Gordon Sim wrote:
>
>
> So the crash was happening when you were manually accepting. Was autoAck
> set to 0 or left as the default value?
>
>
Yes crash occur, when calling messageAccept()
Left autoAck as default value. Or non-zero. doesn't matter
> Really? I would have expected that to
Gordon Sim wrote:
>
> If not can you try the attached test program? It will listen for a
> configurable number of messages on queues a, b and c (all bound to
> amq.fanout for convenience). I can consume 10 million messages from each
> queue without seeing any issue (neither crashes nor memor
i will try it out now.
thanks
--
View this message in context:
http://n2.nabble.com/Bad-allocation-error-tp4265473p4289738.html
Sent from the Apache Qpid users mailing list archive at Nabble.com.
-
Apache Qpid - AMQP Messagin
Gordon Sim wrote:
>
>
> The simplest way (in my view) is to pass in an instance of
> SubscriptionSettings to each SubscriptionManager::subscribe() call.
>
> However if the default settings work for you (and it sounds like they
> might?) then you don't even need to do that. Just call subscrib
Gordon Sim wrote:
>
> In your first mail you stated you were using ACQUIRE_MODE_NOT_ACQUIRED.
> Are you in fact using ACQUIRE_MODE_PRE_ACQUIRED? If so that is what I
> was suggesting. In that case you do not need to manually acquire at all.
>
Initially, i set to ACQUIRE_MODE_NOT_ACQUIRED and
Gordon Sim wrote:
>
>
> Maybe worth asking as an aside at this point why it is you want to
> subscribe in unacquired mode? Do you want to acquire every message you
> receive? If so the pre-acquired mode will do that more simply and
> efficiently.
>
>
All i need is to confirm each message
denny86 wrote:
>
> I have a similar solution [not tried yet] to my problem.
>
> http://n2.nabble.com/ATL-Project-td2944757.html
>
> But that is referring to an ATL project. mine is not ATL. anyways will try
> out what Steve says.
>
>
didnt work out
--
V
I have a similar solution [not tried yet] to my problem.
http://n2.nabble.com/ATL-Project-td2944757.html
But that is referring to an ATL project. mine is not ATL. anyways will try
out what Steve says.
hope i'm not spamming here. =^D
Regards,
Denny
--
View this message in context:
http://n2.na
Further research, i found that queue with min queue depth will dequeue the
message even if i dont acquire and accept those messages in that queue. How
is that possible? I didnt turn on auto delete.
To verify that if have not gone wrong somewhere, I checked the instances
of subscription class fo
Gordon Sim wrote:
>
> On 01/07/2010 01:19 PM, denny86 wrote:
>>
>>
>>
>> Gordon Sim wrote:
>>>
>>>
>>> Each subscription will have its own Subscription instance. You will need
>>> to acquire and accept messages rec
Gordon Sim wrote:
>
>
> Each subscription will have its own Subscription instance. You will need
> to acquire and accept messages received from each queue through the
> relevant Subscription object.
>
Yes... i am working on that already
Is session.sync(); required after sending Compl
denny86 wrote:
>
>
>
> Yes, got it.. Using session to acquire and accept messages would fail
> after sometime. But using Subscription Instance, i face no problem.
>
> Now enqueue/dequeue operations are working smooth.. No issues so far,
> 200,000 message
Gordon Sim wrote:
>
>
Yes, got it.. Using session to acquire and accept messages would fail after
sometime. But using Subscription Instance, i face no problem.
Now enqueue/dequeue operations are working smooth.. No issues so far,
200,000 messages processed without fail. ;)
Thank you Gordon
Gordon Sim wrote:
>
>
> They are on the Subscription class, you get one of those returned from
> SubscriptionManager::subscribe().
>
>
Yes i found it.. i misread ur early post.
i looked upon SubscriptionManager and not Subscription class... Going to try
that now..
tnx & regards,
Denny
--
Gordon Sim wrote:
>
> Try changing this to use the acquire() and accept() messages on the
> Subscription class (you get an instance of this returned when you call
> subscribe() on SubscriptionManager.
>
Sorry i can't find acquire() and accept() member functions within the
SubscriptionManag
Gordon Sim wrote:
>
> Sounds like your client process is running out of memory(?). This is
> using the 0.5 code, right?
>
Yes. I'm using 0.5 version.
I didnt figure out any memory leaks initially.
Gordon Sim wrote:
>
> Try changing this to use the acquire() and accept() messages on the
>
Hi,
I stumbled upon following error while listening to messages
Listening for messages [Press Ctrl+C to Quit]...2010-jan-07 13:45:13 error
FAIL
[2737 192.168.99.44:5672]: bad allocation (.\qpid\client\Connector.cpp:396)
2010-jan-07 13:45:13 warning Connection closed
Connection closed
I'm usin
Hi,
I found it.. Thanks. Will go through the code.
Regards,
Denny
Gordon Sim wrote:
>
> On 01/06/2010 11:12 AM, denny86 wrote:
>> Couldn't find it in the qpid truck [
>> https://svn.apache.org/repos/asf/qpid/trunk/qpid/ ].
>
> Its within the cpp tree:
&
o and it seems to be working
> fine.
>
> Michel
>
> - Original Message -
> From: "denny86"
> To: users@qpid.apache.org
> Sent: Wednesday, January 6, 2010 11:24:08 AM
> Subject: Re: Qpid persistent storage using Berkeley DB
>
>
> H
Hi,
Thanks for your reply. Well in fact, SQL will do. I'm not sticking to BDB
only. If you [or anyone here] could shed some light regarding this SQL
implementation, I would be grateful. :)
Thanks & Regards,
Denny
Robert Godfrey wrote:
>
> Hi Denny,
>
> 2010/1/6
Hi,
I'm using C++ Broker along with C++ Clients, both running in win32
platforms. I would like to implement Berkeley DB persistent storage. I have
seen an example of persistence library at
http://qpidcomponents.org/download.html#persistence QpidComponents . But the
same is implemented for li
21 matches
Mail list logo