Release v0.2.0 · mesosphere-backup/cassandra-mesos-deprecated

New Features

Multi datacenter support. Full docs can be found at http://mesosphere.github.io/cassandra-mesos/docs/multi-dc.html
Improved health check logic
Fixed executor lost handling
Fixed exception when repairing cluster
DCOS Compatibility improvements

Users who were running Cassandra on Mesos v0.1.0 should be able to update to v0.2.0 with minimal impact. The version of Cassandra stayed at 2.1.4.

Shutdown scheduler running v0.1.0
Start scheduler running v0.2.0
Wait for new scheduler to come up and register with mesos
Manually kill each executor in the cluster one at a time
- If Cassandra on Mesos is configured to write it's data to the sandbox (the default behavior) this will result in potential data loss
Cassandra on Mesos should handle the event and start new executors on each node automatically

#114 Allow configuration of mesos user and advertised host for http server
#115 Allow configuration of default rack and dc for GossipingPropertyFileSnitch
#116 Add overview to Table of Contents
#112 UninitializedMessageException when repairing cluster
#119 Update code to ensure that currently running job is used as base of building new job state.
#118 Update health check logic to be tolerant of nodes that have been terminated
#120 Add functionality to sync seeds between clusters
#123 Clarify purpose of env variables for configuration
#125 Update LD_LIBRARY_PATH of executors running on DCOS
#124 Accept offers for default role * in addition to the specified mesosRole
#121 Cassandra tasks getting LOST status in mesos
#129 Update executor lost handling to properly clear executor information
#130 Add docs for multi-dc support

Thanks goes to the following individuals (in alphabetic order)