Agent communication error referring to an "Invalid prefix". "Prefix value must be in the range of 0 to 32,768."
search cancel

Agent communication error referring to an "Invalid prefix". "Prefix value must be in the range of 0 to 32,768."

book

Article ID: 139999

calendar_today

Updated On:

Products

CA Workload Automation DE - Business Agents (dSeries) CA Workload Automation DE DSERIES- SERVER CA Workload Automation DE - System Agent (dSeries) CA Workload Automation DE - Scheduler (dSeries)

Issue/Introduction

The Manager is unable to send jobs to an Agent. Communication errors similar to the following are logged in the errors.txt file and tracelog.txt file (IP addresses have been changed for security purposes):

 

20191101 17:45:12.426 [essential] [ERROR] DM.InputProcessor_8: [2019-11-01_17:45:12.426] Exception: The conversation 192.168.xxx.xxx:7507->192.168.xxx.xxx:58739 contains invalid prefix 38,489. Prefix

value must be in the range of 0 to 32,768.

com.ca.wa.comp.library.communications.WAConversationException: The conversation 192.168.xxx.xxx:7507->192.168.xxx.xxx:58739 contains invalid prefix 38,489. Prefix value must be in the range of 0 to 32

,768.

        at com.ca.wa.comp.library.communications.WAConversation.validateDataLength(WAConversation.java:613)

        at com.ca.wa.comp.library.communications.WAConversation.receivePrefix(WAConversation.java:555)

        at com.ca.wa.comp.library.communications.WAConversation.receiveBinary(WAConversation.java:453)

        at com.ca.wa.comp.library.communications.WAConversation.receiveText(WAConversation.java:647)

        at com.ca.wa.comp.distributedmanager.communications.WADistributedManagerInputSession.run(WADistributedManagerInputSession.java:142)

        at com.ca.wa.core.library.concurrent.WAThreadPool$ThreadPoolThread.run(WAThreadPool.java:698)

Environment

Any DE release

Any Agent release

Cause

There may be some corruption in a particular Agent'a database directory.
The Agent has a directory under the agent root dir called 'database'. Sometimes, some files can become corrupt if writing to the files was interrupted by an issue with the Agent server's resources (CPU, disk space, memory).

Resolution

Stop the Agent:

cd <agent_root_dir>

./cybAgent -s


Rename the database directory:

mv <agent_root_dir>/database <agent_root_dir>/database_bak


Start the Agent:

cd <agent_root_dir>

./cybAgent -a


A new <agent_root_dir>/database directory will be created.


Note: The Agent will lose the status of any jobs that were running immediately prior to the above steps. You may have to update the status of the jobs manually afterwards. This is called a "cold start" of the Agent.