26 May 2020
Apache Cassandra is a free and open-source, distributed, NoSQL database management system designed to handle large amounts of data across many servers, providing high availability with no single point of failure. Cassandra offers robust support for clusters spanning multiple datacenters, with asynchronous masterless replication allowing low latency operations for all clients.
Install Docker and docker-compose for testing purposes
$ apt-get update
$ apt install docker.io
$ systemctl enable docker
$ apt install docker-compose
Create docker-compose file for cassandra cluster
version: '3'
services:
n1:
image: 'cassandra:4.0'
networks:
- cluster
n2:
image: 'cassandra:4.0'
networks:
- cluster
environment:
- CASSANDRA_SEEDS=n1
depends_on:
- n1
n3:
image: 'cassandra:4.0'
networks:
- cluster
environment:
- CASSANDRA_SEEDS=n1
depends_on:
- n1
networks:
cluster: null
Run Node 1
$ docker-compose up -d n1
$ docker ps
$ docker-compose exec n1 nodetool help
$ docker-compose exec n1 nodetool status
$ docker-compose exec n1 nodetool ring
$ docker-compose exec n1 more /etc/cassandra/cassandra.yaml
Run Node 2
$ docker-compose up -d n2
$ docker ps
$ docker-compose exec n2 nodetool status
$ docker-compose exec n2 nodetool ring
$ docker-compose exec n2 more /etc/cassandra/cassandra.yaml
Run Node 3
$ docker-compose up -d n3
$ docker ps
$ docker-compose exec n3 nodetool status
$ docker-compose exec n3 nodetool ring
$ docker-compose exec n3 more /etc/cassandra/cassandra.yaml
Create docker-compose file for cassandra cluster
version: '3'
services:
n1:
image: 'cassandra:4.0'
networks:
- cluster
environment:
- CASSANDRA_SEEDS=n1
- CASSANDRA_ENDPOINT_SNITCH=GossipingPropertyFileSnitch
- CASSANDRA_DC=DC1
- CASSANDRA_RACK=RAC1
n2:
image: 'cassandra:4.0'
networks:
- cluster
environment:
- CASSANDRA_SEEDS=n1
- CASSANDRA_ENDPOINT_SNITCH=GossipingPropertyFileSnitch
- CASSANDRA_DC=DC1
- CASSANDRA_RACK=RAC2
depends_on:
- n1
n3:
image: 'cassandra:4.0'
networks:
- cluster
environment:
- CASSANDRA_SEEDS=n1
- CASSANDRA_ENDPOINT_SNITCH=GossipingPropertyFileSnitch
- CASSANDRA_DC=DC2
- CASSANDRA_RACK=RAC1
depends_on:
- n1
networks:
cluster: null
Please not that, Snitches allows Cassandra to distribute replicas by grouping machines into data centers and racks. For more info, please check this guide
Run Node 1
$ docker-compose up -d n1
$ docker ps
$ docker-compose exec n1 nodetool help
$ docker-compose exec n1 nodetool status
$ docker-compose exec n1 nodetool ring
$ docker-compose exec n1 more /etc/cassandra/cassandra.yaml
$ docker-compose exec n1 more /etc/cassandra/cassandra-rackdc.properties
Run Node 2
$ docker-compose up -d n2
$ docker ps
$ docker-compose exec n2 nodetool status
$ docker-compose exec n2 nodetool ring
$ docker-compose exec n2 more /etc/cassandra/cassandra.yaml
$ docker-compose exec n2 more /etc/cassandra/cassandra-rackdc.properties
Run Node 3
$ docker-compose up -d n3
$ docker ps
$ docker-compose exec n3 nodetool status
$ docker-compose exec n3 nodetool ring
$ docker-compose exec n3 more /etc/cassandra/cassandra.yaml
$ docker-compose exec n3 more /etc/cassandra/cassandra-rackdc.properties
SimpleStrategy: this is the best used in development environment or single data center clusters
CREATE KEYSPACE IF NOT EXISTS clivern WITH REPLICATION = { 'class' : 'SimpleStrategy', 'replication_factor' : 3 };
Here we are asking cassandra to store three copies for all the partitions in all the tables in clivern keyspace.
