Database Reference
In-Depth Information
Balanced Clusters
If you are starting with Cassandra 1.2 or greater, you will likely be using virtual
nodes, so you won't need to worry about keeping your cluster balanced.
Everything related to keeping the data distribution in the cluster balanced evenly
will happen in the background. You need to worry about the balance of a cluster
only if you are using a partitioner where token generation isn't done for you.
Firewalls
Cassandra primarily operates using three ports: 7000 (or 7001 if you are using
SSL), 7199, and 9160. In order for Cassandra to be able to properly talk to the
other nodes in the ring, it should to be able to speak TCP to ports 7000 or 7001.
This is the internode cluster communication port. Port 7001 is used only if in-
ternode communication is set to be encrypted. This is done by setting
inter-
node_encryption
to either
all
,
dc
, or
rack
. The default setting of
none
will ensure that internode encryption is off and Cassandra will use port 7000 for
communication.
Port 7199 is the JMX port used for the initial JMX handshake. Once the initial
JMX handshake has been completed, Cassandra picks a high port, or a port greater
than 1024, to continue the communication. The last and likely most important port
is the Thrift client port. This is port 9160. When clients or applications connect to
Cassandra, they do so on port 9160.
The other port that should typically be open in a Cassandra cluster is port 22,
generally known as the SSH port. Though you can run
nodetool
queries from
any Cassandra machine, you will typically want to run additional commands such
as
vmstat
or
dstat
, or even look into the logs to get deeper insight into prob-
lems.
Platforms
Cassandra will run on any system that supports the Java Virtual Machine (JVM).
The easiest way to get started with an actual deployment greater than your per-
sonal machine is to use a cloud service. For instance, on Amazon Web Services,
there are ready-made AMIs (Amazon Machine Images) that come preloaded with
Cassandra. You can start one up and basically go right into using your Cassandra
cluster. This is the reason that so much work has been done to create a snitch that
is specific to Amazon's EC2 (Elastic Computer Cloud).
