Cassandra database instance exits after 10 hours of use

Products

• Mainframe Operational Intelligence

Issue/Introduction

The exiting of any Cassandra microservice is not normal. 

There is the Sysview MTC zLinux appliance running and after being up for about 10 hours one of the cassandra database instances exits.

Is this normal? 

Resolution

Root cause of the exited container was insufficient memory.

The kernel killed the cassandra3 process due to lack of memory. This was identified in the syslogs/var/log/messages file.

Sample message for searching was "SVN0_MFOI kernel: Out of memory: Kill process 5320 (java-run) score 212 or sacrifice child" 

Increased memory from 39GB to recommended 64GB.