AGENTDOWN: Name or service not known

Products

• CA Workload Automation DE

Issue/Introduction

We are experiencing intermittent AGENTDOWN status messages for Apps on multiple ESP agents. When this happens we see the following exceptions in the scheduling server's logs: 

20220513 17:46:01.789 [essential] [ERROR] DM.OutputMessageQueue.<agent_name>: [2022-05-13_17:46:01.789] Exception: <agent_hostname>: Name or service not known^M
java.net.UnknownHostException: <agent_hostname>: Name or service not known

 

Environment

Release : 12.3

Component : ESP dSeries Workload Automation DE

Cause

This is the result of DNS queries, sent by the dSeries scheduling server, that are responded to by DNS with problems. For example: 

• Server failure; or
• No such name

Resolution

There are 2 options for fixing these problems. 

• Work with your teams to address the DNS problems
  • This wireshark filter may help: dns.flags.rcode != 0
• On the dSeries scheduling server, add an entry in its hosts file with the ip/hostname of the agent machines that DNS intermittently has problems resolving.