Skip to content

Getting Started

You have a chango cluster up — the master + bundled ZooKeeper running on the master host, a node manager running on every host. The admin UI is reachable on :8080. This page walks through your first login, sanity-checking the topology, and installing the first managed component.

First login

Open http://<master-host>:8080/admin/ in a browser. Sign in with admin / admin. Chango will require a password change on first login.

You can also drive everything from the REST API:

BASE=http://<master-host>:8080
TOK=$(curl -s $BASE/admin/api/auth/login \
  -H 'Content-Type: application/json' \
  -d '{"username":"admin","password":"<your-password>"}' | jq -r .accessToken)

The rest of this page uses the admin UI; every action has an equivalent REST call.

Verify cluster topology

From the admin UI's left sidebar, open Cluster → Nodes. You should see:

  • One Master row with ready = true, isLeader = true.
  • One Node Manager row per host listed in the inventory, all ready = true.

REST equivalent:

curl -s -H "Authorization: Bearer $TOK" $BASE/admin/api/nodes/masters
curl -s -H "Authorization: Bearer $TOK" $BASE/admin/api/nodes/node-managers

If a node is missing, tail its log on the host:

# master host
tail -f /var/log/chango/master.log
tail -f /var/log/chango/zookeeper.log

# any node manager host
tail -f /var/log/chango/nodemanager.log

Install your first component — PostgreSQL

PostgreSQL is the smallest managed component chango ships, and several other layers (Polaris metastore, Trino resource groups, NeoRunBase coordinator catalog) can reuse the same instance.

  1. Sidebar → Components → PostgreSQL → Install.
  2. Pick a target node manager.
  3. Set a strong PostgreSQL superuser password.
  4. Click Install, then Start.

The status moves through INSTALLED → STARTING → RUNNING. The pid and live CPU / memory show up within seconds.

REST equivalent:

curl -sm 240 -X POST -H "Authorization: Bearer $TOK" \
  -H 'Content-Type: application/json' \
  -d '{"instanceId":"pg-1","nodeId":"<nm-nodeId>","password":"<strong-password>"}' \
  $BASE/admin/api/postgres

curl -s -X POST -H "Authorization: Bearer $TOK" $BASE/admin/api/postgres/pg-1/start

The component runs as a system user (postgres) and stores its data under /opt/components/pg-1/data. Chango remembers the superuser password and reuses it when another component (Polaris, NeoRunBase, …) asks for a chango-installed PG.

Install a lakehouse stack — Ontul + ShannonStore + Polaris + Trino

For a minimum lakehouse stack, install these in order:

  1. Ontul — IAM authority that the other components delegate authorization to.
  2. ShannonStore — S3-compatible object storage for Iceberg data and Trino exchange.
  3. Polaris — Iceberg REST catalog (uses the PostgreSQL you installed above).
  4. Trino — Distributed SQL with exchange manager (ShannonStore), resource groups (PostgreSQL), and authz (Ontul) wired in.

Each component has its own admin UI panel. The dependencies are auto-discovered:

  • Trino, Spark, Flink, and the Trino Gateway find Ontul automatically and only ask the operator for a long-lived OTOK token (issued from Ontul's IAM page).
  • Polaris and the Trino resource-group manager find the chango-installed PostgreSQL from a dropdown — no need to type host/port/credentials.

See Component Catalog for the full per-component install flow.

Where things live

Path on each host What it is
/opt/chango Chango master + node manager install dirs
/opt/chango/bin/*.pid Chango process pid files (master / zk / nodemanager, one per running instance)
/opt/components/<instanceId> Each managed component's install dir
/var/lib/chango Chango master's RocksDB stores (KMS, IAM, metadata)
/var/log/chango Chango master + NM + ZK logs

Per-component logs live under /opt/components/<instanceId>/logs/ (or the component-specific log path) and are also tailable from the admin UI's Logs panel.

Day-2 lifecycle

Stop, restart, add another master, add another node manager — all done by the operator running shell scripts directly on the host, not through systemctl. See Cluster Operations.

Next steps