OSA - MARKETPLACE Cluster Have Zookeeper And Kafka Instability After Upgrade (Doc ID 3081292.1)

Last updated on APRIL 11, 2025

Applies to:

Goal

We are performing an upgrade of our OSA installations on Marketplace.
Between other problems, we faced with a huge instability of zookeeper and kafka services, that continuously restarted after the upgrade.
The cluster is a 7 node (1 OSA + 2 Spark master + 4 Spark worker & kafka

The zookeeper service (on all 4 worker nodes) continuously restart giving the following errors in the server.log file:

[2025-04-04 13:00:09,989] WARN Close of session 0x0 (org.apache.zookeeper.server.NIOServerCnxn)
java.io.IOException: ZooKeeperServer not running
and
(org.apache.zookeeper.server.NIOServerCnxnFactory) java.lang.OutOfMemoryError: Java heap space site:community.xxxxxx.com


and the following error in the zookeeper-gc.log.0.current file:

[GC concurrent-mark-abort]

consequently also the kafka services restarted on the 4 worker nodes.

 

Solution

My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.