Snapper
Snapper is a tool created by openSUSE's Arvin Schnell that helps with managing snapshots of Btrfs subvolumes and thin-provisioned LVM volumes. It can create and compare snapshots, revert between snapshots, and supports automatic snapshots timelines.
Installation
Install the snapper package. The development version snapper-gitAUR is also available.
Additionally, GUIs are available with snapper-gui-gitAUR and btrfs-assistantAUR.
Creating a new configuration
Before creating a snapper configuration for a Btrfs subvolume, the subvolume must already exist. If it does not, you should create it before generating a snapper configuration.
To create a new snapper configuration named config for the Btrfs subvolume at /path/to/subvolume, run:
# snapper -c config create-config /path/to/subvolume
This will:
- Create a configuration file at
/etc/snapper/configs/configbased on the default template from . - Create a subvolume at where future snapshots for this configuration will be stored. A snapshot's path is , where is the snapshot number.
- Add
configtoSNAPPER_CONFIGSin/etc/conf.d/snapper.
For example, to create a configuration file for the subvolume mounted at , run:
# snapper -c root create-config /
At this point, the configuration is active. If your cron daemon is running, snapper will take #Automatic timeline snapshots. If you do not use a cron daemon, you will need to use the systemd service and timer. See #Enable/disable.
See also .
Taking snapshots
Automatic timeline snapshots
A snapshot timeline can be created with a configurable number of hourly, daily, weekly, monthly, and yearly snapshots kept. When the timeline is enabled, by default a snapshot gets created once an hour. Once a day the snapshots get cleaned up by the timeline cleanup algorithm. Refer to the variables in for details.
Enable/disable
If you have a cron daemon, this feature should start automatically. To disable it, edit the configuration file corresponding with the subvolume you do not want to have this feature and set:
TIMELINE_CREATE="no"
If you do not have a cron daemon, you can use the provided systemd units. Start and enable to start the automatic snapshot timeline. Additionally, start and enable to periodically clean up older snapshots.
Set snapshot limits
The default settings will keep 10 hourly, 10 daily, 10 monthly and 10 yearly snapshots. You may want to change this in the configuration, especially on busy subvolumes like . See #Preventing slowdowns.
Here is an example section of a configuration named config with only 5 hourly snapshots, 7 daily ones, no monthly and no yearly ones:
Change snapshot and cleanup frequencies
If you are using the provided systemd timers, you can edit them to change the snapshot and cleanup frequency.
For example, when editing the , add the following to make the frequency every five minutes, instead of hourly:
[Timer] OnCalendar= OnCalendar=*:0/5
When editing , you need to change . To make cleanups occur every hour instead of every day, add:
[Timer] OnUnitActiveSec=1h
See systemd/Timers and systemd#Drop-in files.
Single snapshots
By default snapper takes snapshots that are of the single type, having no special relationship to other snapshots.
To take a snapshot of a subvolume manually, do:
# snapper -c config create --description desc
The above command does not use any cleanup algorithm, so the snapshot is stored permanently or until deleted.
To set a cleanup algorithm, use the -c flag after and choose either , timeline, , or . sets snapper to periodically remove snapshots that have exceeded a set number in the configuration file. For example, to create a snaphot that uses the algorithm for cleanup do:
# snapper -c config create -c number
See #Automatic timeline snapshots for how timeline snapshots work and see #Pre/post snapshots on how and work.
Pre/post snapshots
The other type of snapshots - pre/post snapshots - are intended to be created as a pair, one before and one after a significant change (such as a system update).
If the significant change is/can be invoked by a single command, then can be used to invoke the command and automatically create pre/post snapshots:
# snapper -c config create --command cmd
Alternatively, the pre/post snapshots can be created manually.
First create a pre snapshot:
# snapper -c config create -t pre -p
Note the number of the new snapshot (it is required to create the post snapshot).
Now perform the actions that will modify the filesystem (*e.g.*, install a new program, upgrade, etc.).
Finally, create the post snapshot, replacing with the number of the pre snapshot:
# snapper -c config create -t post --pre-number N
Snapshots on boot
To have snapper take a snapshot of the configuration, enable . (These snapshots are of type single.)
