local prefix:label > local prefix > inherited prefix:label > inherited prefix
snapshot.sh -h|--help
snapshot.sh -l|--label=label [-p|--prefix=auto] [-e|--force-empty] [-n|--dry-run] [zfs_dataset]...
-h, --help Shows help
-p, --prefix Default "auto". E.g. "somethingelse" for rpool/USERDATA@somethingelse_20230507-2245_hourly
-l, --label Label for finer resolution. E.g. "hourly" for rpool/USERDATA@auto_20230507-2245_hourly
-e, --force-empty Force creating of empty snapshots
-n, --dry-run Does not actually create snapshot
-d, --debug Debug mode (set -x)
-v, --verbose Verbose mode
Note: On FreeBSD is supported only short form of params.
Creates snapshots by defined prefix and label recursively.
Its behaviour is controlled by zfs dataset property cz.solctech:snapshot:<prefix>:<label>.
Value of property specifies if snapshot creation has to be done.
((on|yes|true)|(off|no|false))[,(no-dive|nodive)]
on
off
on,no-dive
Allow snapshot creation for default prefix auto and label hourly for whole rpool.
zfs set cz.solctech:snapshot:auto:hourly=on rpoolDon't want auto snapshots for anything in rpool/STORAGE/docker, even don't want traverse into child datasets.
zfs set cz.solctech:snapshot:auto:hourly=no-dive rpool/STORAGE/dockerSpecify correct prefix, specify datasets (or omit for all locally imported), and try dry run.
snapshot.sh --dry-run --label=hourly rpoolIf everything seems ok, put in cron.
purge.sh -h|--help \
purge.sh -p|--prefix=snapshot_prefix [-n|--dry-run] [zfs_dataset]...
-h, --help Shows help
-p, --prefix E.g. "mybackup" for rpool/USERDATA@mybackup_20221002-23
-n, --dry-run Calls zfs destroy with -n argument
-d, --debug Debug mode (set -x)
-v, --verbose Verbose mode
Note: On FreeBSD is supported only short form of params.
Purges snapshots by defined prefix recursively.
Its behaviour is controlled by zfs dataset property cz.solctech:purge:<prefix>. Value of property specifies if purging has to be done and how many snapshots and/or how many days have to be kept back.
((on|yes|true)|(off|no|false))[,keepnum=#num[,keepdays=#num]]
on
on,keepdays=10
on,keepnum=3,keepdays=20
off
Turn on of purging for prefix backup for whole rpool, keeping minimal of 5 snapshots per dataset and keeping snapshots not older than 25 days.
zfs set cz.solctech:purge:backup=on,keepnum=5,keepdays=25 rpoolFor USERDATA keep more history (min. 10 snapshots and last 60 days).
zfs set cz.solctech:purge:backup=on,keepnum=10,keepdays=60 rpool/USERDATAFor Projects turn off purging - no snapshots will be destroyed.
zfs set cz.solctech:purge:backup=off rpool/USERDATA/myuser/ProjectsSpecify correct prefix, specify datasets (or omit for all locally imported), and try dry run.
purge.sh --dry-run --prefix=backup rpoolIf everything seems ok, put in daily or weekly cron.
release.sh <snapshot1> [snapshot2 ...]
Releases all holds on given list of snapshots.
Release all holds on snapshot rpool@backup_123.
$ release.sh rpool@backup_123
Releasing holds for snapshot: rpool@backup_123
Hold tag: last_backup