Skip to content
Bash script to automate ZFS snapshots and sync to a local pool.
Shell
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
README.md
_config.yml
zfs-local-sync.sh

README.md

zfs-local-sync

Bash script to automate ZFS snapshot creation and syncing. Tested and working on Proxmox VE 5.2, should be compatible with any other distro using ZFS on Linux.

It may not look super fancy but it does it's job reliably. Comments and contributions are welcome :)

Use cases & requirements

This script can be used to automatically create snapshots and copy them to another local pool.

Limitations:

  • At least 2 ZFS pools are required; one with the content you want to make redundant, and another one to receive the datasets and snapshots.
  • The destination (backup) pool must have enough free space to allocate the source datasets and snapshots.
  • The destination pool must not have datasets with the same name as the source pool.

Installation

cd /opt && git clone https://github.com/delebru/zfs-local-sync.git

Parameters

Required parameters:

  • -s | --source: Name of source ZFS pool.
  • -d | --destination: Name of ZFS pool to use as backup.

Optional parameters:

  • -D | --dry-run => Test run: outputs commands to console but won't run. (overrides silent and verbose modes)
  • -S | --silent => Silent mode: only errors and warnings will be loged.
  • -v | --verbose => Verbose mode: sends output also to console. (overrides silent mode)
  • -k | --keep-snapshots: Number of snapshots to keep. Default: 10 Minimum: 1
  • -vols | --datasets: Limits the sync to only the specified dataset(s). Must enter value(s) between quotes (and separated with spaces). Example: --datasets "vm-1-disk-0 vm-1-disk-1"

First run

Please do a dry run (-D) to make sure the script will be doing what you expect it to.

To sync all the datasets in the source pool to the backup pool:

/bin/bash /opt/zfs-local-sync/zfs-local-sync.sh -s source_pool_name -d backup_pool_name -D

To sync only certain datasets (for example vm-1-disk-0 and vm-1-disk-1):

/bin/bash /opt/zfs-local-sync/zfs-local-sync.sh -s source_pool_name -d backup_pool_name -vols "vm-1-disk-0 vm-1-disk-1" -D

Automation

After doing a dry run and confirming the script will do what you need, add an entry to your crontab.

For example to sync every 15 minutes:

*/15 * * * * root /bin/bash /opt/zfs-local-sync/zfs-local-sync.sh -s source_pool_name -d backup_pool_name

To duplicate multiple pools, one entry per source pool is required. A single destination pool may be used for various source pools as long as it satisfies the requirements specified above (enough free space and non conflicting dataset names):

*/15 * * * * root /bin/bash /opt/zfs-local-sync/zfs-local-sync.sh -s source_pool_name -d backup_pool_name
*/15 * * * * root /bin/bash /opt/zfs-local-sync/zfs-local-sync.sh -s source_pool_name_2 -d backup_pool_name

Logs & output

By default the script will write the executed commands and any output to a log file. One log file will be created for every source pool:

/var/log/zfs-local-sync/source_pool_name.log

Special modes

  • Dry run: nothing will be executed nor any log file will be created. Only the commands to be run will be shown on console.
  • Verbose: commands and output will be go to both: log file and console.
  • Silent: only warnings will be written to the log file.
You can’t perform that action at this time.