Re: ClassNotFoundException : org.apache.flink.api.common.operators.util.UserCodeObjectWrapper, while trying to run locally

2015-12-21 Thread peterpanne
I identified the root-cause of the problem: When the FlinkUserCodeClassLoader ist created it simply instantiates the URLClassLoader which takes the System-Classloader as parent. However in the web-application-setting the current thread-webapp-classloader should be taken. I tried to change the clas

[jira] [Created] (FLINK-3189) Error while parsing job arguments passed by CLI

2015-12-21 Thread Filip Leczycki (JIRA)
Filip Leczycki created FLINK-3189: - Summary: Error while parsing job arguments passed by CLI Key: FLINK-3189 URL: https://issues.apache.org/jira/browse/FLINK-3189 Project: Flink Issue Type: B

Re: Incremental checkpoints for Flink

2015-12-21 Thread Paris Carbone
Hi Marius, This is a pretty good and quite challenging topic for a thesis! Your thoughts are pretty much aligned with strategies that we have discussed in the past. There are some initial steps in the makings towards that direction. For example, Aljoscha proposed specialized operator state (KVSt

Re: No job recovery after job manager failure

2015-12-21 Thread Ufuk Celebi
The issue was that 1) local state backend but loss of VM and 2) recovery did not log any Exception. 2) has been addressed in this PR: https://github.com/apache/flink/pull/1472 – Ufuk > On 17 Dec 2015, at 15:26, Ufuk Celebi wrote: > > As an update: I’m investigating this. Ali sent me the log