---
title: "Discussion on \"keep-daily=7 doesn't mean 7 days of backups\" | Hashnode"
description: "Discussion on \"keep-daily=7 doesn't mean 7 days of backups\". Proxmox and PBS retention keeps the newest backup in each calendar bucket, not a span of days. How a month of failed jobs can prune away your..."
url: "https://hashnode.com/posts/keep-daily-7-doesn-t-mean-7-days-of-backups/6af3d8cd0a79f9d85a3c4e1c"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [keep-daily=7 doesn't mean 7 days of backups](https://homelabtherightway.hashnode.dev/keep-daily-7-doesn-t-mean-7-days-of-backups?utm_source=hashnode&utm_medium=feed)

Proxmox and PBS retention keeps the newest backup in each calendar bucket, not a span of days. How a month of failed jobs can prune away your...

20h ago · 6 min read · homelabtherightway.hashnode.dev  
Read the article

-   [#homelab](https://hashnode.com/tag/homelab)
-   [#proxmox](https://hashnode.com/tag/proxmox)
-   [#backup](https://hashnode.com/tag/backup)

## Written by

[Charles Hartmann](https://hashnode.com/@charleshartmann)

@charleshartmann · New York, USA

  
[View profile](https://hashnode.com/@charleshartmann)

Homelab the Right Way: Proxmox, OPNsense, Home Assistant, backups and Android apps.

I run Proxmox, OPNsense and Home Assistant at home and write about doing it the right way: backups that restore, networks that are locked down, automations that just work. I also ship Android apps on Google Play as Hartmann Studios. Articles here are cross-posts from my blog (canonical links point there), drafted with AI help and edited by me.

## Responses 1

Markdown supported · ⌘↵ to post

[Mario Fang](https://hashnode.com/@mariosfong)

[10h ago](https://hashnode.com/posts/keep-daily-7-doesn-t-mean-7-days-of-backups/6af3d8cd0a79f9d85a3c4e1c/comment/6af46d70852156dd76956225)

The calendar-bucket point caught me out — I had always assumed keep-daily=7 meant seven days of coverage, not seven buckets that can stretch or shrink depending on when jobs actually ran. I have had a backup job fail quietly for weeks before, and the idea that the next successful run could then prune away the older copies is exactly the kind of thing you only discover when you need a restore. I now try to take a protected copy before any big change. Do you set that protected backup manually before an upgrade, or have you automated it somehow?