You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@camel.apache.org by "Ben O'Day (JIRA)" <ji...@apache.org> on 2015/07/13 03:34:04 UTC

[jira] [Commented] (CAMEL-8947) LevelDBAggregatorRepository redelivers incorrectly

    [ https://issues.apache.org/jira/browse/CAMEL-8947?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14624108#comment-14624108 ] 

Ben O'Day commented on CAMEL-8947:
----------------------------------

I'm getting the same results using the HawtDBAggregationRepository and HazelcastAggregationRepository...seems like there is a bug in confirming/removing Exchanges from the repo when certain aggregation strategies are used...

> LevelDBAggregatorRepository redelivers incorrectly
> --------------------------------------------------
>
>                 Key: CAMEL-8947
>                 URL: https://issues.apache.org/jira/browse/CAMEL-8947
>             Project: Camel
>          Issue Type: Bug
>          Components: camel-leveldb
>    Affects Versions: 2.16
>            Reporter: Ben O'Day
>         Attachments: LevelDBAggregatorUseLatestTest.patch
>
>
> I'm using an the aggregator with the LevelDBAggregationRepository and seeing an incorrect redelivery of a message when using the UseLatestAggregationStrategy...
> this very basic route (see attached)...
> 				from("direct:start")
> 						.aggregate(constant(true), new UseLatestAggregationStrategy())
> 							.completionSize(2)
> 							.aggregationRepository(repo)
> 						.to("mock:mock");
> shows the following behavior...
> WARN  LevelDBAggregationRepository   - Unable to confirm exchangeId [ID-localhost-63819-1436483565832-0-6 from repository repo1: Not Found]
> DEBUG LevelDBAggregationRepository   - Scanned and found 1 exchange(s) to recover (note some of them may already be in progress).
> DEBUG LevelDBAggregationRepository   - Recovering exchangeId [ID-localhost-63819-1436483565832-0-3] -> Exchange[Message: test1]
> resulting in a duplicate message being processed through the aggregator route...
> if the default in-memory repo is used, the test behaves as expected...no unnecessary redelivery, etc.  



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