Managing snapshots
List snapshots
To list snapshots taken for a given configuration config do:
# snapper -c config list
Delete a snapshot
To delete a snapshot number do:
# snapper -c config delete N
Multiple snapshots can be deleted at one time. For example, to delete snapshots 65 and 70 of the root configuration do:
# snapper -c root delete 65 70
To delete a range of snapshots, in this example between snapshots 65 and 70 of the root configuration do:
# snapper -c root delete 65-70
To free the space used by the snapshot(s) immediately, use --sync:
# snapper -c root delete --sync 65
Access for non-root users
Each config is created with the root user, and by default, only root can see and access it.
To be able to list the snapshots for a given config for a specific user, simply change the value of in your /etc/snapper/configs/config file. You should now be able to run as a normal user.
Eventually, you want to be able to browse the .snapshots directory with a user, but the owner of this directory must stay root. Therefore, you should change the group owner by a group containing the user you are interested in, such as for example:
# chmod a+rx .snapshots
# chown :users .snapshots
Tips and tricks
Wrapping pacman transactions in snapshots
There are a couple of packages used for automatically creating snapshots upon a pacman transaction:
Backup non-Btrfs boot partition on pacman transactions
If your /boot partition is on a non Btrfs filesystem (e.g. an ESP) you are not able to do snapper backups with it. See System backup#Snapshots and /boot partition to copy the boot partition automatically on a kernel update to your Btrfs root with a hook. This also plays nice together with .
Incremental backup to external drive
Some tools can use snapper to automate backups. See Btrfs#Incremental backup to external drive.
Suggested filesystem layout
Here is a suggested file system layout for easily restoring the subvolume that is mounted at root to a previous snapshot:
| Subvolume | Mountpoint |
|---|---|
| @ | / |
| @home | /home |
| @snapshots | /.snapshots |
| @var_log | /var/log |
subvolid=5 | ├── @ -| | contained directories: | ├── /usr | ├── /bin | ├── /.snapshots | ├── ... | ├── @home ├── @snapshots ├── @var_log └── @...
The subvolumes are mounted to any other directory that should have its own subvolume.
- When taking a snapshot of
@(mounted at the root/), other subvolumes are not included in the snapshot. Even if a subvolume is nested below@, a snapshot of@will not include it. Create snapper configurations for additional subvolumes besides@of which you want to keep snapshots. - Due to a Btrfs limitation, snapshotted volumes cannot contain swap files. Either put the swap file on another subvolume or create a swap partition.
If you were to restore your system to a previous snapshots of , these other subvolumes will remain unaffected. For example, this allows you to restore to a previous snapshot while keeping your unchanged, because of the subvolume that is mounted at .
This layout allows the snapper utility to take regular snapshots of , while at the same time making it easy to restore from an Arch Live CD if it becomes unbootable.
In this scenario, after the initial setup, snapper needs no changes, and will work as expected.
Configuration of snapper and mount point
It is assumed that the subvolume is mounted at root . It is also assumed that is not mounted and does not exist as folder, this can be ensured by the commands:
# umount /.snapshots # rm -r /.snapshots
Then create a new configuration for . Snapper create-config automatically creates a subvolume .snapshots with the root subvolume as its parent, that is not needed for the suggested filesystem layout, and can be deleted.
# btrfs subvolume delete /.snapshots
After deleting the subvolume, recreate the directory .
# mkdir /.snapshots
Now mount to . For example, for a file system located on :
# mount -o subvol=@snapshots /dev/sda1 /.snapshots
To make this mount permanent, add an entry to your fstab.
Or if you have an existing fstab entry remount the snapshot subvolume:
# mount -a
Give the folder 750 permissions.
This will make all snapshots that snapper creates be stored outside of the subvolume, so that can easily be replaced anytime without losing the snapper snapshots.
Restoring / to its previous snapshot
To restore using one of snapper's snapshots, first boot into a live Arch Linux USB/CD.
