diff --git a/scripts/cloudron-support b/scripts/cloudron-support
index 390294e7b..6dc731e8a 100755
--- a/scripts/cloudron-support
+++ b/scripts/cloudron-support
@@ -476,7 +476,6 @@ function do_recreate_containers() {
     echo -n "This takes a while ."
     line_count=$(cat "${logfile}" | wc -l)
     sed -e 's/"version": ".*",/"version":"48.0.0",/' -i /home/yellowtent/platformdata/INFRA_VERSION
-    systemctl enable -q box
     systemctl restart -q box # will re-create docker network
 
     while ! tail -n "+${line_count}" "${logfile}" | grep -q "platform is ready"; do
@@ -516,31 +515,29 @@ function ask_reboot() {
 
     [[ ! $choice =~ ^[Yy]$ ]] && exit 1
     reboot
+    exit
 }
 
 function recreate_docker() {
     readonly logfile="/home/yellowtent/platformdata/logs/box.log"
     readonly stagefile="/home/yellowtent/platformdata/recreate-docker-stage"
 
-    docker_root=/var/lib/docker
-    if ! docker_root=$(docker info -f '{{ .DockerRootDir }}'); then
+    if ! docker_root=$(docker info -f '{{ .DockerRootDir }}' 2>/dev/null); then
         warning "Unable to detect docker root. Assuming /var/lib/docker"
     fi
+    [[ -z "${docker_root}" ]] && docker_root="/var/lib/docker"
 
     if [[ ! -e "${stagefile}" ]]; then
         echo -e "Use this command when docker storage (at $docker_root) is corrupt. It will delete the docker storage, re-download docker images and re-create containers. Dashboard and apps will be unreachable for a while. No data will be lost.\n"
-        echo -e "The server will have to be rebooted twice for this process.\n"
+        echo -e "The server may have to be rebooted twice for this process. If so, re-run this command after every reboot.\n"
         read -p "Do you want to proceed? (y/N) " -n 1 -r choice
         echo -e "\n"
 
         [[ ! $choice =~ ^[Yy]$ ]] && exit 1
 
-        info "Stopping box"
-        systemctl disable -q box || true
-        systemctl stop -q box || true
-        info "Stopping docker"
-        systemctl disable -q docker || true # for the reboot situation, we don't want it start again
-        systemctl stop -q docker || true
+        info "Stopping box and docker"
+        systemctl stop -q box docker containerd docker.socket || true
+        systemctl disable -q box docker containerd docker.socket || true
 
         echo -e "clearing_storage" > "${stagefile}" # init
     fi
@@ -548,24 +545,24 @@ function recreate_docker() {
     if grep -q "clearing_storage" "${stagefile}"; then
         info "Clearing docker storage at ${docker_root}"
         if ! rm -rf "${docker_root}/"*; then
-            echo -e "\nThe server has to be rebooted to clear the docker storage. After reboot, run this command again.\n"
+            echo -e "\nThe server has to be rebooted to clear the docker storage. After reboot, run 'cloudron-support --recreate-docker' again.\n"
             ask_reboot
         fi
         echo -e "cleared_storage" > "${stagefile}"
     fi
 
     if grep -q "cleared_storage" "${stagefile}"; then
-        info "Starting docker afresh"
-        systemctl enable -q docker
-        systemctl start -q docker
+        info "Starting docker afresh at ${docker_root}"
+        systemctl enable --now -q docker.socket docker containerd
         sleep 5 # give docker some time to initialize the storage directory
         download_docker_images
         echo -e "downloaded_images" > "${stagefile}"
-        echo -e "\nThe server has to be rebooted again for docker to initialize properly. After reboot, run this command again.\n" # else docker network is not completely functional
+        echo -e "\nThe server has to be rebooted again for docker to initialize properly. After reboot, run 'cloudron-support --recreate-docker' again.\n" # else docker network is not completely functional
         ask_reboot
     fi
 
     if grep -q "downloaded_images" "${stagefile}"; then
+        systemctl enable -q box
         do_recreate_containers
     fi