Jun 28, 2026 · 15 min · 10 segments
In this episode of Scaling Postgres, we discuss a new backup tool called pg\_hardstorage by Cybertech, bottomless storage & ColdFront, monitoring checkpoints and tracking statistics. To get the show…
Creston JamisonHost
There is a blog post we'll look at, but this is a new website that has been spun up called pghardstorage.org, and the actual tool is called pghardstorage.

It's an open-source PostgreSQL backup tool, and this is a project backed by CyberTech, and it says it is version 1.

And for a project I haven't heard of, it has the most comprehensive website for it I've ever seen, talking about all the capabilities and features and why it was done.

You connect up to Postgres via replication slot that receives the wall stream.

And just the number of docs and examples is truly impressive for, again, something I haven't really heard of before.

And looking at the how it works, you basically create a dedicated role on your database and grant it pgReadAllData.

And it has a content address stored, so every backup is a full, essentially, and dedupe is free.

They set up the connection and where you want it to write the files, the repo, essentially.

And this is just a command line tool to take a base back up and then to start reading from the wall stream.

And looking at some of the examples, they give an example of a Linux server and a Petroni cluster.

There is a blog post we'll look at, but this is a new website that has been spun up called pghardstorage.org, and the actual tool is called pghardstorage.

It's an open-source PostgreSQL backup tool, and this is a project backed by CyberTech, and it says it is version 1.

And for a project I haven't heard of, it has the most comprehensive website for it I've ever seen, talking about all the capabilities and features and why it was done.

You connect up to Postgres via replication slot that receives the wall stream.

And just the number of docs and examples is truly impressive for, again, something I haven't really heard of before.

And looking at the how it works, you basically create a dedicated role on your database and grant it pgReadAllData.

And it has a content address stored, so every backup is a full, essentially, and dedupe is free.

They set up the connection and where you want it to write the files, the repo, essentially.

And this is just a command line tool to take a base back up and then to start reading from the wall stream.

And looking at some of the examples, they give an example of a Linux server and a Petroni cluster.
Every episode on Radar is fully transcribed, speaker-labeled, and rich with metadata. Here is a taste of this one. Try Radar for free to see the rest.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
All 10 segments — the transcript broken into labeled sections, every ad read marked
All 9 topics — jump to every other episode discussing the same subject
Every related episode — other shows Radar links to this one
No account is needed to search Radar.