Further improvements to the new service installation code. (#13774)

* Improve OpenRC detection.

- Check for rc-update command, without that we don’t work so it’s ‘not
  OpenRC’.
- Sort checks so that less expensive ones run first.
- If we see `/etc/init.d/local` and it’s got a shebang line with
  `openrc-run`, then assume we’re dealing with OpenRC. This file gets
  installed by default as part of OpenRC itself.

* Fix runit detection.

- `/lib/rc/sv.d` is only used on some runit systems (notably Artix), not
  all, so don’t bail if we can’t find it.
- Check not just for a `runit` command, but also `runsvdir`.

* Add support for systemd in WSL.

THis just involves running the normal systemd check if we detected WSL,
and updating the service manager type appropriately.

* Fix tmpdir handling for service installation.

* Fix invocation of service install script.

* Add fallback to detected but not running service managers.

Currently supported for systemd and OpenRC. If no other service manager
is detected, but we detected that either systemd or OpenRC is present
but not running, fall back to them instead of outright failing.

This does not allow starting/stopping the agent using the service manager,
but _does_ happen to work for installing the service properly on an
offline system.

* Fix service-type fallback handling.

If we see systemd, even if it not booted into it, we should not install
init.d or LSB init scripts.

* Fix recursion bug in initd install type checking.

* Skip initd if OpenRC is found.

Just like systemd, any OpenRC on the system automatically means that
it’s _not_ a classic init.d setup.

* Fall back to init.d in WSL in some cases.

If there is no _usable_ systemd, and the service manager otherwise looks
like an init.d setup, fall back to an init.d setup.

* Support LSB style init scripts in WSL as well.

Also consolidate and update the error message that gets printed when
running under WSL.

* Fix typo in LSB init script handling code.

* Fix command handling for LSB and initd service management.

Instead of using `service`, just run the init scripts directly. This way
we avoid any bugs in `service` that would cause it to pretend that
systemd is in use even though it isn’t.

* Fix botched partial commit.

* Restructure service type detection to not rely on return codes.

* Added option to display service type detection results.

This should simplify debugging.

1 files changed, 199 insertions, 118 deletions

diff --git a/system/install-service.sh.in b/system/install-service.sh.in
index 6c7c018878..ecaf931ff4 100755
--- a/system/install-service.sh.in
+++ b/system/install-service.sh.in
@@ -29,10 +29,12 @@ DUMP_CMDS=0
 ENABLE="auto"
 EXPORT_CMDS=0
 INSTALL=1
-LINUX_INIT_TYPES="wsl systemd openrc lsb initd runit"
+LINUX_INIT_TYPES="systemd openrc lsb initd runit"
 PLATFORM="$(uname -s)"
+SHOW_SVC_TYPE=0
 SVC_SOURCE="@libsysdir_POST@"
 SVC_TYPE="detect"
+WSL_ERROR_MSG="We appear to be running in WSL and were unable to find a usable service manager. We currently support systemd, LSB init scripts, and traditional init.d style init scripts when running under WSL."
 
 # =====================================================================
 # Utility functions
@@ -157,6 +159,7 @@ USAGE: install-service.sh [options]
 
   --source      Specify where to find the service files to install (default ${SVC_SOURCE}).
   --type        Specify the type of service file to install. Specify a type of 'help' to get a list of valid types for your platform.
+  --show-type   Display information about what service managers are detected.
   --cmds        Additionally print a list of commands for starting and stopping the agent with the detected service type.
   --export-cmds Export the variables that would be printed by the --cmds option.
   --cmds-only   Don't install, just handle the --cmds or --export-cmds option.
@@ -176,30 +179,27 @@ HEREDOC
 # =====================================================================
 # systemd support functions
 
-issystemd() {
+_check_systemd() {
   pids=''
   p=''
   myns=''
   ns=''
-  systemctl=''
 
   # if the directory /lib/systemd/system OR /usr/lib/systemd/system (SLES 12.x) does not exit, it is not systemd
   if [ ! -d /lib/systemd/system ] && [ ! -d /usr/lib/systemd/system ]; then
-    return 1
+    echo "NO" && return 0
   fi
 
   # if there is no systemctl command, it is not systemd
-  systemctl=$(command -v systemctl 2> /dev/null)
-  if [ -z "${systemctl}" ] || [ ! -x "${systemctl}" ]; then
-    return 1
-  fi
+  [ -z "$(command -v systemctl 2>/dev/null || true)" ] && echo "NO" && return 0
 