Mount the toplevel subvolume (subvolid=5). That is, omit any or mount flags.
Find the number of the snapshot that you want to recover:
grep -r '<date>' /mnt/@snapshots/*/info.xml
The output should look like so, there is one line for each snapshot, so you can easily match up number and date of each snapshot.
/mnt/@snapshots/number/info.xml: <date>2021-07-26 22:00:00</date>
Remember the .
Now, move to another location (e.g. ) to save a copy of the current system. Alternatively, simply delete using .
Create a read-write snapshot of the read-only snapshot snapper took:
# btrfs subvolume snapshot /mnt/@snapshots/number/snapshot /mnt/@
Where is the number of the snapper snapshot you wish to restore.
If subvolid was used for the mount entry option in fstab, instead of , change subvolid in the to the new subvolid.
# btrfs subvolume list /mnt | grep @$ # Check the new subvolid # grep "/ " /mnt/@/etc/fstab # Check the old subvolid # vim /mnt/@/etc/fstab # Edit
Also change the bootloader configuration such as refind_linux.conf, if it contains the subvolid.
Finally, unmount the top-level subvolume (ID=5), then mount to /mnt and your ESP or boot partition to the appropriate mount point. Change root to your restored snapshot in order to regenerate your initramfs image.
Your has now been restored to the previous snapshot. Now just simply reboot.
Deleting files from snapshots
If you want to delete a specific file or folder from past snapshots without deleting the snapshots themselves, is a script that adds this functionality to Snapper. This script can also be used to manipulate past snapshots in a number of other ways that Snapper does not currently support.
If you want to remove a file without using an extra script, you just need to make your snapshot subvolume read-write, which you can do with:
# btrfs property set /path/to/.snapshots/<snapshot_num>/snapshot ro false
Verify that ro=false:
# btrfs property get /path/to/.snapshots/<snapshot_num>/snapshot ro=false
You can now modify files in like normal. You can use a shell loop to work on your snapshots in bulk.
Preventing slowdowns
Keeping many snapshots for a large timeframe on a busy filesystem like , where many system updates happen over time, can cause serious slowdowns. You can prevent it by:
- Creating subvolumes for things that are not worth being snapshotted, like , , , and .
- Editing the default settings for hourly/daily/monthly/yearly snapshots when using #Automatic timeline snapshots.
updatedb
By default, (see mlocate) will also index the .snapshots directory created by snapper, which can cause serious slowdown and excessive memory usage if you have many snapshots. You can prevent from indexing over it by editing:
Disable quota groups
There are reports of significant slow downs being caused by quota groups, if for instance snapper ls takes many minutes to return a result this could be the cause. See .
To determine whether or not quota groups are enabled use the following command:
# btrfs qgroup show /
Quota groups can then be disabled with:
# btrfs quota disable /
Count the number of snapshots
If disabling quota groups did not help with slow down, it may be helpful to count the number of snapshots, this can be done with:
# btrfs subvolume list -s / | wc -l
Preserving log files
It is recommended to create a subvolume for so that snapshots of exclude it. That way if a snapshot of is restored your log files will not also be reverted to the previous state. This makes it easier to troubleshoot.
Troubleshooting
Snapper logs
Snapper writes all activity to /var/log/snapper.log - check this file first if you think something goes wrong.
If you have issues with hourly/daily/weekly snapshots, the most common cause for this so far has been that the cronie service (or whatever cron daemon you are using) was not running.
IO error
If you get an 'IO Error' when trying to create a snapshot please make sure that the .snapshots directory associated to the subvolume you are trying to snapshot is a subvolume by itself.
Another possible cause is that .snapshots directory does not have root as an owner (You will find in the /var/log/snapper.log).
Orphaned snapshots causing wasted disk space
It is possible for snapshots to get 'lost', where they still exist on disk but are not tracked by snapper. This can result in a large amount of wasted, unaccounted-for disk space. To check for this, compare the output of
# snapper -c <config> list
to
# btrfs subvolume list -o <parent subvolume>/.snapshots
Any subvolume in the second list which is not present in the first is an orphan and can be deleted manually.