[GH-ISSUE #36] [Feature] pg_dump optimization #27

New issue

Open

opened 2026-02-26 21:34:04 +03:00 by kerem · 0 comments

kerem commented

2026-02-26 21:34:04 +03:00

Owner

Originally created by @berkayyildi on GitHub (Sep 4, 2024).
Original GitHub issue: https://github.com/eduardolat/pgbackweb/issues/36

Hello eduardo,
This is one of the project which I want to develop. Thanks for developing this project.
I couldn't try it now my environments are not ready now, but I want to give suggestion to you.
pg_dump without parameter is not efficient, as I see in project codes you get dump then compress it.
Instead of this, you can use directory format backup.

https://www.reddit.com/r/PostgreSQL/comments/1ckewh8/comment/l2mwerg/?utm_source=share&utm_medium=web3x&utm_name=web3xcss&utm_term=1&utm_content=share_button

pg_dump -v -Fd -Z1 --blobs --jobs=40

This is the most efficient backup method most of the time. And you don't need additional compression.
I think adding these methods to UI (plain backup or job count with directory format) will decrease backup size and increase backup speed for TB size databases.

Thanks

Originally created by @berkayyildi on GitHub (Sep 4, 2024). Original GitHub issue: https://github.com/eduardolat/pgbackweb/issues/36 Hello eduardo, This is one of the project which I want to develop. Thanks for developing this project. I couldn't try it now my environments are not ready now, but I want to give suggestion to you. pg_dump without parameter is not efficient, as I see in project codes you get dump then compress it. Instead of this, you can use directory format backup. https://www.reddit.com/r/PostgreSQL/comments/1ckewh8/comment/l2mwerg/?utm_source=share&utm_medium=web3x&utm_name=web3xcss&utm_term=1&utm_content=share_button pg_dump -v -Fd -Z1 --blobs --jobs=40 This is the most efficient backup method most of the time. And you don't need additional compression. I think adding these methods to UI (plain backup or job count with directory format) will decrease backup size and increase backup speed for TB size databases. Thanks