   # if pid 1 is systemd, it is systemd
-  [ "$(basename "$(readlink /proc/1/exe)" 2> /dev/null)" = "systemd" ] && return 0
+  [ "$(basename "$(readlink /proc/1/exe)" 2> /dev/null)" = "systemd" ] && echo "YES" && return 0
 
-  # if systemd is not running, it is not systemd
+  # it ‘is’ systemd at this point, but systemd might not be running
+  # if not, return 2 to indicate ‘systemd, but not running’
   pids=$(safe_pidof systemd 2> /dev/null)
-  [ -z "${pids}" ] && return 1
+  [ -z "${pids}" ] && echo "OFFLINE" && return 0
 
   # check if the running systemd processes are not in our namespace
   myns="$(readlink /proc/self/ns/pid 2> /dev/null)"
@@ -207,20 +207,19 @@ issystemd() {
     ns="$(readlink "/proc/${p}/ns/pid" 2> /dev/null)"
 
     # if pid of systemd is in our namespace, it is systemd
-    [ -n "${myns}" ] && [ "${myns}" = "${ns}" ] && return 0
+    [ -n "${myns}" ] && [ "${myns}" = "${ns}" ] && echo "YES" && return 0
   done
 
   # else, it is not systemd
-  return 1
+  echo "NO"
 }
 
 check_systemd() {
   if [ -z "${IS_SYSTEMD}" ]; then
-    issystemd
-    IS_SYSTEMD="$?"
+    IS_SYSTEMD="$(_check_systemd)"
   fi
 
-  return "${IS_SYSTEMD}"
+  echo "${IS_SYSTEMD}"
 }
 
 get_systemd_service_dir() {
@@ -259,50 +258,61 @@ install_systemd_service() {
     exit 4
   fi
 
-  if ! systemctl daemon-reload; then
-    warning "Failed to reload systemd unit files."
-  fi
+  if check_systemd; then
+    if ! systemctl daemon-reload; then
+        warning "Failed to reload systemd unit files."
+    fi
 
-  if ! systemctl ${ENABLE} netdata; then
-    warning "Failed to ${ENABLE} Netdata service."
+    if ! systemctl ${ENABLE} netdata; then
+        warning "Failed to ${ENABLE} Netdata service."
+    fi
   fi
 }
 
 systemd_cmds() {
-  NETDATA_START_CMD='systemctl start netdata'
-  NETDATA_STOP_CMD='systemctl stop netdata'
+  if check_systemd; then
+    NETDATA_START_CMD='systemctl start netdata'
+    NETDATA_STOP_CMD='systemctl stop netdata'
+  else # systemd is not running, use external defaults by providing no commands
+    warning "Detected systemd, but not booted using systemd. Unable to provide commands to start or stop Netdata using the service manager."
+  fi
 }
 
 # =====================================================================
 # OpenRC support functions
 
-isopenrc() {
+_check_openrc() {
   # if /lib/rc/sh/functions.sh does not exist, it's not OpenRC
-  [ ! -f /lib/rc/sh/functions.sh ] && return 1
+  [ ! -f /lib/rc/sh/functions.sh ] && echo "NO" && return 0
 
   # if there is no /etc/init.d, it's not OpenRC
-  [ ! -d /etc/init.d ] && return 1
+  [ ! -d /etc/init.d ] && echo "NO" && return 0
 
-  # if PID 1 is openrc-init, it's OpenRC
-  [ "$(basename "$(readlink /proc/1/exe)" 2> /dev/null)" = "openrc-init" ] && return 0
+  # if there is no rc-update command, it's not OpenRC
+  [ -z "$(command -v rc-update 2>/dev/null || true)" ] && echo "NO" && return 0
 
   # If /run/openrc/softlevel exists, it's OpenRC
-  [ -f /run/openrc/softlevel ] && return 0
+  [ -f /run/openrc/softlevel ] && echo "YES" && return 0
+
+  # if PID 1 is openrc-init, it's OpenRC
+  [ "$(basename "$(readlink /proc/1/exe)" 2> /dev/null)" = "openrc-init" ] && echo "YES" && return 0
 
-  # if there is an openrc command, it's OpenRC
-  command -v openrc > /dev/null 2>&1 && return 0
+  # if there is an openrc command, it's OpenRC, but not booted as such
+  [ -n "$(command -v openrc 2>/dev/null || true)" ] && echo "OFFLINE" && return 0
+
+  # if /etc/init.d/local exists and has `openrc-run` in it's shebang line, it’s OpenRC, but not booted as such
+  [ -r /etc/init.d/local ] && head -n 1 /etc/init.d/local | grep -q openrc-run && echo "OFFLINE" && return 0
 
