Sanoid is a policy-driven snapshot management tool for ZFS filesystems. When combined with the Linux KVM hypervisor, you can use it to make your systems functionally immortal.

^{(Real time demo: rolling back a full-scale cryptomalware infection in seconds!)}

More prosaically, you can use Sanoid to create, automatically thin, and monitor snapshots and pool health from a single eminently human-readable TOML config file at /etc/sanoid/sanoid.conf. (Sanoid also requires a "defaults" file located at /etc/sanoid/sanoid.defaults.conf, which is not user-editable.) A typical Sanoid system would have a single cron job but see INSTALL.md for more details:

* * * * * TZ=UTC /usr/local/bin/sanoid --cron

Note: Using UTC as timezone is recommended to prevent problems with daylight saving times

And its /etc/sanoid/sanoid.conf might look something like this:

[data/home]
    use_template = production
[data/images]
    use_template = production
    recursive = yes
    process_children_only = yes
[data/images/win7]
    hourly = 4

#############################
# templates below this line #
#############################

[template_production]
        frequently = 0
        hourly = 36
        daily = 30
        monthly = 3
        yearly = 0
        autosnap = yes
        autoprune = yes

Which would be enough to tell sanoid to take and keep 36 hourly snapshots, 30 dailies, 3 monthlies, and no yearlies for all datasets under data/images (but not data/images itself, since process_children_only is set). Except in the case of data/images/win7, which follows the same template (since it's a child of data/images) but only keeps 4 hourlies for whatever reason.

For more full details on sanoid.conf settings see Wiki page.

Note: Be aware that if you don't specify some interval options the defaults will be used (from /etc/sanoid/sanoid.defaults.conf)

Sanoid Command Line Options

• --cron

  This will process your sanoid.conf file, create snapshots, then purge expired ones.

• --configdir

  Specify a location for the config file named sanoid.conf. Defaults to /etc/sanoid

• --cache-dir

  Specify a directory to store the zfs snapshot cache. Defaults to /var/cache/sanoid

• --run-dir

  Specify a directory for temporary files such as lock files. Defaults to /var/run/sanoid

• --take-snapshots

  This will process your sanoid.conf file, create snapshots, but it will NOT purge expired ones. (Note that snapshots taken are atomic in an individual dataset context, not a global context - snapshots of pool/dataset1 and pool/dataset2 will each be internally consistent and atomic, but one may be a few filesystem transactions "newer" than the other.)

• --prune-snapshots

  This will process your sanoid.conf file, it will NOT create snapshots, but it will purge expired ones.

• --force-prune

  Purges expired snapshots even if a send/recv is in progress

• --monitor-snapshots

  This option is designed to be run by a Nagios monitoring system. It reports on the health of your snapshots.

• --monitor-health

  This option is designed to be run by a Nagios monitoring system. It reports on the health of the zpool your filesystems are on. It only monitors filesystems that are configured in the sanoid.conf file.

• --monitor-capacity

  This option is designed to be run by a Nagios monitoring system. It reports on the capacity of the zpool your filesystems are on. It only monitors pools that are configured in the sanoid.conf file.

• --force-update

  This clears out sanoid's zfs snapshot listing cache. This is normally not needed.

• --version

  This prints the version number, and exits.

• --quiet

  Suppress non-error output.

• --verbose

  This prints additional information during the sanoid run.

• --debug

  This prints out quite a lot of additional information during a sanoid run, and is normally not needed.

• --readonly

  Skip creation/deletion of snapshots (Simulate).

• --help

  Show help message.

Sanoid script hooks

There are three script types which can optionally be executed at various stages in the lifecycle of a snapshot:

pre_snapshot_script

Will be executed before the snapshot(s) of a single dataset are taken. The following environment variables are passed:

If the script returns a non-zero exit code, the snapshot(s) will not be taken unless no_inconsistent_snapshot is false.

post_snapshot_script

Will be executed when:

• The pre-snapshot script succeeded or
• The pre-snapshot script failed and force_post_snapshot_script is true.

pruning_script

Will be executed after a snapshot is successfully deleted. The following environment variables will be passed:

example

sanoid.conf:

...
[sanoid-test-0]
    use_template = production
    recursive = yes
    pre_snapshot_script = /tmp/debug.sh
    post_snapshot_script = /tmp/debug.sh
    pruning_script = /tmp/debug.sh
...

verbose sanoid output:

...
executing pre_snapshot_script '/tmp/debug.sh' on dataset 'sanoid-test-0'
taking snapshot sanoid-test-0@autosnap_2020-02-12_14:49:33_yearly
taking snapshot sanoid-test-0@autosnap_2020-02-12_14:49:33_monthly
taking snapshot sanoid-test-0@autosnap_2020-02-12_14:49:33_daily
taking snapshot sanoid-test-0@autosnap_2020-02-12_14:49:33_hourly
executing post_snapshot_script '/tmp/debug.sh' on dataset 'sanoid-test-0'
...

