Michael,

What you are describing, if true, is a serious bug. My gut is that the
bug isn't with the GenericObjectPool code because I expect it would
have been reported by now.

Without code that demonstrates the bug (aka:unit test) it's hard to
know anything for sure but the first thing that comes to mind is:

GenericObjectPool is supposed to be thread-safe out of the box, if you
are adding any synchronization, such as calling pool.borrowObject from
with in your own synchronized block, then you could get the behavior
you describe. When a thread acquires a couple of locks by passing into
a synchronized block, calling Object.wait() only releases that one
lock and not all that have been acquired.  GenericObjectPool only
knows about it's internal locks and cannot release any of your code's
locks while it waits. This could create the inactivity you describe.

Also, make sure you are using the latest Pool release (1.3) and if you
can generate a javacore (`kill -3 javapid` or ctrl+\ on unix, or
ctrl+break on windows) you should be able to see where the Threads are
waiting to acquire locks.

Hope that helps some.

On 10/18/06, Michael Thomson <[EMAIL PROTECTED]> wrote:
Hi,

I have an app running in JBOSS

I am trying to use the GenericObjectPool.

I have two "user" objects and one "pool" object

MaxActive =1
pool.setWhenExhaustedAction(pool.WHEN_EXHAUSTED_BLOCK)

It appears that when one "user" object is blocked waiting on a
borrowObject() when the other "user" object attempts to returnObject(),
both "user" objects "hang" until the maxWait time is reached?

I can go into more detail if anyone thinks it will help.

Anyone seen this problem, or have any suspicions, is JBoss likely to be
interfering with anything here, anyone doing anything similar, it seems
like a very basic use case?

Thanks in Advance,
Michael

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]




--
Sandy McArthur

"He who dares not offend cannot be honest."
- Thomas Paine

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to