Pivotal Knowledge Base

Follow

Elastic Runtime deploy fails with a UDP port error

Environment 

Pivotal Elastic Runtime: 1.6 to 1.12

Symptom

When deploying an Elastic Runtime job, the deploy fails and returns the following error: 

Failed to start UDP listener. Err: listen udp 10.0.0.10:1234: bind: address already in use

Cause

dhclient, the Linux process that sends Dynamic Host Configuration Protocol (DHCP) client requests, claims two UDP ports at random when a VM starts. Because the Linux process selects the ports rather than Elastic Runtime, it is possible for dhclient to select a port that is already in use by a job in your deployment.

If dhclient selects a port that one of your Elastic Runtime jobs is configured to use, deploying that job fails and returns the above error.

Resolution

Restart the VM that produces this error. Upon restart, the dhclient process selects two new UDP ports.

Comments

Powered by Zendesk