docker-flink not showing all log statements - docker

Question

I am using 2 docker flink images with AMIDST and the following sample code. AMIDST is a probabilistic graphical model framework which supports Flink.
One image is running as JobManager the other as TaskManager. JM is reachable via DNS and I provide my own log4j.properties based on the startup script in bin/flink-console.sh used by these images.
public class ParallelMLExample {
private static final Logger LOG = LoggerFactory.getLogger(ParallelMLExample.class);
public static void main(String[] args) throws Exception {
final ExecutionEnvironment env;
//Set-up Flink session
env = ExecutionEnvironment.getExecutionEnvironment();
env.getConfig().disableSysoutLogging();
//generate a random dataset
DataFlink<DataInstance> dataFlink = new DataSetGenerator().generate(env, 1234, 1000, 5, 0);
//Creates a DAG with the NaiveBayes structure for the random dataset
DAG dag = DAGGenerator.getNaiveBayesStructure(dataFlink.getAttributes(), "DiscreteVar4");
LOG.info(dag.toString());
//Create the Learner object
ParameterLearningAlgorithm learningAlgorithmFlink = new ParallelMaximumLikelihood();
//Learning parameters
learningAlgorithmFlink.setBatchSize(10);
learningAlgorithmFlink.setDAG(dag);
//Initialize the learning process
learningAlgorithmFlink.initLearning();
//Learn from the flink data
LOG.info("########## BEFORE UPDATEMODEL ##########");
learningAlgorithmFlink.updateModel(dataFlink);
LOG.info("########## AFTER UPDATEMODEL ##########");
//Print the learnt Bayes Net
BayesianNetwork bn = learningAlgorithmFlink.getLearntBayesianNetwork();
LOG.info(bn.toString());
}
}
The problem is that I only see LOG.info() entries up until the updateModel call. After that silence. If I comment out this call, I can see the other entries. I am silencing Flink entries on purpose here.
Creating flink_jobmanager_1 ... done
Creating flink_jobmanager_1 ...
Creating flink_taskmanager_1 ... done
Attaching to flink_jobmanager_1, flink_taskmanager_1
jobmanager_1 | Starting Job Manager
jobmanager_1 | config file:
taskmanager_1 | Starting Task Manager
jobmanager_1 | jobmanager.rpc.address: jobmanager
taskmanager_1 | config file:
jobmanager_1 | jobmanager.rpc.port: 6123
jobmanager_1 | jobmanager.heap.mb: 1024
taskmanager_1 | jobmanager.rpc.address: jobmanager
jobmanager_1 | taskmanager.heap.mb: 1024
taskmanager_1 | jobmanager.rpc.port: 6123
jobmanager_1 | taskmanager.numberOfTaskSlots: 1
taskmanager_1 | jobmanager.heap.mb: 1024
jobmanager_1 | taskmanager.memory.preallocate: false
taskmanager_1 | taskmanager.heap.mb: 1024
jobmanager_1 | parallelism.default: 1
taskmanager_1 | taskmanager.numberOfTaskSlots: 2
jobmanager_1 | web.port: 8081
taskmanager_1 | taskmanager.memory.preallocate: false
jobmanager_1 | blob.server.port: 6124
taskmanager_1 | parallelism.default: 1
jobmanager_1 | query.server.port: 6125
taskmanager_1 | web.port: 8081
jobmanager_1 | Starting jobmanager as a console application on host c16d9156ff68.
taskmanager_1 | blob.server.port: 6124
taskmanager_1 | query.server.port: 6125
taskmanager_1 | Starting taskmanager as a console application on host 76c78378d35c.
jobmanager_1 | 2018-02-18 15:31:42,809 INFO akka.event.slf4j.Slf4jLogger - Slf4jLogger started
taskmanager_1 | 2018-02-18 15:31:43,897 INFO akka.event.slf4j.Slf4jLogger - Slf4jLogger started
jobmanager_1 | 2018-02-18 15:32:18,667 INFO com.ness.ParallelMLExample - DAG
jobmanager_1 | DiscreteVar0 has 1 parent(s): {DiscreteVar4}
jobmanager_1 | DiscreteVar1 has 1 parent(s): {DiscreteVar4}
jobmanager_1 | DiscreteVar2 has 1 parent(s): {DiscreteVar4}
jobmanager_1 | DiscreteVar3 has 1 parent(s): {DiscreteVar4}
jobmanager_1 | DiscreteVar4 has 0 parent(s): {}
jobmanager_1 |
jobmanager_1 | 2018-02-18 15:32:18,679 INFO com.ness.ParallelMLExample - ########## BEFORE UPDATEMODEL ##########
The updateModel method starts with a new Configuration() then retrieves the data set. It then runs a map, reduce and collect against the supplied data set but does not seem to be messing with root loggers...
What am I missing?