I am trying to configure a small Hadoop3 test cluster on Oracle VM instances.
The worker nodes (datanodes) are not connecting to the masternode (namenode).
The following lists the steps that are treid already (but still the problem persists):
- Deleted namenode and datanode directories, format namenode and restart the deamons
- Changed public ips to private ips in /etc/hosts file of master and slave nodes (assuming Oracle prevents nodes accessing each other on their public IPs, owing to internal subnetting)
- Changed the ingress/egress security rules for the compartment subnet (to open all ports required for Hadoop)
When the cluster is up, nmap on namenode shows that port 9000 is open and is listening.
Note that ssh and ping from slaves to master node work correctly, but traceroute from slave to master is not working.
None of the worker datanodes are able to establish a connection to the masternode on port 9000.
I am stuck at this issue for a few days now.
Any help appreicated!