Unclear on parallel execution about dnr-editor HOT 1 CLOSED

Hi there,

Thanks for your interest. Your observation is right. It is an interesting and known limitation on the constraint definition where you cannot have an excluding constraint. I suppose adding something like "run this node on device other than X" will be something in my todo list.

The main reason behind this is when a node doesn't have a constraint, it will be deployed and available on every participating NR instances. Consequently, when the inject node realizes that there is an exec node (it's peer) that can run in the same device as its own, it won't send a message out to other devices.

At the current situation, there is a hack to make this happen I guess. This is by using the location requirement, essentially we want all the nodes to have a constraint. So you pretending to have 3 NR instances in 3 different locations, let say the inject node is constrained to run in Vancouver, the exec to run in Seattle and the debug to run in SanFran (could be the same as Vancouver). You can do this with location-based node requirement.

Then you configure your NR instances so that one located in Vancouver, the ones that needs to run the exec node to be located in Seattle, and the one who run the debug node to be in SanFran. You can do this by configuring the Dnr Daemon node and checking the Location Fixed checkbox.

You could try this and let me know.

from dnr-editor.