ScyllaDB Cluster Docker Example

An example ScyllaDB cluster setup for the InSpire Posts Service:

• 5 ScyllaDB nodes
• Prometheus for scraping metrics

---

1. Start the Cluster

docker-compose up -d

Node	CQL Port	REST API Port
scylla-node1	19042	19100
scylla-node2	19043	19101
scylla-node3	19044	19102
scylla-node4	19045	19103
scylla-node5	19046	19104

---

2. Verify All Nodes Are Running

docker exec -it scylla-node1 nodetool status                         

Example output:

Datacenter: DC1
===============
Status=Up/Down
|/ State=Normal/Leaving/Joining/Moving
-- Address      Load      Tokens Owns Host ID                              Rack 
UN 192.168.97.2 607.50 KB 256    ?    94e070c4-ba13-48ce-95a9-ba329100e672 Rack1
UN 192.168.97.3 612.60 KB 256    ?    6e58defa-8717-4686-bd82-562fea6d640b Rack1
UN 192.168.97.4 574.76 KB 256    ?    3d07157f-189d-4212-83f0-1cee2bd433ff Rack1
UN 192.168.97.6 602.65 KB 256    ?    a89a542c-c0b0-4af5-b9c4-732de7f64f67 Rack1
UN 192.168.97.7 567.55 KB 256    ?    4b7a37e0-aca3-4531-939f-6943d7066c2e Rack1

---

3. Create a Keyspace and posts Table

docker exec scylla-node1 cqlsh -f migrations/create_posts.cql

---

4. Insert Test Data

docker exec scylla-node1 cqlsh -f migrations/insert_data.cql

---

5. Verify Data Distribution Across Nodes

docker exec -it scylla-node2 nodetool status --keyspace posts_service

You should see all 5 nodes with effective token ownership.

Datacenter: DC1
===============
Status=Up/Down
|/ State=Normal/Leaving/Joining/Moving
-- Address      Load      Tokens Owns (effective) Host ID                              Rack 
UN 192.168.97.2 607.50 KB 256    59.8%            94e070c4-ba13-48ce-95a9-ba329100e672 Rack1
UN 192.168.97.3 581.48 KB 256    61.2%            6e58defa-8717-4686-bd82-562fea6d640b Rack1
UN 192.168.97.4 574.76 KB 256    61.8%            3d07157f-189d-4212-83f0-1cee2bd433ff Rack1
UN 192.168.97.6 602.65 KB 256    56.8%            a89a542c-c0b0-4af5-b9c4-732de7f64f67 Rack1
UN 192.168.97.7 567.55 KB 256    60.5%            4b7a37e0-aca3-4531-939f-6943d7066c2e Rack1

---

6. Run Read Queries

cat migrations/read_data.cql | docker exec -i scylla-node1 cqlsh

Example result:

(10 rows)

 post_id                              | author_id                            | created_at                      | description     | image_height | image_url                    | image_width
--------------------------------------+--------------------------------------+---------------------------------+-----------------+--------------+------------------------------+-------------
 b6c0e97b-6dfa-4001-9319-957ed6f18464 | 95c867bd-214c-4ebe-bc85-38f48a1538c1 | 2025-05-16 22:11:32.541000+0000 | Instagram style |         1350 | https://example.com/img7.jpg |        1080
 d2d68894-a0e7-4618-adf9-aea24db859a2 | 47bb52ba-bac6-4179-a0de-8fe8e4d9aed4 | 2025-05-16 22:11:32.539000+0000 |     Profile pic |         1080 | https://example.com/img2.jpg |        1080
 ...

You can also run queries manually (see read_data.cql):

SELECT * FROM posts WHERE post_id = <uuid>;

SELECT * FROM posts WHERE image_width = 1080 ALLOW FILTERING;

SELECT * FROM posts WHERE image_height = 1080 ALLOW FILTERING;

SELECT * FROM posts WHERE created_at > 1715731200000 ALLOW FILTERING;

---

Destroy the Cluster

docker-compose down --volumes
docker-compose rm -f

---

Prometheus + Grafana

Prometheus is exposed on port 19090 and scrapes metrics from all ScyllaDB nodes. See dashboard.json for an example Grafana dashboard configuration.

---

Contributing

Feel free to open issues or submit pull requests to improve this setup. To contribute:

1. Fork the repository
2. Create a new branch: git checkout -b my-feature
3. Commit your changes
4. Push to your fork and submit a PR

---

TODO

• Grafana dashboard
• Add client connection example