   # Otherwise, it’s not OpenRC
-  return 1
+  echo "NO" && return 0
 }
 
 check_openrc() {
   if [ -z "${IS_OPENRC}" ]; then
-    isopenrc
-    IS_OPENRC="$?"
+    IS_OPENRC="$(_check_openrc)"
   fi
 
-  return "${IS_OPENRC}"
+  echo "${IS_OPENRC}"
 }
 
 enable_openrc() {
@@ -329,39 +339,49 @@ install_openrc_service() {
 }
 
 openrc_cmds() {
-  NETDATA_START_CMD='rc-service netdata start'
-  NETDATA_STOP_CMD='rc-service netdata stop'
+  if check_openrc; then
+    NETDATA_START_CMD='rc-service netdata start'
+    NETDATA_STOP_CMD='rc-service netdata stop'
+  else # Not booted using OpenRC, use external defaults by not providing commands.
+    warning "Detected OpenRC, but the system is not booted using OpenRC. Unable to provide commands to start or stop Netdata using the service manager."
+  fi
 }
 
 # =====================================================================
 # LSB init script support functions
 
-islsb() {
+_check_lsb_ignore_systemd() {
   # if there is no /etc/init.d directory, it’s not an LSB system
-  [ ! -d /etc/init.d ] && return 1
+  [ ! -d /etc/init.d ] && echo "NO" && return 0
 
   # If it's an OpenRC system, then it's not an LSB system
-  check_openrc && return 1
+  [ "$(check_openrc)" != "NO" ] && echo "NO" && return 0
 
   # If /lib/lsb/init-functions exists, it’s an LSB system
-  [ -f /lib/lsb/init-functions ] && return 0
+  [ -f /lib/lsb/init-functions ] && echo "YES" && return 0
+
+  echo "NO" && return 0
+}
 
-  return 1
+_check_lsb() {
+  # if there is _any_ systemd, it’s not an LSB system
+  [ "$(check_systemd)" != "NO" ] && echo "NO" && return 0
+
+  _check_lsb_ignore_systemd
 }
 
 check_lsb() {
   if [ -z "${IS_LSB}" ]; then
-    islsb
-    IS_LSB="$?"
+    IS_LSB="$(_check_lsb)"
   fi
 
-  return "${IS_LSB}"
+  echo "${IS_LSB}"
 }
 
 enable_lsb() {
   if ! update-rc.d netdata defaults; then
     warning "Failed to enable Netdata service."
-  elif ! update-rc.d netdata defaults-disable; then
+  elif ! update-rc.d netdata defaults-disabled; then
     warning "Failed to fully enable Netdata service."
   fi
 }
@@ -377,38 +397,42 @@ install_lsb_service() {
 }
 
 lsb_cmds() {
-  if command -v service >/dev/null 2>&1; then
-    NETDATA_START_CMD='service netdata start'
-    NETDATA_STOP_CMD='service netdata stop'
-  else
-    NETDATA_START_CMD='/etc/init.d/netdata start'
-    NETDATA_STOP_CMD='/etc/init.d/netdata stop'
-  fi
+  NETDATA_START_CMD='/etc/init.d/netdata start'
+  NETDATA_STOP_CMD='/etc/init.d/netdata stop'
 }
 
 # =====================================================================
 # init.d init script support functions
 
-isinitd() {
+_check_initd_ignore_systemd() {
   # if there is no /etc/init.d directory, it’s not an init.d system
-  [ ! -d /etc/init.d ] && return 1
+  [ ! -d /etc/init.d ] && echo "NO" && return 1
 
   # if there is no chkconfig command, it's not a (usable) init.d system
-  command -v chkconfig >/dev/null 2>&1 || return 1
+  [ -z "$(command -v chkconfig 2>/dev/null || true)" ] && echo "NO" && return 0
+
+  # if there is _any_ openrc, it’s not init.d
+  [ "$(check_openrc)" != "NO" ] && echo "NO" && return 0
 
   # if it's not an LSB setup, it’s init.d
-  check_initd || return 0
+  [ "$(check_lsb)" != "NO" ] && echo "NO" && return 0
 
-  return 1
+  echo "YES" && return 0
+}
+
+_check_initd() {
+  # if there is _any_ systemd, it’s not init.d
+  [ "$(check_systemd)" != "NO" ] && echo "NO" && return 0
+
+  _check_initd_ignore_systemd
 }
 
 check_initd() {
   if [ -z "${IS_INITD}" ]; then
-    isinitd
-    IS_INITD="$?"
+    IS_INITD="$(_check_initd)"
   fi
 
