activemq-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Gary Tully (JIRA)" <>
Subject [jira] Resolved: (AMQ-2103) Memory leak when marshaling ActiveMQTextMessage to persistent store
Date Mon, 11 Oct 2010 19:18:40 GMT


Gary Tully resolved AMQ-2103.

       Resolution: Fixed
    Fix Version/s:     (was: 5.3.0)

patch reverted and new policy based solution applied in r1021466

The patch, setting the text attribute to null, was unprotected. Currently there are no destructive
operations, on a message, allowing it to be shared across topic consumers without synchronization,
which allows concurrent dispatch to be fast. This is important to maintain.

A new destination policy for queues, boolean attribute reduceMemoryFootprint has been introduced
to do meet this use case.
When true, the marshaled state of a message will be cleared once the message is persisted
in the store. This is a natural, contention free sync point.
Note:  The kahaDB concurrentStoreAndForwardQueues option (which introduces concurrency) and
the memory persistence adapter (which will not marshall) are not compatible with this new

Trevor, can you validate that this change meet your use case?

> Memory leak when marshaling ActiveMQTextMessage to persistent store
> -------------------------------------------------------------------
>                 Key: AMQ-2103
>                 URL:
>             Project: ActiveMQ
>          Issue Type: Bug
>          Components: Broker
>    Affects Versions: 5.0.0
>         Environment: ActiveMQ
>            Reporter: Trevor Pounds
>            Assignee: Gary Tully
>             Fix For: 5.5.0
>         Attachments: AMQ-2103.diff, Duplicate Message Data (Internal Marshalling).png,
heap_100_1MB_messages.png, jhat_ActiveMQTextMessage_0xe837a478.htm, jhat_ByteSequence_0xe837a5c0.htm,
> When an org.apache.activemq.command.ActiveMQTextMessage is marshaled into the persistence
store some portion of the messages are stored in memory (i.e. pending cursor/consumer dispatch
queue).  The messages stored in memory have the potential to cause the broker to run out of
memory because org.apache.activemq.command.ActiveMQTextMessage objects can store the data
twice, once in the 'text' field and once in the 'content' field.  Normally this isn't a problem
since the 'content' field is cleared when the message is being used in a client application
(i.e. by calling getText() clears content).  The problem occurs when a consumer is slow and
a large number of messages are sitting around on the broker in pending/dispatch memory space.
 The message is marshaled for the store and then persisted to disk and copied to pending memory
when space is available.
> This bug affects any ActiveMQ*Message object that does not clear its temporary data (i.e.
'text' field) once it has been marshaled.  When a message is marshaled we should null the
derived objects memory space once the data has been written to the parent object's 'content'

This message is automatically generated by JIRA.
You can reply to this email to add a comment to the issue online.

View raw message