I have reproduce you issue. Seem that is java heap issue.
If the test is producing in one physical machine the issue does not reproduce, because costs of network communications are small. But when you are dedicate nodes by others computers, overheads are growing and heap space becomes insufficient. In result long GC pause or OOM exception and breakdown topology. If you wan to avoid this issue you should add enough memory for jvm and additional tuning you application. Look in his article[1]. [1]: https://apacheignite.readme.io/docs/jvm-and-system-tuning#section-jvm-tuning-for-clusters-with-on_heap-caches -- View this message in context: http://apache-ignite-users.70518.x6.nabble.com/Exceptions-while-running-Closure-with-task-count-1000-tp11989p12174.html Sent from the Apache Ignite Users mailing list archive at Nabble.com.
