源码分析_Shinken-2.4.0001.启动脚本/etc/init.d/shinken源码分析?
Posted
tags:
篇首语：本文由小常识网(cha138.com)小编为大家整理，主要介绍了源码分析_Shinken-2.4.0001.启动脚本/etc/init.d/shinken源码分析?相关的知识，希望对你有一定的参考价值。
简单介绍:
说明: Shinken是一个网络监控平台,可以通过一系列直观的方式监控网络内的各种健康状况.Shinken脱胎于Nagios,其实Shinken这个项目本身就是一帮Nagios项目的人无法忍受Nagios,自己跳出来重新用纯Python重构了一下,甚至完全兼容Nagios的配置文件.
相关地址:
官网地址: http://www.shinken-monitoring.org/
官网文档: http://shinken.readthedocs.io/en/latest/
论坛地址: http://shinken.readthedocs.io/en/latest/
架构一览:
组件介绍:
shinken-arbiter
说明: shinken-arbiter节点读取配置,然后将配置切分后分发到多个shinken-scheduler节点
shinken-broker
说明: shinken-broker节点负责导出和管理shinken-scheduler节点中的数据
shinken-scheduler
说明: shinken-scheduler节点负责管理shinken-poller节点和shinken-reactionner节点任务调度
shinken-poller
说明: shinken-poller节点通过各类插件执行shinken-scheduler节点的任务,获取各种监控指标
shinken-reactionner
说明: shinken-reactionner节点负责一旦满足要求则触发预警通知事件
shinken-receiver
说明: shinken-receiver可选节点,特殊环境下的数据汇总统一转发
源码分析:
#!/bin/sh # Copyright (C) 2009-2014: # Gabes Jean, [email protected] # Gerhard Lausser, [email protected] # Gregory Starck, [email protected] # Hartmut Goebel, [email protected] # # This file is part of Shinken. # # Shinken is free software: you can redistribute it and/or modify # it under the terms of the GNU Affero General Public License as published by # the Free Software Foundation, either version 3 of the License, or # (at your option) any later version. # # Shinken is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU Affero General Public License for more details. # # You should have received a copy of the GNU Affero General Public License # along with Shinken. If not, see <http://www.gnu.org/licenses/>. ### BEGIN INIT INFO # Provides: shinken # Required-Start: $network $remote_fs # Required-Stop: $network $remote_fs # Default-Start: 2 3 4 5 # Default-Stop: 0 1 6 # Short-Description: Shinken monitoring daemon # Description: Shinken is a monitoring tool composed of many separated modules: # - arbiter : the main one : control everything else. # - scheduler : receives checks/actions from arbiter. Schedules & forwards them to pollers. # - poller : receives the checks from a scheduler. Launch them and returns results # - broker : manage results by looking at scheduler. Like export to flat file or db. # - reactionner : manage the failed checks by looking at scheduler. # - receiver : manage all passive data ### END INIT INFO ### Chkconfig Header # Shinken Starts Shinken daemons # # chkconfig: 345 99 01 # description: Start Shinken daemons # Reference: # http://refspecs.linuxfoundation.org/LSB_4.1.0/LSB-Core-generic/LSB-Core-generic/iniscrptact.html NAME="shinken" # 分析: 所有组件 AVAIL_MODULES="scheduler poller reactionner broker receiver arbiter" ## SHINKEN_MODULE_FILE is set by shinken-* if it‘s one of these that‘s calling us. if [ -z "$SHINKEN_MODULE_FILE" ]; then SCRIPTNAME=$0 _usage_mods_="[ <$AVAIL_MODULES> ]" else SCRIPTNAME=$SHINKEN_MODULE_FILE fi # 分析: /etc/init.d/shinken restart启动时curpath为/etc/init.d/ curpath=$(cd $(dirname "$0") && pwd) #echo curpath is $curpath filename is $(basename "$0") # 分析: # bin - /etc 是否有问题 ??? # var - /var # etc - /etc ## Default paths: test "$BIN" || BIN=$(cd $curpath/.. && pwd) test "$VAR" || VAR=$(cd $curpath/../../var && pwd) test "$ETC" || ETC=$(cd $curpath/../../etc && pwd) # 分析: 设置相关环境变量 