Hello all,

I have enabled snapshots of a number of ZFS datasets. Now I want to monitor the space consumed by the snapshots and adjust the number of monthly snapshots I will keep so as to not run out of space.

The command "zfs list -t snapshot" gives me the revelant information for individual snapshots, but I am looking for a way to get the consolidated used space of all snapshots.

Thank you.

snapshot plugin and zfs master plugin

user script plugin cron monthly:

https://forums.unraid.net/topic/178033-bmartino1-user-scripts/#findComment-1478664

we can even edit it based on sanpshot size...

Edited January 8Jan 8 by bmartino1

exampls updated script with size storage space restrictions.

#!/usr/bin/env bash
set -euo pipefail

######################## simple-snapshot-zfs #######################
###################### User Defined Options #######################

DATASETS=("vm-zfs/Backups")
SNAPSHOT_QTY=5
MAX_SNAPSHOT_SPACE="500G"   # HARD LIMIT PER DATASET

##################################################################

timestamp=$(date "+%Y-%m-%d-%H:%M")
echo "Starting Snapshot ${timestamp}"
echo "_____________________________________________________________"

create_snapshot_if_changed() {
  local DATASET="$1"
  local WRITTEN

  WRITTEN=$(zfs get -H -o value written "${DATASET}")

  if [[ "${WRITTEN}" != "0" ]]; then
    local TS
    TS="$(date '+%Y-%m-%d-%H%M')"
    zfs snapshot -r "${DATASET}@${TS}"
    echo "Recursive snapshot created: ${DATASET}@${TS}"
  else
    echo "No changes detected in ${DATASET}. No snapshot created."
  fi
}

prune_by_count() {
  local DATASET="$1"
  local KEEP="${SNAPSHOT_QTY}"

  mapfile -t SNAPSHOTS < <(
    zfs list -t snapshot -o name -s creation -r "${DATASET}" |
    grep "^${DATASET}@"
  )

  local COUNT=${#SNAPSHOTS[@]}
  echo "Snapshot count for ${DATASET}: ${COUNT}"

  if (( COUNT > KEEP )); then
    local TO_DELETE=$((COUNT - KEEP))
    for snap in "${SNAPSHOTS[@]:0:${TO_DELETE}}"; do
      zfs destroy "${snap}"
      echo "Deleted snapshot (count limit): ${snap}"
    done
  fi
}

prune_by_space() {
  local DATASET="$1"

  while true; do
    local USED
    USED=$(zfs get -H -o value usedbysnapshots "${DATASET}")

    echo "Snapshot space used by ${DATASET}: ${USED}"

    if ! zfs get -H -o value usedbysnapshots "${DATASET}" \
         | awk -v max="${MAX_SNAPSHOT_SPACE}" '
           BEGIN { exit (system("numfmt --from=iec " $0) > system("numfmt --from=iec " max)) }'
    then
      break
    fi

    local OLDEST
    OLDEST=$(zfs list -t snapshot -o name -s creation -r "${DATASET}" |
             grep "^${DATASET}@" | head -n 1)

    [[ -z "${OLDEST}" ]] && break

    zfs destroy "${OLDEST}"
    echo "Deleted snapshot (space limit): ${OLDEST}"
  done
}

for dataset in "${DATASETS[@]}"; do
  create_snapshot_if_changed "${dataset}"
  prune_by_count "${dataset}"
  prune_by_space "${dataset}"
  echo "_____________________________________________________________"
done

echo "----------------------------Done!----------------------------"

What This Script Now Guarantees

✔ Snapshots only created if data changed
✔ Snapshot count limit enforced
✔ Snapshot space limit enforced
✔ Oldest snapshots removed first
✔ Fully recursive-safe
✔ Uses correct ZFS accounting (usedbysnapshots)

Per-snapshot USED is not additive → don’t sum it.

1. The only authoritative metric is:
   zfs get usedbysnapshots DATASET

   

Edited January 8Jan 8 by bmartino1

Thanks a lot, I'll take a look at these scripts.