DISNIX-RESTORE(1)

📦 version 0.10.2, in nixos 26.05, nixpkgs unstable

💅 disnix-restore — the disnix-restore(1) glow-up

okay so bestie, welcome to disnix-restore, a User Commands girlie from the disnix fam. this is the page where you learn how to put the state of your components back the way it was, and honestly? it's giving time machine energy ✨

🌸 SYNOPSIS 🌸

disnix-restore [OPTION] [MANIFEST]

🦋 DESCRIPTION 🦋

So. You've got a whole network of machines and components deployed across them, and you wanna restore the state of components deployed in that network. That's literally the whole job. 🫶

By default — and this is the part people miss, pay attention — it only restores the state of components that are recently activated, i.e. the ones defined in the current deployment manifest but not the previous one. Lowkey selective. It's not doing a full sweep unless you tell it to.

Wanna do a full restore of the state of the entire system? iconic, just pass the --no-upgrade parameter and it'll do the whole thing. 👑

And if no manifest has been provided at all? no stress, it just uses the last deployed one. slay.

✨ OPTIONS ✨

okay here's the menu, bestie, order whatever you want:

  • -c, --container=CONTAINER — name of the container in which the mutable component is deployed 🎀
  • -C, --component=COMPONENT — name of the mutable component to take snapshots from
  • -o, --old-manifest=MANIFEST — Nix profile path where the manifest should be stored, so that Disnix knows the current configuration of a distributed system. By default it's stored in the profile directory of the user. iykyk 💜
  • --no-upgrade — indicates that no upgrade should be performed and the state of all components should be restored. this is the full-send flag 🔥
  • --transfer-only — transfers the snapshot from the target machines, but doesn't actually restore them. it's like ordering the food and not eating it, no cap 🙈
  • --depth-first — snapshots components depth-first as opposed to breadth-first. this approach is more space efficient, but slower. trade-offs, bestie, always trade-offs 🦋
  • --all — transfers all snapshot generations of the target machines, not just the latest 🧚
  • --keep=NUM — amount of snapshot generations to keep. defaults to: 1 🌸
  • -p, --profile=PROFILE — name of the profile in which the services are registered. defaults to: default 🎀
  • --coordinator-profile-path=PATH — path to the manifest of the previous configuration. by default this tool will use the manifest stored in the disnix coordinator profile instead of the specified one, which is usually sufficient in most cases 💖
  • -m, --max-concurrent-transfers=NUM — maximum amount of concurrent closure transfers. and here's the tea: the original manpage says "Defauls to: 2" — typo and all, we're keeping it, it's part of the lore now 💀
  • -h, --help — shows the usage of this command to the user. the classic 🫶

🌈 ENVIRONMENT 🌈

  • DISNIX_PROFILE — sets the name of the profile that stores the manifest on the coordinator machine and the deployed services per machine on each target. defaults to: default ⭐

Copyright © 2008-2022 Sander van der Burg

💅 real talk: those two "defaults to" values are load-bearing, okay? --keep defaults to 1, -p / --profile and DISNIX_PROFILE default to default, and -m defaults to 2. don't come for me if you ignored them.

and that's the whole vibe check, bestie 🎀 this page came to you courtesy of disnix 0.10.2, dated January 1980 (yes, really, 1980, we don't question the lore 💀) — go forth and restore responsibly, mwah 💖

📦 where this page comes from