Deploying cassandra-mesos framework with marathon

Question

I'm using the latest jar files of cassandra-mesos framework (by using this jason file: https://teamcity.mesosphere.io/repository/download/Oss_Mesos_Cassandra_CassandraFramework/97399:id/marathon.json), but getting the following errors:

I0310 13:19:34.699774 16389 sched.cpp:264] No credentials provided. Attempting to register without authentication I0310 13:19:34.701026 16389 sched.cpp:819] Got error 'Completed framework attempted to re-register' I0310 13:19:34.701038 16389 sched.cpp:1625] Asked to abort the driver I0310 13:19:34.701364 16389 sched.cpp:861] Aborting framework '20160309-183453-2497969674-5050-19271-0001' I0310 13:19:34.719744 16373 sched.cpp:1591] Asked to stop the driver I0310 13:19:34.719784 16389 sched.cpp:835] Stopping framework '20160309-183453-2497969674-5050-19271-0001'

Any idea?

Answer 1

Thanks a lot :-). It's working now. but when I tried to check zookeeper for cassandra-mesos, I got the following error: mesos-resolve zk://mesos-master-2:2181/cassandra-mesos/cassandra-mesos-fw 2016-03-13 12:46:22,428:26613(0x7fa4fa843700):ZOO_INFO@log_env@712: Client environment:zookeeper.version=zookeeper C client 3.4.5 2016-03-13 12:46:22,428:26613(0x7fa4fa843700):ZOO_INFO@log_env@716: Client environment:host.name=mesos-slave-1 2016-03-13 12:46:22,428:26613(0x7fa4fa843700):ZOO_INFO@log_env@723: Client environment:os.name=Linux 2016-03-13 12:46:22,428:26613(0x7fa4fa843700):ZOO_INFO@log_env@724: Client environment:os.arch=3.10.0-327.4.4.el7.x86_64 2016-03-13 12:46:22,428:26613(0x7fa4fa843700):ZOO_INFO@log_env@725: Client environment:os.version=#1 SMP Tue Jan 5 16:07:00 UTC 2016 2016-03-13 12:46:22,428:26613(0x7fa4fa843700):ZOO_INFO@log_env@733: Client environment:user.name=root 2016-03-13 12:46:22,428:26613(0x7fa4fa843700):ZOO_INFO@log_env@741: Client environment:user.home=/root 2016-03-13 12:46:22,428:26613(0x7fa4fa843700):ZOO_INFO@log_env@753: Client environment:user.dir=/ephemeral/cassandra-mesos 2016-03-13 12:46:22,428:26613(0x7fa4fa843700):ZOO_INFO@zookeeper_init@786: Initiating client connection, host=mesos-master-2:2181 sessionTimeout=10000 watcher=0x7fa5023200b0 sessionId=0 sessionPasswd= context=0x7fa4d8001ec0 flags=0 2016-03-13 12:46:22,429:26613(0x7fa4f6628700):ZOO_INFO@check_events@1703: initiated connection to server [10.254.227.148:2181] 2016-03-13 12:46:22,434:26613(0x7fa4f6628700):ZOO_INFO@check_events@1750: session establishment complete on server [10.254.227.148:2181], sessionId=0x25364fb9f3a0020, negotiated timeout=10000 WARNING: Logging before InitGoogleLogging() is written to STDERR I0313 12:46:22.434587 26616 group.cpp:313] Group process (group(1)@10.254.235.46:56890) connected to ZooKeeper I0313 12:46:22.434659 26616 group.cpp:787] Syncing group operations: queue size (joins, cancels, datas) = (0, 0, 0) I0313 12:46:22.434670 26616 group.cpp:385] Trying to create path '/cassandra-mesos/cassandra-mesos-fw' in ZooKeeper Failed to detect master from 'zk://mesos-master-2:2181/cassandra-mesos/cassandra-mesos-fw' within 5secs

Answer 2

The error says Completed framework attempted to re-register which means the framework keeps its state somewhere (probably in Zookeeper, but cannot access your URL with marathon.json to verify), and thus tries to start with the framework ID stored in this state. However, that framework ID is already deregistered, and Mesos does not allow you to start the framework with the same ID again.

The solution to this would be either to pick a different znode for framework storage or remove the existing znode before starting the framework.