Hey,

Yes. I am however synchronizing. I basically made my own connection
manager. Then using synchronization to ensure that only one thread
gets a reference to a connection. When it does it calls the my
connection objec'ts setActive(true). Other threads would do something
like "getActive() == false".

My only explanation is that I'm not synchronizing properly somewhere.

Q

On Thu, Jun 19, 2008 at 11:44 PM, Oleg Kalnichevski <[EMAIL PROTECTED]> wrote:
> Quintin Beukes wrote:
>>
>> I'm not sure why accessing it from different threads would cause this
>> to happen? Could you explain this please.
>>
>
> Quintin
>
> I am only guessing here. All I am saying HTTP connection objects are not
> threading safe. The only method one can safely can from a different thread
> is #shutdown.
>
> I would certainly help if you managed to reproduce the problem with a test
> case.
>
> Cheers
>
> Oleg
>
>
>> It is definitely a concurrency issue. I am still narrowing them all down.
>>
>> thanks,
>> Q
>>
>> On Thu, Jun 19, 2008 at 12:02 AM, Oleg Kalnichevski <[EMAIL PROTECTED]>
>> wrote:
>>>
>>> On Wed, 2008-06-18 at 22:58 +0200, Quintin Beukes wrote:
>>>>
>>>> Hey,
>>>>
>>>> Is it possible for this method (DefaultHttpClientConnection.isOpen())
>>>> to return "true" when the socket is closed.
>>>>
>>> If you are not accessing the same connection instance from multiple
>>> threads, I do not see how this could happen.
>>>
>>> Please note only HttpConnection#shutdown method is expected to be
>>> threading safe.
>>>
>>> This sounds like a concurrency issue to me
>>>
>>> Oleg
>>>
>>>> I have a case where netstat doesn't return ANY open connections to the
>>>> target server, but isOpen returns true.
>>>>
>>>> It's not very easy to reproduce my problem, but I've been able to do
>>>> it whenever needed by using high concurrency. I'm not sure if isOpen()
>>>> ever returns true when there is no socket when it seems like my
>>>> problem isn't there. I've only checked when my problem is occurring,
>>>> and this seems very odd.
>>>>
>>>> Thanks
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: [EMAIL PROTECTED]
>>> For additional commands, e-mail: [EMAIL PROTECTED]
>>>
>>>
>>
>>
>>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
>



-- 
Quintin Beukes

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

Reply via email to