-  return "${IS_INITD}"
+  echo "${IS_INITD}"
 }
 
 enable_initd() {
@@ -428,13 +452,8 @@ install_initd_service() {
 }
 
 initd_cmds() {
-  if command -v service >/dev/null 2>&1; then
-    NETDATA_START_CMD='service netdata start'
-    NETDATA_STOP_CMD='service netdata stop'
-  else
-    NETDATA_START_CMD='/etc/init.d/netdata start'
-    NETDATA_STOP_CMD='/etc/init.d/netdata stop'
-  fi
+  NETDATA_START_CMD='/etc/init.d/netdata start'
+  NETDATA_STOP_CMD='/etc/init.d/netdata stop'
 }
 
 # =====================================================================
@@ -442,29 +461,28 @@ initd_cmds() {
 #
 # Currently not supported, this exists to provide useful error messages.
 
-isrunit() {
-  # if there is no /lib/rc/sv.d, then it's not runit
-  [ ! -d /lib/rc/sv.d ] && return 1
+_check_runit() {
+  # if there is no runsvdir command, then it's not runit
+  [ -z "$(command -v runsvdir 2>/dev/null || true)" ] && echo "NO" && return 0
 
   # if there is no runit command, then it's not runit
-  command -v runit >/dev/null 2>&1 || return 1
+  [ -z "$(command -v runit 2>/dev/null || true)" ] && echo "NO" && return 0
 
   # if /run/runit exists, then it's runit
-  [ -d /run/runit ] && return 0
+  [ -d /run/runit ] && echo "YES" && return 0
 
   # if /etc/runit/1 exists and is executable, then it's runit
-  [ -x /etc/runit/1 ] && return 0
+  [ -x /etc/runit/1 ] && echo "YES" && return 0
 
-  return 1
+  echo "NO" && return 0
 }
 
 check_runit() {
   if [ -z "${IS_RUNIT}" ]; then
-    isrunit
-    IS_RUNIT="$?"
+    IS_RUNIT="$(_check_runit)"
   fi
 
-  return "${IS_RUNIT}"
+  echo "${IS_RUNIT}"
 }
 
 install_runit_service() {
@@ -482,34 +500,33 @@ runit_cmds() {
 #
 # Cannot be supported, this exists to provide useful error messages.
 
-iswsl() {
+_check_wsl() {
   # If uname -r contains the string WSL, then it's WSL.
-  uname -r | grep -q 'WSL' && return 0
+  uname -r | grep -q 'WSL' && echo "YES" && return 0
 
   # If uname -r contains the string Microsoft, then it's WSL.
   # This probably throws a false positive on CBL-Mariner, but it's part of what MS officially recommends for
   # detecting if you're running under WSL.
-  uname -r | grep -q "Microsoft" && return 0
+  uname -r | grep -q "Microsoft" && echo "YES" && return 0
 
-  return 1
+  echo "NO" && return 0
 }
 
 check_wsl() {
   if [ -z "${IS_WSL}" ]; then
-    iswsl
-    IS_WSL="$?"
+    IS_WSL="$(_check_wsl)"
   fi
 
-  return "${IS_WSL}"
+  echo "${IS_WSL}"
 }
 
 install_wsl_service() {
-  error "We appear to be running in WSL. Netdata cannot be automatically installed as a service under WSL."
+  error "${WSL_ERROR_MSG}"
   exit 3
 }
 
 wsl_cmds() {
-  error "We appear to be running in WSL. Netdata cannot be automatically installed as a service under WSL."
+  error "${WSL_ERROR_MSG}"
   exit 3
 }
 
@@ -564,21 +581,47 @@ darwin_cmds() {
 
 detect_linux_svc_type() {
   if [ "${SVC_TYPE}" = "detect" ]; then
-    for t in ${LINUX_INIT_TYPES}; do
-      if "check_${t}"; then
-        SVC_TYPE="${t}"
-        break
-      fi
+    found_types=''
+
+    for t in wsl ${LINUX_INIT_TYPES}; do
+      case "$("check_${t}")" in
+        YES)
+          SVC_TYPE="${t}"
+          break
+          ;;
+        NO) continue ;;
+        OFFLINE)
+          if [ -z "${found_types}" ]; then
+            found_types="${t}"
+          else
+            found_types="${found_types} ${t}"
+          fi
+          ;;
+      esac
     done
 