Let’s create a docker-compose.yml file for multi-node cluster
version: '3'
services:
n1:
build: .
image: cassandra-with-cqlshrc
networks:
- cluster
n2:
image: 'cassandra:4.0'
networks:
- cluster
environment:
- CASSANDRA_SEEDS=n1
depends_on:
- n1
n3:
image: 'cassandra:4.0'
networks:
- cluster
environment:
- CASSANDRA_SEEDS=n1
depends_on:
- n1
networks:
cluster: null
and create Dockerfile
FROM cassandra:4.0
COPY cqlshrc /root/.cqlshrc
and create cqlshrc file
[connection]
;; a timeout in seconds for opening a new connection
timeout = 60
;; a timeout in seconds to execute queries
request_timeout = 60
Bring up the cluster node and wait
$ docker-compose up -d
Get cluster nodes status
$ docker-compose exec n1 nodetool status
Use cqlsh command line tool
$ docker-compose exec n1 cqlsh
# Create the keyspace
>> CREATE KEYSPACE IF NOT EXISTS clivern WITH REPLICATION = { 'class' : 'SimpleStrategy', 'replication_factor' : 3 };
$ docker-compose exec n1 nodetool describering clivern
$ docker-compose exec n3 nodetool status clivern
$ docker-compose exec n1 cqlsh
>> drop keyspace clivern;
>> CREATE KEYSPACE IF NOT EXISTS clivern WITH REPLICATION = { 'class' : 'SimpleStrategy', 'replication_factor' : 1 };
NetworkTopologyStrategy: this is the best used in production with multi datacenter clusters
CREATE KEYSPACE IF NOT EXISTS clivern WITH REPLICATION = { 'class' : 'NetworkTopologyStrategy', 'DC1' : 3, 'DC2' : 1 };
Here we store four copies of the data, 3 copies in DC1 and 1 copy in DC2. You can do something similar to what i did in SimpleStrategy
Consistency level determines how many nodes in the replica must respond for the coordinator node to successfully process a non-lightweight transaction. The consistency level defaults to ONE for all write and read operations.
For a list of all supported Write Consistency Levels, please check this table.
For a list of all supported Read Consistency Levels, please check this table.
Please note that:
Hinted Handoff: On occasion, a node becomes unresponsive while data is being written. a hinted handoff inherently allows Cassandra to continue performing the same number of writes even when the cluster is operating at reduced capacity. for more info, please check this guide.Read Repair: Read repair improves consistency in a Cassandra cluster with every read request. for more info, please check this guide.To create a table
cqlsh:clivern> create table if not exists user (id uuid primary key, username varchar);
cqlsh:clivern> alter table user add firstname varchar;
cqlsh:clivern> desc table user;
CREATE TABLE clivern.user (
id uuid PRIMARY KEY,
firstname text,
username text
) WITH additional_write_policy = '99p'
AND bloom_filter_fp_chance = 0.01
AND caching = {'keys': 'ALL', 'rows_per_partition': 'NONE'}
AND cdc = false
AND comment = ''
AND compaction = {'class': 'org.apache.cassandra.db.compaction.SizeTieredCompactionStrategy', 'max_threshold': '32', 'min_threshold': '4'}
AND compression = {'chunk_length_in_kb': '16', 'class': 'org.apache.cassandra.io.compress.LZ4Compressor'}
AND crc_check_chance = 1.0
AND default_time_to_live = 0
AND extensions = {}
AND gc_grace_seconds = 864000
AND max_index_interval = 2048
AND memtable_flush_period_in_ms = 0
AND min_index_interval = 128
AND read_repair = 'BLOCKING'
AND speculative_retry = '99p';
To insert rows
cqlsh:clivern> insert into user (id, username) values (92d99616-a899-487a-852c-40bee1d9b6eb, 'joe');
cqlsh:clivern> insert into user (id, username) values (92d9962e-9710-4c63-81c7-fc7fb31b3162, 'doe');
To select rows
cqlsh:clivern> select * from user where id = 92d99616-a899-487a-852c-40bee1d9b6eb limit 1;
cqlsh:clivern> select * from user where username = 'joe' limit 1;
InvalidRequest: Error from server: code=2200 [Invalid query] message="Cannot execute this query as it might involve data filtering and thus may have unpredictable performance. If you want to execute this query despite the performance unpredictability, use ALLOW FILTERING"
cqlsh:clivern> select * from user where username = 'joe' ALLOW FILTERING;
To update rows
cqlsh:clivern> update user set firstname = 'Joe' where id = 92d99616-a899-487a-852c-40bee1d9b6eb;
cqlsh:clivern> select * from user where id = 92d99616-a899-487a-852c-40bee1d9b6eb limit 1;
id | firstname | username
--------------------------------------+-----------+----------
92d99616-a899-487a-852c-40bee1d9b6eb | Joe | joe
To delete rows
cqlsh:clivern> delete from user where id = 92d99616-a899-487a-852c-40bee1d9b6eb;
cqlsh:clivern> select * from user where id = 92d99616-a899-487a-852c-40bee1d9b6eb limit 1;
id | firstname | username
----+-----------+----------
(0 rows)