activemq-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Gerdes, Mike" <Mike.Ger...@airbus.com>
Subject AW: still strange memory problems
Date Fri, 02 Jun 2006 10:26:00 GMT

After a bit looking into the code, I wonder if I understood the concept right, that there
is a referencecounter that is raised for each consumer and then decresed by one when a consumer
gets a message. The memory if freed if the referencecounter reaches 0.
So now after reading the counter out and send it to the console, I noticed that it normally
doesn't get set back to 0, when there is more then one consumer.

The method decreaseUsage is never called in the case of more then one consumer and only increased.

Is this a correct obversion and is this the core of the memory problem I encounter?

-----Urspr√ľngliche Nachricht-----
Von: Gerdes, Mike
Gesendet: Freitag, 2. Juni 2006 09:59
An: activemq-users@geronimo.apache.org
Betreff: still strange memory problems




Hi,

I still have memory problems with AMQ. These problems mostly only happen, when I have more
then one consumer listening to a queue and more then one producer sending messages at a high
rate to the broker. These messages start at a size of 256 byte and are send for 30sec, then
they double their size, this is reapeated some time. The producers read the messages all the
time till they recieve no messages for 30sec.
I have tried this in a few configuration and always hit this problem. First everything locally
on one pc with 512mb ram and a celeron 2GHz, then with two pcs of the same configuration and
divided the senders and producers in any configuration on them, then two brokers locally.
I use AMQ 10.05.2006 and have also tried AMQ RC2.

Everytime I observed that the memoryusage raises all the time and it doesn't get freed when
all the consumers and producers are shutdown. Also it happens that the memoryusage hits 100
and normally then the system stops to work, I get no error message or drop, the producers
just seem to block. The consumers then exit after 30seconds. This happens when the message
size is about 256kb, sometimes lower sometimes higher. With topics this problem doesn't occure.

What I have tried to solve this problem is, that I have set the memory manager to use 100mb,
used mysql persistence adapter, user journaled adapter with 1gb available, used no persistence
adapter, persistence adapters with useasync=true.
Nothing solved this problem.

So now my question, is this a bug or is my machine for the case to weak? Am I the only one
who gets this kind of problem or are there others and have you found solutions?

mike

This mail has originated outside your organization, either from an external partner or the
Global Internet. Keep this in mind if you answer this message.



This mail has originated outside your organization,
either from an external partner or the Global Internet.
Keep this in mind if you answer this message.

This mail has originated outside your organization, either from an external partner or the
Global Internet. Keep this in mind if you answer this message.

Mime
View raw message