Hi, You will get the message "Task was not deployed or was redeployed" when topology is instability.
You need to look at heap utilization and GC work. I recommend to try G1 garbage collector and jdk 1.8 with option like this: -server -XX:SurvivorRatio=12 -XX:ConcGCThreads=8 -XX:+AlwaysPreTouch -XX:+UseG1GC -XX:MaxGCPauseMillis=2000 -XX:InitiatingHeapOccupancyPercent=30 -XX:G1HeapRegionSize=8m -XX:+ScavengeBeforeFullGC -XX:+DisableExplicitGC and looking after behavior of heap by log -XX:+PrintGCDetails -XX:+PrintGCTimeStamps -XX:+PrintGCDateStamps -XX:+UseGCLogFileRotation -XX:+PrintAdaptiveSizePolicy -XX:+PrintGCApplicationStoppedTime -XX:NumberOfGCLogFiles=10 -XX:GCLogFileSize=100M -Xloggc:/var/gridgain/logs/gc/gc.log Also look at jvm dump, in case OOM through out again: -XX:+HeapDumpOnOutOfMemoryError -XX:HeapDumpPath=/path/to/heapdump -- View this message in context: http://apache-ignite-users.70518.x6.nabble.com/Exceptions-while-running-Closure-with-task-count-1000-tp11989p12193.html Sent from the Apache Ignite Users mailing list archive at Nabble.com.