activemq-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Michael Feingold (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (AMQ-5762) Severe memory leak in the MQTT connector
Date Tue, 05 May 2015 22:10:59 GMT

    [ https://issues.apache.org/jira/browse/AMQ-5762?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14529385#comment-14529385
] 

Michael Feingold commented on AMQ-5762:
---------------------------------------

What you are saying is that the redundant subscribe creates the memory leak? I would expect
that when the same client tries to subscribe once more to the same topic the subscription
should be considered redundant and ignored. Anyway this is important info and we will run
the test without the subscribe to confirm that this is the case. 

I also think that the extra 230K per hard close is still a concern. We will run a few more
tests and I will keep you posted

> Severe memory leak in the MQTT connector
> ----------------------------------------
>
>                 Key: AMQ-5762
>                 URL: https://issues.apache.org/jira/browse/AMQ-5762
>             Project: ActiveMQ
>          Issue Type: Bug
>          Components: MQTT
>    Affects Versions: 5.10.1, 5.10.2, 5.11.0, 5.11.1
>            Reporter: Michael Feingold
>            Assignee: Timothy Bish
>            Priority: Critical
>
> When an MQTT connection is closed some memory is not released. In case of the soft close
(calling connection close on the client) memory leaks at the rate of 70K per connection close.
In case of the hard close (pulling the wire, or cutting off the network connection) the leak
goes at the rate of 300K per close. we also checked it on version 5.11 - same problem



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message