Sorry for the false alarm - I can't reproduce it either now! That's wierd.
And, as you say Roman, it does try the stack.pop so shouldn't fail.

We actually found a way of retrieving the actual connection via the commons
pool, so we have reverted to using that. When we switched to use the
firebird pool, we were failing under load with couldn't get connection
errors, even with a very long timeout, which we don't see with commons
pool.

Oh, well, all's well that ends well! Again, sorry for the false alarm, and
thanks for all your help.

BTW, is there any update on the status of the table reservation issue,
Roman? Did you manage to find time to submit to dev?

cheers,

David




Steven Jardine
<steve@mjnservices To: Firebird-Java@yahoogroups.com
.com> cc:
Sent by: Subject: Re: [Firebird-Java] Pool bug?
Firebird-Java@yaho
ogroups.com


02/06/2006 03:59
PM
Please respond to
Firebird-Java






I just ran my own test and I cannot duplicate your problem, so we will
need a test case as Roman suggests to proceed any further.

Steve

Roman Rokytskyy wrote:

>> In here, repeated calls are made to the addConnection(stack) method.

This

>> will never return a connection, as the maxConnections has been reached,
>> and so it simply returns false, waits, and then tries again. It never
>> checks the stack, however, to see if any of the busy connections have

been

>> returned. Eventually it fails with a timeout error, even though
>> connections are sitting there waiting to be used.
>>
>
> Did you see 'stack.pop' below?
>
> I do not believe that you have descovered an error, since this situation

is

> checked in one of our test cases. But to be sure, please create a
> reproducable test case and post it either here or send me per email.
>
> Roman
>
>
>
>
> Yahoo! Groups Links
>
>
>
>
>
>

Yahoo! Groups Links