[
https://issues.apache.org/activemq/browse/CAMEL-2854?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=60350#action_60350
]
Claus Ibsen commented on CAMEL-2854:
------------------------------------
The LRUCache cannot be changed.
The endpoints should be safe to re-create again. And the cache keeps preventing
having Camel hooking up memory when end users uses many different endpoints.
So it keep the last 1000 endpoints around in the cache.
Its just the direct component which had an issue, storing state on the
endpoint. Which should have been stored on the component instead.
The seda component works this way btw.
> No consumers available on "direct://xyz" endpoint
> -------------------------------------------------
>
> Key: CAMEL-2854
> URL: https://issues.apache.org/activemq/browse/CAMEL-2854
> Project: Apache Camel
> Issue Type: Bug
> Components: camel-core
> Affects Versions: 2.2.0
> Environment: ActiveMQ 5.3.2 / Camel 2.2.0
> Reporter: Michael Chen
> Assignee: Claus Ibsen
> Priority: Critical
> Fix For: 2.4.0
>
>
> The following warning in log4j log shows up with no predictable reason:
> 2010-06-24 14:07:01,090 WARN
> [DefaultMessageListenerContainer-7:DirectProducer.java:42] No consumers
> available on endpoint: Endpoint[direct://UserCancelJobRequest] to process:
> Exchange[JmsMessage:
> RunnerBase{locationToRunId='DV02YcABTE43HTRPATEGCBFk0m0', jobrun_id=1234,
> requestOrigId='null', secretKey='null', replyToQueueOrTopic='null',
> corrolationId='null', requestCreationTime=Thu Jun 24 14:07:01 PDT 2010}]
> The endpoint [direct://UserCancelJobRequest] could be a number of different
> endpoints in our application and all of them are of "direct:" type. When
> this happens, all messages that run through that endpoint will fail, and it
> will never recover unless the application is restarted.
> Also, if I restart the application and use that route immediately, that
> warning will not occur and the request will finish successfully. It seems to
> happen to routes that have not been used for a while after startup.
> When this happens, I can find the exact Camel endpoint/route/processor in
> JConsole, and nothing seems to be wrong. How can a "direct:" endpoint runs
> out of consumer? Isn't the thread that product this warning THE thread to
> process the message?
> We recently upgraded from 5.2 to 5.3.2 and this starts to show up randomly.
> It is a show stopper!
> --Michael
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
