Problem: common Cassandra pods don't start failing with the following message 

log -f selffish-cassandra-cassandra-0

• • • buffer overflow detected ***: java terminated

======= Backtrace: =========

/lib/x86_64-linux-gnu/libc.so.6(+0x731af)[0x7fd315f481af]

/lib/x86_64-linux-gnu/libc.so.6(__fortify_fail+0x37)[0x7fd315fcdaa7]

/lib/x86_64-linux-gnu/libc.so.6(+0xf6cc0)[0x7fd315fcbcc0]

/usr/lib/jvm/java-7-openjdk-amd64/jre/lib/amd64/libnet.so(Java_java_net_Inet4AddressImpl_getLocalHostName+0x1a4)[0x7fd2f7087624]

[0x7fd30c3c9e98]

======= Memory map: ========

This seems to be due to a JDK7 bug which has issue resolving long hostnames.

WORK-AROUND: I have updated the image of the shared Cassandra to 

"library/cassandra:2.2.14" and the Cassandra image boots up fine!

IDEAL SOLUTION: SDC should upgrade titangraph library to JanusGraph library which should support Cassandra 3.x. and then change the image back to any recent 3.x cassandra image.