pre script env variables:

SANOID_SCRIPT=pre
SANOID_TARGET=sanoid-test-0/b/bb
SANOID_TARGETS=sanoid-test-0/b/bb
SANOID_SNAPNAME=autosnap_2020-02-12_14:49:32_yearly
SANOID_SNAPNAMES=autosnap_2020-02-12_14:49:32_yearly,autosnap_2020-02-12_14:49:32_monthly,autosnap_2020-02-12_14:49:32_daily,autosnap_2020-02-12_14:49:32_hourly
SANOID_TYPES=yearly,monthly,daily,hourly

post script env variables:

SANOID_SCRIPT=post
SANOID_TARGET=sanoid-test-0/b/bb
SANOID_TARGETS=sanoid-test-0/b/bb
SANOID_SNAPNAME=autosnap_2020-02-12_14:49:32_yearly
SANOID_SNAPNAMES=autosnap_2020-02-12_14:49:32_yearly,autosnap_2020-02-12_14:49:32_monthly,autosnap_2020-02-12_14:49:32_daily,autosnap_2020-02-12_14:49:32_hourly
SANOID_TYPES=yearly,monthly,daily,hourly
SANOID_PRE_FAILURE=0

Syncoid

Sanoid also includes a replication tool, syncoid, which facilitates the asynchronous incremental replication of ZFS filesystems. A typical syncoid command might look like this:

syncoid data/images/vm backup/images/vm

Which would replicate the specified ZFS filesystem (aka dataset) from the data pool to the backup pool on the local system, or

syncoid data/images/vm root@remotehost:backup/images/vm

Which would push-replicate the specified ZFS filesystem from the local host to remotehost over an SSH tunnel, or

syncoid root@remotehost:data/images/vm backup/images/vm

Which would pull-replicate the filesystem from the remote host to the local system over an SSH tunnel.

Syncoid supports recursive replication (replication of a dataset and all its child datasets) and uses mbuffer buffering, lzop compression, and pv progress bars if the utilities are available on the systems used. If ZFS supports resumable send/receive streams on both the source and target those will be enabled as default.

As of 1.4.18, syncoid also automatically supports and enables resume of interrupted replication when both source and target support this feature.

Syncoid Dataset Properties

• syncoid:sync

  Available values:

  • true (default if unset)

  This dataset will be synchronised to all hosts.

  • false

  This dataset will not be synchronised to any hosts - it will be skipped. This can be useful for preventing certain datasets from being transferred when recursively handling a tree.

  • host1,host2,...

  A comma separated list of hosts. This dataset will only be synchronised by hosts listed in the property.

  Note: this check is performed by the host running syncoid, thus the local hostname must be present for inclusion during a push operation // the remote hostname must be present for a pull.

  Note: this will also prevent syncoid from handling the dataset if given explicitly on the command line.

  Note: syncing a child of a no-sync dataset will currently result in a critical error.

  Note: empty properties will be handled as if they were unset.

Syncoid Command Line Options

• [source]

  This is the source dataset. It can be either local or remote.

• [destination]

  This is the destination dataset. It can be either local or remote.

• --identifier=

  Adds the given identifier to the snapshot and hold name after "syncoid" prefix and before the hostname. This enables the use case of reliable replication to multiple targets from the same host. The following chars are allowed: a-z, A-Z, 0-9, , -, : and . .

• -r --recursive

  This will also transfer child datasets.

• --skip-parent

  This will skip the syncing of the parent dataset. Does nothing without '--recursive' option.

• --compress

  Compression method to use for network transfer. Currently accepted options: gzip, pigz-fast, pigz-slow, zstd-fast, zstd-slow, lz4, xz, lzo (default) & none. If the selected compression method is unavailable on the source and destination, no compression will be used.

• --source-bwlimit <limit t|g|m|k>

  This is the bandwidth limit in bytes (kbytes, mbytes, etc) per second imposed upon the source. This is mainly used if the target does not have mbuffer installed, but bandwidth limits are desired.

• --target-bwlimit <limit t|g|m|k>

  This is the bandwidth limit in bytes (kbytes, mbytes, etc) per second imposed upon the target. This is mainly used if the source does not have mbuffer installed, but bandwidth limits are desired.

• --no-command-checks

  Does not check the existence of commands before attempting the transfer, providing administrators a way to run the tool with minimal overhead and maximum speed, at risk of potentially failed replication, or other possible edge cases. It assumes all programs are available, and should not be used in most situations. This is an not an officially supported run mode.

