I'm currently getting ThingWorx to run on a kubernetes cluster. I basically have it running, but so far only in a non HA setup. The database -- I'm using postgreSQL -- is using a multiple nodes with master/slave failover etc., but ThingWorx runs in a single POD.
I want to change that -- we need ThingWorx in a HA and Load Balanced setup. For this to work I'm in the process of creating a kubernetes operator for ThingWorx.
Does anyone have some insight as of how the ZooKeeper / ThingWorx interaction works in the standard PTC HA setup? I want to use native kubernetes primitives instead of having to spin up a ZooKeeper cluster.
Not quite sure how you are looking to do the HA setup.
But with the zookeeper setup, there is a specific handle that allows Thingworx to be running but not connected to the Thingworx Storage. I apologize for not having the deep details on this.
So there is one instance fully up and running and others (1 or more) that are designated as 'standby' that have tomcat up and running, Thingworx initialized but are not connected to the Store so they don't have the model deployed etc.
A simpler HA setup would be to have tomcat NOT running and have it started when the leader goes down.
Hi -- sorry for the late answer :(
As for the HA setup -- that's what I feared. Essentially TWX runs in a "hot spare" configuration. I hoped that we could use the spares for read-only queries.
Today I saw this post:
This is essentially what I *actually* need/want to do, will try that.
Could you please confirm if your reply is answer of this post, So that we can mark it accept as solution for the benefit of others.