In our implementation, the master node submits tasks to the grid. How can we achieve high availability by having multiple master nodes so that if the master node dies or leaves the grid, another node can collect the results - without having to resubmit the jobs. Also, is it possible to implement load balancing and have multiple master nodes within the same grid
There is the interface ComputeJobMasterLeaveAware but I'm not sure it allows for failover https://ignite.apache.org/releases/mobile/org/apache/ignite/compute/ComputeJobMasterLeaveAware.html