     if [ "${SVC_TYPE}" = "detect" ]; then
-      error "Failed to detect what type of service manager is in use."
-    else
-      echo "${SVC_TYPE}"
+      if [ -z "${found_types}" ]; then
+        error "Failed to detect what type of service manager is in use."
+      else
+        SVC_TYPE="$(echo "${found_types}" | cut -f 1 -d ' ')"
+        warning "Failed to detect a running service manager, using detected (but not running) ${SVC_TYPE}."
+      fi
+    elif [ "${SVC_TYPE}" = "wsl" ]; then
+      if [ "$(check_systemd)" = "YES" ]; then
+        # Support for systemd in WSL.
+        SVC_TYPE="systemd"
+      elif [ "$(_check_lsb_ignore_systemd)" = "YES" ]; then
+        # Support for LSB init.d in WSL.
+        SVC_TYPE="lsb"
+      elif [ "$(_check_initd_ignore_systemd)" = "YES" ]; then
+        # Support for ‘generic’ init.d in WSL.
+        SVC_TYPE="initd"
+      fi
     fi
-  else
-    echo "${SVC_TYPE}"
   fi
+
+  echo "${SVC_TYPE}"
 }
 
 install_linux_service() {
@@ -602,6 +645,38 @@ linux_cmds() {
 }
 
 # =====================================================================
+# Service type display function
+
+show_service_type() {
+  info "Detected platform: ${PLATFORM}"
+
+  case "${PLATFORM}" in
+    FreeBSD)
+      info "Detected service managers:"
+      info "  - freebsd: YES"
+      info "Would use freebsd service management."
+      ;;
+    Darwin)
+      info "Detected service managers:"
+      info "  - launchd: YES"
+      info "Would use launchd service management."
+      ;;
+    Linux)
+      [ "$(check_wsl)" = "YES" ] && info "Detected WSL environment."
+      info "Detected service managers:"
+      for t in ${LINUX_INIT_TYPES}; do
+        info "  - ${t}: $("check_${t}")"
+      done
+      info "Would use $(detect_linux_svc_type) service management."
+      ;;
+    *)
+      info "${PLATFORM} is not supported by this script. No service file would be installed."
+  esac
+
+  exit 0
+}
+
+# =====================================================================
 # Argument handling
 
 parse_args() {
@@ -620,6 +695,7 @@ parse_args() {
           shift 1
         fi
         ;;
+      "--show-type") SHOW_SVC_TYPE=1 ; INSTALL=0 ;;
       "--save-cmds")
         if [ -z "${2}" ]; then
           info "No path specified to save command variables."
@@ -671,28 +747,33 @@ main() {
 
   parse_args "${@}"
 
-  case "${PLATFORM}" in
-    FreeBSD)
-      [ "${INSTALL}" -eq 1 ] && install_freebsd_service
-      freebsd_cmds
-      ;;
-    Darwin)
-      [ "${INSTALL}" -eq 1 ] && install_darwin_service
-      darwin_cmds
-      ;;
-    Linux)
-      [ "${INSTALL}" -eq 1 ] && install_linux_service
-      linux_cmds
-      ;;
-    *)
-      error "${PLATFORM} is not supported by this script."
-      exit 5
-      ;;
-  esac
+  if [ "${SHOW_SVC_TYPE}" -eq 1 ]; then
+    show_service_type
+  else
+    case "${PLATFORM}" in
+      FreeBSD)
+        [ "${INSTALL}" -eq 1 ] && install_freebsd_service
+        freebsd_cmds
+        ;;
+      Darwin)
+        [ "${INSTALL}" -eq 1 ] && install_darwin_service
+        darwin_cmds
+        ;;
+      Linux)
+        [ "${INSTALL}" -eq 1 ] && install_linux_service
+        linux_cmds
+        ;;
+      *)
+        error "${PLATFORM} is not supported by this script."
+        exit 5
+        ;;
+    esac
+
+    [ "${DUMP_CMDS}" -eq 1 ] && dump_cmds
+    [ "${EXPORT_CMDS}" -eq 1 ] && export_cmds
+    [ -n "${SAVE_CMDS_PATH}" ] && save_cmds
+  fi
 
-  [ "${DUMP_CMDS}" -eq 1 ] && dump_cmds
-  [ "${EXPORT_CMDS}" -eq 1 ] && export_cmds
-  [ -n "${SAVE_CMDS_PATH}" ] && save_cmds
   exit 0
 }