• --no-stream

  This argument tells syncoid to use -i incrementals, not -I. This updates the target with the newest snapshot from the source, without replicating the intermediate snapshots in between. (If used for an initial synchronization, will do a full replication from newest snapshot and exit immediately, rather than starting with the oldest and then doing an immediate -i to the newest.)

• --no-sync-snap

  This argument tells syncoid to restrict itself to existing snapshots, instead of creating a semi-ephemeral syncoid snapshot at execution time. Especially useful in multi-target (A->B, A->C) replication schemes, where you might otherwise accumulate a large number of foreign syncoid snapshots.

• --keep-sync-snap

  This argument tells syncoid to skip pruning old snapshots created and used by syncoid for replication if '--no-sync-snap' isn't specified.

• --create-bookmark

  This argument tells syncoid to create a zfs bookmark for the newest snapshot after it got replicated successfully. The bookmark name will be equal to the snapshot name. Only works in combination with the --no-sync-snap option. This can be very useful for irregular replication where the last matching snapshot on the source was already deleted but the bookmark remains so a replication is still possible.

• --use-hold This argument tells syncoid to add a hold to the newest snapshot on the source and target after replication succeeds and to remove the hold after the next succesful replication. Setting a hold prevents the snapshots from being destroyed. The hold name incldues the identifier if set. This allows for separate holds in case of replication to multiple targets.

• --preserve-recordsize

  This argument tells syncoid to set the recordsize on the target before writing any data to it matching the one set on the replication src. This only applies to initial sends.

• --delete-target-snapshots

  With this argument snapshots which are missing on the source will be destroyed on the target. Use this if you only want to handle snapshots on the source. Note that snapshot deletion is only done after a successful synchronization. If no new snapshots are found, no synchronization is done and no deletion either.

• --no-clone-rollback

  Do not rollback clones on target

• --no-rollback

  Do not rollback anything (clones or snapshots) on target host

• --exclude=REGEX

  The given regular expression will be matched against all datasets which would be synced by this run and excludes them. This argument can be specified multiple times.

• --no-resume

  This argument tells syncoid to not use resumable zfs send/receive streams.

• --force-delete

  Remove target datasets recursively (WARNING: this will also affect child datasets with matching snapshots/bookmarks), if there are no matching snapshots/bookmarks. Also removes conflicting snapshots if the replication would fail because of a snapshot which has the same name between source and target but different contents.

• --no-clone-handling

  This argument tells syncoid to not recreate clones on the target on initial sync, and do a normal replication instead.

• --dumpsnaps

  This prints a list of snapshots during the run.

• --no-privilege-elevation

  Bypass the root check and assume syncoid has the necessary permissions (for use with ZFS permission delegation).

• --sshport

  Allow sync to/from boxes running SSH on non-standard ports.

• --sshcipher

  Instruct ssh to use a particular cipher set.

• --sshoption

  Passes option to ssh. This argument can be specified multiple times.

• --sshkey

  Use specified identity file as per ssh -i.

• --insecure-direct-connection=IP:PORT[,IP:PORT,[TIMEOUT,[mbuffer]]]

  WARNING: This is an insecure option as the data is not encrypted while being sent over the network. Only use if you trust the complete network path. Use a direct tcp connection (with socat and busybox nc/mbuffer) for the actual zfs send/recv stream. All control commands are still executed via the ssh connection. The first address pair is used for connecting to the target host from the source host and the second pair is for listening on the target host. If the later isn't provided the same as the former is used. This can be used for saturating high throughput connection like >= 10GBe network which isn't easy with the overhead off ssh. It can also be useful for encrypted datasets to lower the cpu usage needed for replication but be aware that metadata is NOT ENCRYPTED in this case. The default timeout is 60 seconds and can be overridden by providing it as third argument. By default busybox nc is used for the listeing tcp socket, if mbuffer is preferred specify its name as fourth argument but be aware that mbuffer listens on all interfaces and uses an optionally provided ip address for access restriction (This option can't be used for relaying between two remote hosts)

• --quiet

  Suppress non-error output.

• --debug

  This prints out quite a lot of additional information during a sanoid run, and is normally not needed.

• --help

  Show help message.

• --version

  Print the version and exit.

• --monitor-version

  This doesn't do anything right now.

Note that the sync snapshots syncoid creates are not atomic in a global context: sync snapshots of pool/dataset1 and pool/dataset2 will each be internally consistent, but one may be a few filesystem transactions "newer" than the other. (This does not affect the consistency of snapshots already taken in other ways, which syncoid replicates in the overall stream unless --no-stream is specified. So if you want to manually zfs snapshot -R pool@1 before replicating with syncoid, the global atomicity of pool/dataset1@1 and pool/dataset2@1 will still be intact.)