Our cluster sizes are getting bigger and bigger [40-50machines] & there are times when we need to do a full cluster restart
It is really painful to log into each machines and restart ES processes running on them
For mitigating that, we were writing a daemon process in each machine that would be responsible for starting/stopping ES nodes on that machine thru a rest API [Java daemon]
But the problem we are facing is, whenever we spawn a ES process [./bin/elasticsearch -d] from the Java daemon, we are unable to get back the PID of the ES process. Can someone explain how ES as a daemon works?
We use RuntimeExec.exec for executing the script