Database guides

Greenplum

Updated 2026-08-01 · 2 min read

Massively-parallel analytics built on PostgreSQL. AddisDB connects with the Postgres driver, with introspection relaxed for Greenplum’s older catalog.

Greenplum is PostgreSQL forked and rebuilt for massively parallel processing: one coordinator plans a query and many segment servers execute it in parallel. It speaks the PostgreSQL wire protocol, so AddisDB connects with its Postgres driver on port 5432.

Who it is for

Greenplum suits analytical workloads too large for a single Postgres node but where you would rather keep Postgres SQL, Postgres types and Postgres tooling than move to a proprietary warehouse.

The whole performance model comes down to one clause. Every table has a distribution key, and rows are hashed across segments by it. A join between two tables distributed on the same key runs entirely inside each segment; a join on anything else moves data across the network first. DISTRIBUTED RANDOMLY is the fallback when no column is a good spreader.

It is a poor fit for high-concurrency transactional traffic — the parallel execution model is built for scanning a lot of data per query, not for many small ones.

Set up the cluster

  1. Note the coordinator host — that is the only node clients connect to; the segments are internal.
  2. Confirm the coordinator port. It is 5432 by default, though many deployments move it to 6432 to sit alongside a stock Postgres.
  3. Make sure your client address is allowed in pg_hba.conf and that the role you will use has SELECT on the schemas you care about.
  4. Ask which Greenplum major version you are on — 6 and 7 differ enough that the catalog answers different questions.
# Confirm the coordinator is reachable on the port you were given
psql -h coordinator.example.com -p 5432 -U gpadmin -d gpadmin -c 'SELECT version();'

Connect from AddisDB

  1. New Connection → Greenplum. Port prefills to 5432, database and username to gpadmin.
  2. Enter the coordinator host and your role’s credentials.
  3. Set SSL mode to require if the cluster terminates TLS.
  4. Test, then Save.

What AddisDB gives you

  • Schema browsing, the foreign-key diagram, and in-grid editing behind the Edit toggle.
  • Mock data generated from your real schema, plus AI querying and query repair.
  • Distribution and partition clauses are preserved when AddisDB shows you existing DDL — it reads definitions rather than reconstructing them.
  • The Chart view, ⌘K search, and the full safety model.

That last point matters here more than on most engines: DISTRIBUTED BY and the partition hierarchy are the two things you most need to see when reading someone else’s Greenplum schema, and a tool that reconstructs DDL from catalog columns usually drops both.