Troubleshooting Startup of Workers from the Worker Agent
When the worker has successfully connected to the Worker Agent using a manual start of the worker, attempt to start it by selecting the entry in the adapter drop-down list. When the entry is selected, the menu stays selected, until the startup timeout is exceeded, or a connection from the worker is achieved.
If this fails, check to see if the command specified to execute it is in fact correct.
Local Worker on Windows – In the case of a local worker on Windows, it is worth trying to run the command using cmd.exe. For example, a Creo Parametric worker can be started as:
cmd.exe /C start "PROEWORKER" /MIN c:\ptc\program_files
\productview_adapters\proe_setup\proeworker.bat
\productview_adapters\proe_setup\proeworker.bat
Remote Worker on Windows – Workers that are running on a remote Windows client must be started using the Worker Daemon.
Remote Worker on UNIX – For remote UNIX workers, you must specify nohup and put the task in the background. Also, Telnet is used to connect to the remote worker. To test, use Telnet from the Worker Agent host, to the worker host. Specify the hostname, username, and password defined (in the agent.ini file) during the configuration of the worker. Assuming a successful connection, it is important that the system prompt (for that user) should not change. Manually executing the specified worker command from this environment should create a connection to the Worker Agent. If not, there is probably a difference in the environment used by Telnet and that of the default user login. Adjust the environment to ensure that the command causes a connection via Telnet, and look for things like DISPLAY, path, and shell type. The Worker Agent should then be able to start the worker.
When the worker fails to start during publishing even after multiple attempts, then an email notification is sent to the specified Windchill users or groups. Worker start is attempted thrice by default and the number of attempts are configured using the following properties:.
• cadagent.maxstartattempts = <integer>, or
• cadagent.maxstartattempts.<workertype> = <integer>. For example,cadagent.maxstartattempts.PROE=5
Configuring Notification If Worker Fails To Start
When the worker fails to start during publishing, all publish jobs sent to that worker keep on failing until the administrator fixes the worker issue and reloads the worker. The administrator user can configure email notification to be triggered if the worker fails to start after multiple attempts. To configure an email notification if the worker fails to start worker:
• Make sure the configuration required for sending an email from Windchill is enabled using WT Properties file like wt.mail.mailhost.
• Set the WVS property cadagent.workerfailedtostart.emailnotification with valid Windchill user or group names. The property value is a list of comma separated Windchill users or group names. The email notification contains the link to the Worker Agent Administration where the administrator user can enable the worker after handling the errors encountered during the publishing job.
cadagent.workerfailedtostart.emailnotification = Administrators [OOTB property value]
where OOTB property value is a list of comma separated Windchill users or group names.
 |
The email notification will not be sent in the following events:
• If the email properties are not set in the WT properties files like wt.mail.mailhost
• If the cadagent.workerfailedtostart.emailnotification property value is not set
• If the cadagent.workerfailedtostart.emailnotification property is set with invalid user or group name(s).
• If the administrator tries to start the worker manually and it fails to start after pre-configured retry attempts.
|