[ 
https://issues.apache.org/jira/browse/CAMEL-16104?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Claus Ibsen resolved CAMEL-16104.
---------------------------------
    Fix Version/s:     (was: 3.4.6)
       Resolution: Fixed

Thanks for the detailed report

> ProducerCache does not close producers when cacheSize is 1 (potential memory 
> leak)
> ----------------------------------------------------------------------------------
>
>                 Key: CAMEL-16104
>                 URL: https://issues.apache.org/jira/browse/CAMEL-16104
>             Project: Camel
>          Issue Type: Bug
>          Components: camel-core
>    Affects Versions: 3.3.0, 3.7.1
>            Reporter: Benoit AVERTY
>            Assignee: Claus Ibsen
>            Priority: Major
>             Fix For: 3.7.2, 3.8.0
>
>
> I'm using a route with a dynamic endpoint using toD and a cache size of 1.
> In my case, the dynamic URI changes once a day. When the day changes, I want 
> a new producer to be created and the old one closed.
> However, the new producer is correctly created but the old one is not closed.
> When I use a cache size of 2, then when the day ends two new producers are 
> created and the two old ones are closed correctly.
> I've tracked down the problem to the last line of the ServicePool 
> constructor, where a LRU cache is created only if the cache size is greater 
> than one. This LRU cache is the one calling the evict method, so when there 
> is no cache there is no eviction on the ServicePool used by the ProducerCache.
> Using the debugger, i confirmed that the service pool contains two producers 
> even though its capacity is 1. I'm afraid this could be a memory leak.
> I'm not sure of the best way to resolve this, but it's a problem for me 
> because if the producer is not closed, then an operation I need is note 
> performed (I'm using the HDFS component, and it needs to rename a file when 
> the producer is closed. Currently, the file is not renamed and I'm afraid 
> some data is not flushed to the server)
> I hope I was clear, please do ask if there is something unclear in my 
> description of the problem.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

Reply via email to