Skip to main content
B
bmills1-Visitor
1-Visitor
June 29, 2010
Question

Unable to create new Worker Agents

  • June 29, 2010
  • 4 replies
  • 1176 views

Intralink 9.1 M050

Pro/E WF4 M110

I went live with the above configuration this week after a successful migration from 3.4 and am in the process of setting up a worker to perform Productview conversions. I goto the Worker Agent Administrator and see the default OOTB entry. I then click on the Configure link to create a new agent (Add Worker). Afterrunning through all of the steps in the Wizard and selecting Apply, the new agent appears in the list in the Wizard but when I click on "Return to Worker Agent Administrator" the newly created agent does not appear in the list, only the default OOTB entry. Is there a step I am missing?

Bob Mills

DRS Technologies

4 replies

avillanueva
23-Emerald I
avillanueva23-Emerald I
23-Emerald I
June 29, 2010
Did you reload the agent.ini file?

D
ddemay1-Visitor
1-Visitor
March 25, 2013

Hello all,



I can customize the heck out of this system. Visualization is not my forte. We have a PTC call open to add a new worker to the agent.ini file. I came across this post and do not see a really obvious solution or answer. Antonio, when you say reload the agent.ini file, could you elaborate? Thinking since PTC told us to edit this file, that perhaps the changes are still not being picked up properly. One thing I would note is we are experiencing this on our production cluster, when the cad worker machine is added with the exact same steps to a monolithic test server, there are no issues at all. We are on PDMLink 9.1 M050, which in theory, is a very similar codebase.


Anyone out there got any ideas? I haven't really dived into this myself, but one of our team members has been banging head on keyboard to figure this out.


Curious to know what type of communication takes place (if any) that anyone is aware of during the configuration of a new worker (between the background method server and the cad worker system)



Other useful info:


When you go to edit workers, the new worker does not appear. It also doe snot appear on the worker agent administration page.


We deleted the original cad worker from the worker agent administration page, no luck. Adding it back, no luck.


The original worker that was there experienced hardware failure, so we are staging a new system for the cad worker. It was a fresh install, but when things did not work, certain configuration files would've been copied over. Pretty sure this isn't the problem, but wanted to share.


The new worker was added as a new system and hostname. However, we were able to salvage the old hardware temporarily while the new worker is brought online.


Verified all username and passwords already, as well as file permissions.



In the meantime, we are going to compare an agent.ini file from our test server where the cad worker was successfully added to see what is different from the production server agent.ini file.



Any insights would be greatly appreciated!!



Thanks,


David DeMay





In Reply to Antonio Villanueva:


Did you reload the agent.ini file?

D
ddemay1-Visitor
1-Visitor
March 25, 2013
Update: Issue appears somewhat resolved. Agent.ini file corrupt. We removed all workers. Made sure the agent.ini file was identical on both parts of our production cluster. We then added back only the new host/worker and it showed up on one screen, but not the other, even with reload. Only one of the agent.ini files was updated. We then copied the agent.ini file to our master cluster node which has the background methodserver. The worker administrator UI behaves correctly, but now have to resolve the fails to start issue. (Which is further then we got before.


Sent from my Verizon Wireless 4G LTE Smartphone
T
tmcnaney14-Alexandrite
14-Alexandrite
March 26, 2013

If you are configuring in a Windchill clustered environment, you initially cannot use the Worker Agent Administrator from any node.


Visualization publishing and thumbnailgeneration is processed using background queues. In a cluster, generally only one member processes the queue using a dedicatedBackground MethodServer. Manual configuration is required.


The "host" parameter enables the Worker Agent Administration Pages, running on Foreground MethodServers, to connect to the Background MethodServer (BGMS), which might be running on a different host machine, by identifying the hostname of the machine running the BGMS.


Additional information is available in the Windchill Visualization Services Guide.


https://www.ptc.com/appserver/cs/view/solution.jsp?n=149063


https://www.ptc.com/appserver/cs/view/solution.jsp?n=CS17218


Make one of the following host entries in the [agent] section of the agent.ini file of the slave servers:



host=<background method=" server=" machine's=" host=" name=">


or


host=<background method=" server=" machine's=" ip=" address=">



  • Define the port entry in the [agent] section:

port=<port number=" that=" the=" cad=" (worker)=" agent=" is=" listening=" to=" on=" the=" background=" method=" server=">



  • The agent.ini file of the slave should look as follows:

[agent]
host=<bgms_hostname_or_ip_address>
port=<bgms_cadagent_port>
transferpath=$(wt.temp)\transfer
numworkers=0
;

No need to define any worker entries in the agent.ini file of the slaves


  • Additional Information

    • The agent.ini file of the master should be edited manually to define the different workers

    • The CAD (Worker) Agent Administration Page viewed from the slave servers should show the workers configured on the master (BG MS) host

    • You can start and stop workers from this window

    • The workers configuration has to be done manually

    • Setting the debug=true parameter in the [agent] enables sending the the worker messages to the Background Method Server and logs

    • BGMS_CADAgent_Port value is 5600


Hope that helps!

Terms & ConditionsCookie settingsAccessibility statement