Adding code to treat a suddenly nil tracker as an unavailable partition error.
We see this error in our client code before this line of code panics and crashes due to tracker being null.
This happens almost 100% of the time a server node is restarted for any reason (maintenance, updating, etc.) and we are doing a partition scan.
Adding code to treat a suddenly nil tracker as an unavailable partition error. We see this error in our client code before this line of code panics and crashes due to tracker being null. This happens almost 100% of the time a server node is restarted for any reason (maintenance, updating, etc.) and we are doing a partition scan.