diff --git a/sysadmin/grid5000/cassandra/99-clean-zfs.sh b/sysadmin/grid5000/cassandra/99-clean-zfs.sh
new file mode 100755
index 0000000..a8b040b
--- /dev/null
+++ b/sysadmin/grid5000/cassandra/99-clean-zfs.sh
@@ -0,0 +1,33 @@
+#!/usr/bin/env bash
+
+# set -eu
+
+SCRIPT_DIR="$(pwd $(dirname @0))"
+
+source "${SCRIPT_DIR}/environment.cfg"
+
+CASSANDRA_NODES="$(cat ${SCRIPT_DIR}/cassandra_seeds.lst | tr ',' '\n') "
+
+echo "*********************"
+echo "WARNING"
+echo "*********************"
+echo "Running this script will wipe out all the data of the cassandra cluster."
+echo "cassandra nodes: $(cat ${SCRIPT_DIR}/cassandra_seeds.lst)"
+echo "Are you sure?"
+read -r
+
+echo "Stopping cassandra"
+echo -n ${CASSANDRA_NODES} | parallel -v -d' ' ssh "${SSH_USER}"@{} systemctl stop cassandra 
+
+for node in ${CASSANDRA_NODES}; do
+    echo "Removing zfs datasets on ${node}"
+    POOLS="$(ssh "${SSH_USER}"@${node} zpool list -o name | grep -v -e NAME -e "no pool")"
+    echo ${POOLS}
+    for pool in ${POOLS}; do
+        echo "   Removing ${pool}"
+        ssh "${SSH_USER}"@${node} zpool destroy $pool
+    done
+done
+
+echo "Done"
+echo "Don't forget to re-run ansible or 03-deploy_nodes.sh to reconfigure the environment"