Hi, Apparently, Pulsar jobs went wild. There are 14 processes which consume 9000+ threads:
22058 jenkins 20 0 19.514g 2.156g 33960 S 36.8 2.3 2032:55 /usr/local/asfpackages/java/jdk1.8.0_191/jre/bin/java -Xmx1G -XX:+UseG1GC -Dpulsar.allocator.pooled=false -Dpulsar.allocator.leak_detection=Advanced -Dpulsar.allocator.exit_on_oom=false -Dlog4j.configurationFile=log4j2.xml -jar /home/jenkins/jenkins-slave/workspace/pulsar_precommit_java8/pulsar-broker/target/surefire/surefirebooter5673414172185975509.jar /home/jenkins/jenkins-slave/workspace/pulsar_precommit_java8/pulsar-broker/target/sur22058 jenkins 20 0 19.514g 2.156g 33960 S 36.8 2.3 2032:55 /usr/local/asfpackages/java/jdk1.8.0_191/jre/bin/java -Xmx1G -XX:+UseG1GC -Dpulsar.allocator.pooled=false -Dpulsar.allocator.leak_detection=Advanced -Dpulsar.allocator.exit_on_oom=false -Dlog4j.configurationFile=log4j2.xml -jar /home/jenkins/jenkins-slave/workspace/pulsar_precommit_java8/pulsar-broker/target/surefire/surefirebooter5673414172185975509.jar /home/jenkins/jenkins-slave/workspace/pulsar_precommit_java8/pulsar-broker/target/sur What do you think if we kill those processes, and kindly ask Pulsar team to refrain from executing that kind of workload? Vladimir
