Community Tip - New to the community? Learn how to post a question and get help from PTC and industry experts! X
Dear All,
We need to set up a Windchill 10.2 cluster environment with 2 masters and 1 slave.One master would be dedicated to handle the workflow queues,while the other master would handle the rest of the queues. Is such a configuration possible? If yes, how can it be set up?
Thank you.
Regards,
Amit
Solved! Go to Solution.
In a Windchill cluster environment, the master node is the node which runs the master cache server manager. This server manager synchronizes caches between all nodes of the cluster. Only one server manager can act as the cache master at any given time, so you can't have multiple master nodes. But, you can absolutely split your queues between multiple nodes. What you could do is have one master node which runs a BGMS that handles all the non-workflow queues, one slave node that handles user requests, and then third background node that doesn't handle user requests, but just runs a BGMS that handles the workflow queues. This node would be set up the same way as a slave node. But, in the wt.properties file, you'd configure it not to start any Method Servers, and instead just start a BGMS process configured to run the workflow queue group.
In a Windchill cluster environment, the master node is the node which runs the master cache server manager. This server manager synchronizes caches between all nodes of the cluster. Only one server manager can act as the cache master at any given time, so you can't have multiple master nodes. But, you can absolutely split your queues between multiple nodes. What you could do is have one master node which runs a BGMS that handles all the non-workflow queues, one slave node that handles user requests, and then third background node that doesn't handle user requests, but just runs a BGMS that handles the workflow queues. This node would be set up the same way as a slave node. But, in the wt.properties file, you'd configure it not to start any Method Servers, and instead just start a BGMS process configured to run the workflow queue group.
Thanks for the detailed answer, Chris. I believe we may also need to exclude the slave node running background queues (with no foreground Method server running) from the load balancer so that the user requests- http/https are not forwarded to it?