hn-support · March 25, 2020 14:33 · Jul 12, 2017 · Jul 12, 2017 · Jul 12, 2017
diff --git a/searchd-watchdog.py b/searchd-watchdog.py
@@ -80,12 +80,8 @@ def pid_exists(pid):
     """
     if pid < 0:
         return False
-    if pid == 0:
-        # According to "man 2 kill" PID 0 refers to every process
-        # in the process group of the calling process.
-        # On certain systems 0 is a valid PID but we have no way
-        # to know that in a portable fashion.
-        raise ValueError('invalid PID 0')
+    if pid == 0 or pid == 1:
+        raise ValueError('invalid PID 0 or 1')
     try:
         os.kill(pid, 0)
     except OSError as err:

diff --git a/searchd-watchdog.py b/searchd-watchdog.py
@@ -14,7 +14,7 @@
     * * * * * flock -n ~/.searchd_live    /data/web/searchd-watchdog.py --environment live    --version 2   (live)
     * * * * * flock -n ~/.searchd_staging /data/web/searchd-watchdog.py --environment staging --version 2   (staging)
 
-    July 12 2017 - Flip Hes <flip@byte.nl>
+    July 12 2017 - Flip Hes <support@byte.nl>
 """
 import argparse
 import errno

diff --git a/searchd-watchdog.py b/searchd-watchdog.py
@@ -0,0 +1,130 @@
+#!/usr/bin/env python
+""" Hypernode searchd cron watchdog
+
+    This script checks the pid of the searchd daemon and restarts it when the service is not running
+    It works for both Magento 1 and Magento 2 and for both live and staging.
+
+    To run it, add it to the crontab of your hypernode.
+
+    For Magento 1:
+    * * * * * flock -n ~/.searchd_live    /data/web/searchd-watchdog.py --environment live    --version 1   (live)
+    * * * * * flock -n ~/.searchd_staging /data/web/searchd-watchdog.py --environment staging --version 1   (staging)
+
+    For Magento 2:
+    * * * * * flock -n ~/.searchd_live    /data/web/searchd-watchdog.py --environment live    --version 2   (live)
+    * * * * * flock -n ~/.searchd_staging /data/web/searchd-watchdog.py --environment staging --version 2   (staging)
+
+    July 12 2017 - Flip Hes <[email protected]>
+"""
+import argparse
+import errno
+import logging
+import os
+import sys
+import subprocess
+
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger('searchd-watchdog')
+
+
+CONFIG_DICT = {
+    "staging": {
+        1: {
+            "config": "/data/web/staging/var/sphinx/sphinx.conf",
+            "pid": "/data/web/staging/var/sphinx/searchd.pid"
+        },
+        2: {
+            "config": "/data/web/magento2_staging/var/sphinx/sphinx.conf",
+            "pid": "/data/web/magento2_staging/var/sphinx/searchd.pid"
+        },
+    },
+    "live": {
+        1: {
+            "config": "/data/web/public/var/sphinx/sphinx.conf",
+            "pid": "/data/web/public/var/sphinx/searchd.pid"
+        },
+        2: {
+            "config": "/data/web/magento2/var/sphinx/sphinx.conf",
+            "pid": "/data/web/magento2/var/sphinx/searchd.pid"
+        }
+    }
+}
+
+
+def parse_arguments():
+    parser = argparse.ArgumentParser(description="Searchd watchdog")
+    parser.add_argument("--env", dest="environment", type=str, required=True, choices=CONFIG_DICT.keys(),
+                        help="The environment to watch")
+    parser.add_argument('--version', dest='version', type=int, required=True, choices=[1, 2],
+                        help="The magento version to use")
+    parser.add_argument("--log", dest="logfile", type=str, default="/data/web/sphinx.log",
+                        help="The log file to write all output to")
+    arguments = parser.parse_args()
+
+    return arguments
+
+
+def read_pid_file(filename):
+    if not filename or not os.path.isfile(filename):
+        return None
+    try:
+        with open(filename) as fh:
+            return int(fh.read())
+    except:
+        return None
+
+
+def pid_exists(pid):
+    """Check whether pid exists in the current process table.
+       UNIX only.
+    """
+    if pid < 0:
+        return False
+    if pid == 0:
+        # According to "man 2 kill" PID 0 refers to every process
+        # in the process group of the calling process.
+        # On certain systems 0 is a valid PID but we have no way
+        # to know that in a portable fashion.
+        raise ValueError('invalid PID 0')
+    try:
+        os.kill(pid, 0)
+    except OSError as err:
+        if err.errno == errno.ESRCH:
+            # ESRCH == No such process
+            return False
+        elif err.errno == errno.EPERM:
+            # EPERM clearly means there's a process to deny access to
+            return True
+        else:
+            # According to "man 2 kill" possible error values are
+            # (EINVAL, EPERM, ESRCH)
+            raise
+    else:
+        return True
+
+
+def start_sphinx(config, logfile):
+    command = '/usr/bin/nohup /usr/bin/searchd --config {} 2>&1 >> {} &'.format(config, logfile)
+    try:
+        subprocess.check_call(command, shell=True, stdout=subprocess.PIPE, stderr=subprocess.STDOUT, preexec_fn=os.setpgrp)
+    except subprocess.CalledProcessError:
+        logger.error("Failed to start searchd for config {}".format(config), file=sys.stderr)
+
+
+if __name__ == "__main__":
+    arguments = parse_arguments()
+    try:
+        sphinx_config = CONFIG_DICT[arguments.environment][arguments.version]["config"]
+        sphinx_pid = CONFIG_DICT[arguments.environment][arguments.version]["pid"]
+    except KeyError:
+        logger.error('Configuration for version {} and environment {} not found!'.format(arguments.version, arguments.environment))
+        sys.exit(1)
+
+    if not os.path.isfile(sphinx_config):
+        logger.error('Sphinx configuration file {} not found!'.format(sphinx_config))
+        sys.exit(1)
+
+    pid = read_pid_file(filename=sphinx_pid)
+    if not pid or not pid_exists(pid=pid):
+        logger.info('Restarting searchd daemon for {}'.format(arguments.environment))
+        start_sphinx(config=sphinx_config, logfile=arguments.logfile)