Re: Start a route with aggregation fails due to NPE in AggregateProcessor

2023-02-20 Thread Sydney Henrard
email, a username and a display name. Il lun 20 feb 2023, 19:04 Andrea Cosentino ha scritto: > Hello, > > Please follow the instructions here > https://infra.apache.org/jira-guidelines.html#who to create an account on > your behalf. > > Thanks > > Il lun 20 feb 202

Re: Start a route with aggregation fails due to NPE in AggregateProcessor

2023-02-20 Thread Sydney Henrard
aggregation fails due to NPE in AggregateProcessor Hi Yes that sounds plausible. So it would be good in the camel code, to check for null exchange On Sat, Feb 18, 2023 at 6:46 PM Sydney Henrard wrote: > Hello, > > In my application I have a route with an aggregation using an ag

Start a route with aggregation fails due to NPE in AggregateProcessor

2023-02-18 Thread Sydney Henrard
Hello, In my application I have a route with an aggregation using an aggregation repository ClusteredJdbcAggregationRepository and completionTimeout. This route is part of a micro service that can be scaled. From time to time, the micro service does not start due to a NPE: Caused by: java.lang

Clustered JDBC Aggregation in Kubernetes

2022-09-05 Thread Sydney Henrard
Hello, We are currently migrating a microservice that contains a route that is using aggregation with persistence (LevelDB) from VMs to Kubernetes. We want to be able to deploy the micro service as a Kubernetes Deployment. That's why we switched to ClusteredJdbcAggregationRepository so this way

Re: Aggregator stops route consumption

2022-09-04 Thread Sydney Henrard
Hello, The unit test is using the latest version 3.18.1. I did not really get your point on using LevelDB as the unit test reproduces the issue without persistence. The aggregation is used to remove duplicates from 25 millions messages, knowing that there are 600K unique ids. I wonder if using

Aggregator stops route consumption

2022-09-03 Thread Sydney Henrard
Hello, In my application I have a route that consumes a JMS queue, then aggregate on a header and finally process. The issue is that the consumption of the queue is stopped until the processing is done. The queue contains 25 millions messages (600K unique messages based on aggregation rule), ag