Hazelcast cluster member crash results in loosing all scheduled tasks

Question

We are running 4 instances of our java application in hazelcast cluster. We scheduled around 2000 task using schedule executor service schedule method. Hazelcast partition all these 2000 tasks across the 4 instances. Due to some reason one of the cluster member crashes then all the task that are assign to the partition that are owned by the crashed node are lost, rest all 3 cluster member completed their assign task.

So how can we overcome this problem to avoid the lost tasks.

Answer 1

Try using the Durable Executor

Probably a good idea also to find why the process crashed in the first place.