAIT
state. It is not very clear why.

>>The real problem is that I am expecting to reuse connections a lot 
>>    
>>
>
>By "reuse" do you mean serve the same connection lots of times? Serving
>(and returning) a connection should be very quick.
>  
>
Yes - the pattern of use is as I said before: create-use-returnToPool

>Have you tried switching trace on? That will give you the performance of
>your database calls. Could it be that the database is performing slowly
>and that is what is causing many connections to be built?
>
I switched the tracing on for a short test, but I will do it again and 
will take a
closer look.

Thanks,
Jean

>
>Regards,
>Bill Horsman
>
>
>
>-------------------------------------------------------
>This SF.net email is sponsored by: SF.net Giveback Program.
>Does SourceForge.net help you be more productive?  Does it
>help you create better code?  SHARE THE LOVE, and help us help
>YOU!  Click Here: http://sourceforge.net/donate/
>_______________________________________________
>Proxool-developer mailing list
>Pro...@li...
>https://lists.sourceforge.net/lists/listinfo/proxool-developer
>  
>

Re: [Proxool-developer] Connection release latency when using proxool with PostgreSQL

[Bill H. <bi...@lo...>]

Hi Jean,

On Wed, 2003-12-10 at 11:52, Jean-Henry Berevoescu wrote:

> There is a problem in this new setup: the unused connections are released
> at a slower pace

Can you clarify what you mean by "released"? The life cycle is that a
connection is built, [served, returned,]* destroyed. Which bit is taking
a long time?

> Is there any known issue with proxool when connecting to a PostgreSQL DB 

No.

> The real problem is that I am expecting to reuse connections a lot 

By "reuse" do you mean serve the same connection lots of times? Serving
(and returning) a connection should be very quick.

Have you tried switching trace on? That will give you the performance of
your database calls. Could it be that the database is performing slowly
and that is what is causing many connections to be built?

Regards,
Bill Horsman

[Proxool-developer] Connection release latency when using proxool with PostgreSQL

[Jean-Henry B. <ber...@ne...>]

Hi,
I am using proxool for some time in my project and I am very pleased with
the performance and functionality I get from it. I mainly used my server to
connect to an Informix database, but I recently did a port to handle 
PostgreSQL
connections also.
There is a problem in this new setup: the unused connections are released
at a slower pace, much slower than in the Informix case. Is there any
known issue with proxool when connecting to a PostgreSQL DB in the
line I am describing here?
The real problem is that I am expecting to reuse connections a lot and
if they are released with big delays, this generates a lot of new 
connections.
The effect is that everything slows down and at certain points I reach
system limitations (max connection number for example)

Best regards,
Jean

RE: [Proxool-developer] finally block throws NPE

[Bill H. <bi...@lo...>]

Hi John,

On Tue, 2003-12-09 at 15:09, Hume, John - NA US HQ Delray wrote:

> Apologies for the fomat being screwy, 

None necessary. That looks like a good patch. I'll commit it to CVS
later today.

Thanks again,
Bill Horsman

RE: [Proxool-developer] finally block throws NPE

[Hume, J. - NA US HQ D. <JH...@my...>]

Apologies for the fomat being screwy, but I'm not familiar with CVS or diff.
Thanks!  
-john.

$ diff -c1 ConnectionPool.java_old ConnectionPool.java
*** ConnectionPool.java_old Tue Dec  9 08:54:24 2003
--- ConnectionPool.java Tue Dec  9 09:05:56 2003
***************
*** 287,292 ****
              } finally {
!                 try {
!                     s.close();
!                 } catch (SQLException e) {
!                     // Ignore
                  }
--- 287,294 ----
              } finally {
!                 if( s!=null ) {
!                     try {
!                         s.close();
!                     } catch (SQLException e) {
!                         // Ignore
!                     }
                  }




-----Original Message-----
From: Bill Horsman [mailto:bi...@lo...]
Sent: Tuesday, December 09, 2003 3:51 AM
To: Hume, John - NA US HQ Delray
Cc: 'Pro...@li...'
Subject: Re: [Proxool-developer] finally block throws NPE


Hi John,

On Mon, 2003-12-08 at 21:58, Hume, John - NA US HQ Delray wrote:

> I added "if(s!=null)" to the beginning of line 288 to get around this.
Let
> me know if you'd like me to submit this as a bug and/or patch on
> sourceforge.  Thanks.  

Thanks for the contribution. Yes please, either submit a patch on
sourceforge or just email me the diff and I will be happy to incorporate
the fix into our next release (which isn't that far away).

Regards,
Bill Horsman

Re: [Proxool-developer] finally block throws NPE

[Bill H. <bi...@lo...>]

Hi John,

On Mon, 2003-12-08 at 21:58, Hume, John - NA US HQ Delray wrote:

> I added "if(s!=null)" to the beginning of line 288 to get around this.  Let
> me know if you'd like me to submit this as a bug and/or patch on
> sourceforge.  Thanks.  

Thanks for the contribution. Yes please, either submit a patch on
sourceforge or just email me the diff and I will be happy to incorporate
the fix into our next release (which isn't that far away).

Regards,
Bill Horsman

[Proxool-developer] finally block throws NPE

[Hume, J. - NA US HQ D. <JH...@my...>]

Greetings ... I'm new to Proxool so forgive me if this has been discussed or
addressed, but I didn't see anything in the archives that seemed related.  

In ConnectionPool.testConnection(ProxyConnectionIF), if the
createStatement() on line 278 throws anything (for example because the
connection was unexpectedly closed), then s never gets assigned, so the
s.close() in the finally block throws a NullPointerException, preventing
testConnection from returning false.  

I know this should be a rare occurence, but it's currently happening to us
with great frequency for some unknown reason.  (We were blaming Jakarta
DBCP, which is why I was working on implementing proxool, but now it looks
like our problem lies elsewhere.)  

I added "if(s!=null)" to the beginning of line 288 to get around this.  Let
me know if you'd like me to submit this as a bug and/or patch on
sourceforge.  Thanks.  

-john.
_____________
john d. hume
 561-266-6256
 john.hume at cruisesonly.com

RE: [Proxool-developer] [Fwd: DataSource and JDK 1.2]

[Bill H. <bi...@lo...>]

Hi Martin!

On Mon, 2003-12-08 at 13:40, Martin Crawford wrote:

> I've just woken up from a long coma

Glad to see you awake now! Welcome back.

> Can someone elaborate what are the
> requirements for Hibernate and issues with cglib?

Cglib 2 isn't backwards compatible with Cglib 1. Different API.
Hibernate 2.1 plans to ship with Cglib 2. It also ships Proxool so we
need to provide a Cglib2 compatible version. 

We hope to be able to support Cglib 1 and 2 with the same binary - that
way we don't have to choose between maintaining two versions or cutting
off people who still use Cglib 1. 

Hibernate and Spring are the main users of Cglib at the moment.

Depending on timing, we will either release 0.8.3 to support this, or
include it in the upcoming 0.9.0 release (with Christian's
JNDI/DataSource code :)

That's all for now. Got to get a flight... Should really leave time to
say goodbye to my family...

Bill Horsman