0

This question was previously asked here,

VIP publish function not timing out when remote platform is dead,

but unfortunately went unanswered.

This gist of the problem is that the vip.pubsub.publish method does not timeout if a remote platform drops off the network.

Specifically, I have an agent that publishes both locally and to several different external platforms. If individual agents drop off the remote platform there is no problem. However, if the entire remote platform drops off the network, the agent will hang when it attempts to publish to that platform. This keeps the agent from publishing locally or to the other remote platforms.

The method I am using to publish is, 

master.vip.pubsub.publish('pubsub', topic, headers, message).get(timeout=10)

Is there anything that can be done within the volttron platform to enforce a timeout?

I am using the develop branch

Chris
  • 5
  • 1
  • I added a response to the other question. please let me know if that helps your case as well. I suspect not however so I will look into this. – Craig Jan 31 '17 at 21:41
  • My set up is that the agent connects to the remote platform on start. I am assuming this is a successful connection, so the timeout in your response does not come in to play. If the remote platform then drops off the network, the agent gets stuck trying to publish to the downed platform. Is it possible to arrange another timeout using gevent for the volttron publish method? Or would it be better to connect and then disconnect from the remote platform when I want to send a message to ensure I'm publishing to a live platform? – Chris Feb 01 '17 at 16:22

0 Answers0