Ask HN: Any good tools to pgdump multi tenant database?

3•polishdude20•12mo ago

Just joined a new company and they run a multi tenant database in postgres where each tenant is distinguished by their account ID. I'd like to be able to dump the data of a single account to my local instance to assist in debugging.

Problem is, many of the entities do not include an account ID. For example, a user can have an account ID but a user post will not but it will be tied to a user ID.

Also foreign keys aren't really implemented fully.

Are there any tools or techniques that could help make this process easier to do?

Comments

tudorg•12mo ago

You can try with data subsetting tools, a few that I know are: greenmask, jailer, tonic condenser, and I think there are more. They are not exactly for your use case, but might help you a bit. The problem is most of these tools can walk foreign keys, but only in one directions.

Also, since you said FK are not fully implemented, then there just won't be enough information for them to work with. So you can either work on fixing the foreign keys or have more complex scripts.

polishdude20•12mo ago

Thanks for the suggestions!

Yeah due to the unavailability of FK's, I'd have to somehow map certain ID columns to the tables they're for, accounting for slight naming convention differences.

woyten•11mo ago

In case you don’t have FKs you could specify virtual foreign keys in Greenmask.

Check out docs for details

https://docs.greenmask.io/latest/database_subset/#virtual-re...

abhisek•12mo ago

This is poor design. Multi-tenant data model design must have a tenant or segment ID for every tenant segmented tables. While it may not be a problem initially but as the business grows, you may need to move tenant data to separate DBs or even different regions based on compliance requirements. IMHO it is a good idea to run DB migrations and have a tenant ID populated in all tenant segmented tables

polishdude20•12mo ago

Oh yes I agree with you. It may become a big push to convince my org to do that

ScyllaDB MCP Server: What AI-Native Developer Distribution Looks Like

Toto 2.0: Time series forecasting enters the scaling era

Show HN: I built a free AI startup validator – it scored a dog walking app 4/10

Coal is rising in China's clean energy transition

MINA BTP

Crypto Prediction Markets Explained

Prolog Basics Explained with Pokémon

Chinese short dramas became AI content machines

ACM ByteCast Interview with Sendmail Developer Eric Allman

O(x)Caml in Space

The Secret World of Tiny Phones That Go Inside Your Butt (2016)

MVRV Regime Monitor – the only crypto strategy that survived held-out validation

The Sigmoids Won't Save You

WxRuby3 v1.70 Released

Hello Robot's Wheeled Home Robot Ditches Humanoid Hype

Clippy can send emails, write excel sheets, and interact with any application

Developer Experience Is a Performance Feature

Steve Jobs Next Computer: His Forgotten Exile Years

SDL Library Adds Support for the New Steam Controller Without Depending on Steam

Justice Department Investigation Determines Yale Discriminated Based on Race

NanoTDB – Golang Append-Only Time Series DB

Known by Their Actions: Fingerprinting LLM Browser Agents via UI Traces

Tachyons Neo – Utility CSS without build step

Show HN: OrcaSheets, local first analytics engine to process billions of rows

How to Enter Side Doors

Nairobi became a nexus for the black market in giant harvester ants

Scryve-tools – Unified wallet auth for CKB, EVM, and BTC in one NPM package

The Earliest Known Dentistry Wasn't Done by Our Species

Show HN: Sanjaya – Academic paper discovery and extraction (OpenAlex/Scrapy)

Rest alone doesn't fix burnout. Here's the structural reason why