I'm facing the grid configuration and doing a workaround about but what I've learned is stopping me about this idea.
Configuration needed:
- You need one selenium server running with the hub role and listening to the nodes.
- N appium server instances as nodes for N devices you want to test which listen appium client and send the data to the hub.
- N appium clients which are connected with each appium server.
So if you need to automate the test launching you need:
- Take the list of all devices you have and get their names.
- Launch one selenium server.
- Launch N appium servers each one with a different JSON file and ports assigned.
- Launch N appium clients each one with the parameters the client needs and the server needs.
(Take into account that each appium client is controlled by the same code, so it seems unnecessary)
So my question is:
Why all this headache is needed if you can simply launch an appium server and an appium client through the same port?
What is really doing the selenium server?
(I'm controlling all the setup from a bash file and I started to feel seek when I saw all this huge configuration)
(huge because is needed to launch it dynamically independent).
I hope you could understand everything I exposed, sorry if my english is not as well as I needed.
Thank you for your time