Cassandra tasks are not restarted when the Mesos cluster is shutdown

[mesosphere-backup]

Cassandra tasks are not restarted when the Mesos cluster is shutdown

[tnn1t1s]

I had this issue earlier today. I removed the cassandraMesos directory from my zookeeper instance and restarted the framework. After that, it started taking offers and launching (not successfully yet ...) tasks.

[Bleakwise]

I have this same issue, been driving me mad. The only solution I've found is to completely re-install Mesos and zookeeper (maybe just zookeeper?). Where is the cassandraMesos directory?

UPDATE:
I was able to get it going by wiping out zookeeper's data directory on my Mesos-master. On Ubuntu/Mesosphere this was /var/lib/zookeeper/version-2.
a-la
sudo rm -R /var/lib/zookeeper/version2

Once I did that, Cassandra fired up on all 3 of my slave-nodes.

[Bleakwise]

UPDATE2:
Using the zkCli.sh utility I went into ZK and deleted
/cassandraMesos/tesserver/app:nodeSet and app:frameworkId which also got cassandra-mesos back running.

I'm going to sleep, but tomorrow I plan to see if just deleting app:nodeset OR just app:frameworkId gets it going as well. Once either the app:nodeset or app:frameworkid has been isolated as the problem I'll better know how to continue in finding some resolution.

PS: Just incase anyone reading this is new
On my Ubuntu 14.04 system zkCli.sh was in...
/usr/share/zookeeper/bin
I belive it's part of the zookeeper-bin package. If you're on Redhat or some other you can always do
find / -name zkCli.sh, that is, if this bug even exists on Redhat or others.

[BenWhitehead]

Obsolete framework has been rewritten.