---
category:
  - "[[Note]]"
  - "[[Software]]"
type:
title: " Sanoid and Syncoid Setup"
created: 2026-05-12T21:09:25
tags:
  - project
---
## resources

- [github](https://github.com/jimsalterjrs/sanoid)
- [forum](https://discourse.practicalzfs.com/)

## intro

I've set this up on on a [[Proxmox]] backup machine, it creates the [[OpenZFS|zfs]] snapshots on the remote machine and them pulls them into it's own zfs pool.

## initial setup

install Sanoid on backup and source host (not strictly required on both, but Sanoid uses tools on both machine for fast and reliable connection)
```
apt install sanoid
```

if you don't run Sanoid on one machine, ensure the following folders are created on it
```
mkdir -p /var/cache/sanoid /var/run/sanoid
```

use root accounts for ssh connections, the zfs commands need root access (tried to get it to work without and it was turtles all the way down)

add conf files and populate from [github](https://github.com/jimsalterjrs/sanoid)
```
/etc/sanoid/sanoid.conf
/etc/sanoid/sanoid.defaults.conf
/usr/local/bin/zfs-nightly-backup.sh
```

run manual backup (add `--debug` flag for more info)
```
syncoid -r --use-hold --preserve-recordsize --preserve-properties root@192.168.1.43:noggapool/music stanleypool/music
```

commands for nightly script
```
/usr/sbin/syncoid -r --use-hold --preserve-recordsize --preserve-properties root@pve-thinkstation:noggapool/dataset stanleypool/dataset
```
```
/usr/sbin/syncoid -r --no-sync-snap --create-bookmark --use-hold --preserve-recordsize --preserve-properties stanleypool/dataset root@pve-shug9:bathpool/dataset
```

## scheduled snapshots and prune

by default, when installing via `apt` it starts a timer service that executes every 15minutes. for my requirements i need to disable this for now
```
systemctl stop timer.sanoid.service
systemctl disable time.sanoid.service
```

## on zfs bookmarks

The creator of sanoid, Jim Salter, gave a good overview of bookmarks [here](https://discourse.practicalzfs.com/t/can-sanoid-leave-bookmarks-behind-after-deleting-snaphots/4381/4)

> [!NOTE] Jim Salter
> Bookmarks are only useful in a **single** scenario:
> 
> When you already have a common snapshot between source and target–note, you must **already** have the common snapshot on both sides–you may first create a bookmark **on the source** and then destroy the snapshot **on the source** without breaking the replication chain.
> 
> However, the snapshot still has to be present **in full** on the target.
> 
> ...
> 
> This allows you to more easily keep much shallower archive depth on a source system with a few (or one) small, fast drives, while keeping much deeper archives on a target system with many more, or larger drives.