[Monit] Monitoring the running status of containers.

files/build_templates/sonic_debian_extension.j2

@@ -309,6 +309,9 @@ sudo cp $IMAGE_CONFIGS/monit/conf.d/* $FILESYSTEM_ROOT/etc/monit/conf.d/
 sudo chmod 600 $FILESYSTEM_ROOT/etc/monit/conf.d/*
 sudo cp $IMAGE_CONFIGS/monit/process_checker $FILESYSTEM_ROOT/usr/bin/
 sudo chmod 755 $FILESYSTEM_ROOT/usr/bin/process_checker
+sudo cp $IMAGE_CONFIGS/monit/container_checker $FILESYSTEM_ROOT/usr/bin/
+sudo chmod 755 $FILESYSTEM_ROOT/usr/bin/container_checker
+
 
 # Install custom-built openssh sshd
 sudo dpkg --root=$FILESYSTEM_ROOT -i $debs_path/openssh-server_*.deb

files/image_config/monit/conf.d/sonic-host

@@ -31,3 +31,5 @@ check program routeCheck with path "/usr/local/bin/route_check.py"
     every 5 cycles
     if status != 0 for 3 cycle then alert repeat every 1 cycles
 
+check program container_checker with path "/usr/bin/container_checker"
+    if status != 0 for 5 times within 5 cycles then alert repeat every 1 cycles

files/image_config/monit/container_checker

@@ -0,0 +1,116 @@
+#!/usr/bin/env python3
+
+"""
+container_checker
+
+This script is intended to be run by Monit. It will write an alerting message into
+syslog if it found containers which were expected to run but were not running. At
+the same time, if some containers were unexpected to run, it also writes an alerting
+syslog message. Note that if print(...) statement in this script was executed, the
+string in it will be appended to Monit syslog messages.
+
+The following is an example in Monit configuration file to show how Monit will run
+this script:
+
+check program container_checker with path "/usr/bin/container_checker"
+    if status != 0 for 5 times within 5 cycles then alert repeat every 1 cycles
+"""
+
+import subprocess
+import sys
+
+import swsssdk
+from sonic_py_common import multi_asic
+
+
+def get_command_result(command):
+    """
+    @summary: This function will execute the command and return the resulting output.
+    @return: A string which contains the output of command.
+    """
+    command_stdout = ""
+
+    try:
+        proc_instance = subprocess.Popen(command, stdout=subprocess.PIPE, stderr=subprocess.PIPE,
+                                         shell=True, universal_newlines=True)
+        command_stdout, command_stderr = proc_instance.communicate()
+        if proc_instance.returncode != 0:
+            print("Failed to execute the command '{}'. Return code: '{}'".format(
+                command, proc_instance.returncode))
+            sys.exit(1)
+    except (OSError, ValueError) as err:
+        print("Failed to execute the command '{}'. Error: '{}'".format(command, err))
+        sys.exit(2)
+
+    return command_stdout.rstrip().split("\n")
+
+
+def get_expected_running_containers():
+    """
+    @summary: This function will get the expected running containers by following the rule:
+              The 'state' field of container in 'FEATURE' table should not be 'disabled'. Then
+              if the device has Multi-ASIC, this function will get container list by determining the
+              value of field 'has_global_scope', the number of ASICs and the value of field
+              'has_per_asic_scope'. If the device has single ASIC, the container name was put into
+              the list.
+    @return:  A set which contains the expected running containers.
+    """
+    config_db = swsssdk.ConfigDBConnector()
+    config_db.connect()
+    feature_table = config_db.get_table("FEATURE")
+
+    expected_running_containers = set()
+
+    for container_name in feature_table.keys():
+        if feature_table[container_name]["state"] != "disabled":
+            if multi_asic.is_multi_asic():
+                if feature_table[container_name]["has_global_scope"] == "True":
+                    expected_running_containers.add(container_name)
+                if feature_table[container_name]["has_per_asic_scope"] == "True":
+                    num_asics = multi_asic.get_num_asics()
+                    for asic_id in range(num_asics):
+                        expected_running_containers.add(container_name + str(asic_id))
+            else:
+                expected_running_containers.add(container_name)
+
+    return expected_running_containers
+
+
+def get_current_running_containers():
+    """
+    @summary: This function will get the current running container list by analyzing the
+              output of command `docker ps`.
+    @return:  A set which contains the current running contianers.
+    """
+    running_containers = set()
+
+    command = "docker ps"
+    command_stdout = get_command_result(command)
+    for line in command_stdout[1:]:
+        running_containers.add(line.split()[-1].strip())
+
+    return running_containers
+
+
+def main():
+    """
+    @summary: This function will compare the difference between the current running containers
+              and the containers which were expected to run. If containers which were exepcted
+              to run were not running, then an alerting message will be written into syslog.
+    """
+    expected_running_containers = get_expected_running_containers()
+    current_running_containers = get_current_running_containers()
+
+    not_running_containers = expected_running_containers.difference(current_running_containers)
+    if not_running_containers:
+        print("Containers not running: " + ", ".join(not_running_containers))
+        sys.exit(3)
+
+    unexpected_running_containers = current_running_containers.difference(expected_running_containers)
+    if unexpected_running_containers:
+        print("Unexpected running containers: " + ", ".join(unexpected_running_containers))
+        sys.exit(4)
+
+
+if __name__ == "__main__":
+    main()