objectbridge-developers

[OJB-developers] Open prepared statements

[Ranjan B. <ran...@fr...>]

Hi --

	Since I'm new, I want to say that I'm quite impressed by
objectbridge:  it's certainly easing a lot of the tedium of writing jdbc
applications.  Congrats, really!

So here's what I'm running into:
	I'm running against Oracle 8i, 8.1.6

	My schema (and I'm trying to get it sufficiently reduced so I 
can share it) has an entity E with a collection attribute of entities E1
which further has a collection of entities E2.  Pretty standard
containment, with the repository set up so that loading E, loads up all
the E1 and E2 attributes.

	Basically it's

		 E +--<- E1 +--<- E2

	When the database got built to around 100 E2 entities I started
getting errors thatthe available cursors had been exhausted.

	Doing some investigation, I found that a prepared statement for
each E2 was being created, but they weren't being closed.  I put a quick
patch into ojb.broker.accesslayer.JdbcAccess to cache prepared statements,
and close them when the RsIterator is cleaning up.

	This solves the immediate problem, but it isn't a long term fix.

	Has this been experienced before?   Can I configure things differently so
I don't need the code change?

	Thanks!

	Ranjan Bagchi

Re: [OJB-developers] Open prepared statements

[Thomas M. <tho...@ho...>]

Hi Ranjan,

Ranjan Bagchi wrote:

> Hi --
> 
> 	Since I'm new, I want to say that I'm quite impressed by
> objectbridge:  it's certainly easing a lot of the tedium of writing jdbc
> applications.  Congrats, really!
> 


thanks !


<snip> 
> 	Doing some investigation, I found that a prepared statement for
> each E2 was being created, but they weren't being closed.  I put a quick
> patch into ojb.broker.accesslayer.JdbcAccess to cache prepared statements,
> and close them when the RsIterator is cleaning up.
> 
> 	This solves the immediate problem, but it isn't a long term fix.
> 
> 	Has this been experienced before?  


Yes, this is a known bug that is already fixed. Will be contained in the 
next release!

 Can I configure things differently so
> I don't need the code change?
> 



No, there is not much to configure here. OJB just must handle this 
correctly!


HTH,


Thomas

Re: [OJB-developers] Open prepared statements

[Eric B. <EBi...@sa...>]

I was experiencing the same problem as you are until I started using OJB
version 0.7.343. If you're using OJB by putting jars on your classpath,
ensure that you've only got the specific jar you want to be using in your
classpath. Fyi, I'm using the ODMG interface against Oracle 8i, 8.1.6 using
the 'thin' oracle JDBC driver.

- Eric

----- Original Message -----
Subject: [OJB-developers] Open prepared statements
<snip>

When the database got built to around 100 E2 entities I started
getting errors thatthe available cursors had been exhausted.

Doing some investigation, I found that a prepared statement for
each E2 was being created, but they weren't being closed.  I put a quick
patch into ojb.broker.accesslayer.JdbcAccess to cache prepared statements,
and close them when the RsIterator is cleaning up.

This solves the immediate problem, but it isn't a long term fix.

Has this been experienced before?   Can I configure things differently so
I don't need the code change?

<snip>

Re: [OJB-developers] Open prepared statements

[Ranjan B. <ran...@fr...>]

Eric, thanks --
 Yes, the new version's working just fine (after I got 7.310 out of my
classpath)

 Ranjan



On Fri, 22 Feb 2002, Eric Bieschke wrote:

> I was experiencing the same problem as you are until I started using OJB
> version 0.7.343. If you're using OJB by putting jars on your classpath,
> ensure that you've only got the specific jar you want to be using in your
> classpath. Fyi, I'm using the ODMG interface against Oracle 8i, 8.1.6 using
> the 'thin' oracle JDBC driver.
> 
> - Eric
> 
> ----- Original Message -----
> Subject: [OJB-developers] Open prepared statements
> <snip>
> 
> When the database got built to around 100 E2 entities I started
> getting errors thatthe available cursors had been exhausted.
> 
> Doing some investigation, I found that a prepared statement for
> each E2 was being created, but they weren't being closed.  I put a quick
> patch into ojb.broker.accesslayer.JdbcAccess to cache prepared statements,
> and close them when the RsIterator is cleaning up.
> 
> This solves the immediate problem, but it isn't a long term fix.
> 
> Has this been experienced before?   Can I configure things differently so
> I don't need the code change?
> 
> <snip>
> 
>