export PATH="${PATH:+$PATH:}/usr/sbin:/bin:/sbin" export LANG=en_US.UTF8 export LC_ALL=en_US.UTF8 export PYTHONIOENCODING=utf8 export PYTHONUNBUFFERED="0" export TZ=:/etc/localtime # also unset http proxy, because pycurl is using it and this is bad, very bad :) unset http_proxy unset https_proxy # 分析: 判断当前系统PY版本 # We try to find the LAST possible Python VERSION pythonver() { versions="2.4 2.5 2.6 2.7" LASTFOUND="" # Is there any python here? for v in $versions do which python$v > /dev/null 2>&1 if [ $? -eq 0 ] then LASTFOUND="python$v" fi done if [ -z "$LASTFOUND" ] then # Finaly try to find a default python which python > /dev/null 2>&1 if [ $? -ne 0 ] then echo "No python interpreter found!" exit 2 else echo "python found" LASTFOUND=$(which python) fi fi # 分析: PYTHON为当前PY版本 PYTHON=$LASTFOUND } # Ok, go search this Python version pythonver # Uncomment the line below if you got the **lib** shinken installed # on a non standard place (not in /usr/lib/python*) #export PYTHONPATH="${PATH:+$PATH:}/opt/shinken" # Or uncommentif you want to force the Python version #export PYTHON=python2.7 # default DEBUG=false CMD="" SUBMODULES="" # 分析: 如果$SHINKEN_DEFAULT_FILE未配置则设置为/etc/default/shinken文件,就是所有组件的默认配置 ## This permits to overhidde the default "default shinken cfg file": [ -z "$SHINKEN_DEFAULT_FILE" ] && SHINKEN_DEFAULT_FILE="/etc/default/$NAME" ## so you can now do: ## bash -c "SHINKEN_DEFAULT_FILE=$your_own_default_file $init_path/shinken $action $args" ## to easily use your own config #echo "Using $SHINKEN_DEFAULT_FILE .." usage() { cat << END Usage: $SCRIPTNAME [ -d ] {start|stop|status|restart|reload|force-reload|check} $_usage_mods_ -d start requested module(s) in debug mode, only useful with start|restart END } if [ "$1" = "-d" ]; then DEBUG="1" shift fi if [ $# -eq 0 ]; then usage >&2 exit 2 fi # 分析: CMD就是start stop status restart reload force-reload check这些命令 CMD=$1 shift # 分析: SUBMODULES就是scheduler poller reactionner broker receiver arbiter这些组件 SUBMODULES=$* # 分析: 加载/etc/default/shinke中的所有组件的默认配置 # Reads configuration variable file if it is present [ -r "$SHINKEN_DEFAULT_FILE" ] && . "$SHINKEN_DEFAULT_FILE" # 分析: 通过参数形式来设置CMD是作用于一个组件还是多个组件 if [ -z "$SUBMODULES" ]; then SUBMODULES=$AVAIL_MODULES else # check given modules for mod1 in $SUBMODULES; do found=0 for mod2 in $AVAIL_MODULES; do [ $mod1 = $mod2 ] && found=1; done [ $found = 0 ] && { usage >&2 ; exit 2 ; } done fi # 分析: 如果正确加载/etc/default/shinke上面的$ETC被重置,设置通用对象配置文件SHINKENCFG # Now look if some required variables are pre defined: if ! test "$SHINKENCFG" then SHINKENCFG="$ETC/shinken.cfg" fi # If var or run dir is missing, create them and chown them #[ ! -d $VAR ] && mkdir -p $VAR && chown $SHINKENUSER:$SHINKENGROUP $VAR #[ ! -d $RUN ] && mkdir -p $RUN && chown $SHINKENUSER:$SHINKENGROUP $RUN # Now place us in our var directory so even our arbiter will be # happy for opening its pid and cmd files # 分析: 进入运行目录/var下 cd $VAR #echo BIN=$BIN #echo VAR=$VAR #echo ETC=$ETC #set -xv echo_success() { log_end_msg 0 $* } echo_failure() { log_end_msg 1 $* } #log_end_msg # 分析: 加载默认一些配置以及SHELL函数库 # Load the VERBOSE setting and other rcS variables [ -f /etc/default/rcS ] && . /etc/default/rcS # Source function library. [ -f /etc/rc.d/init.d/functions ] && . /etc/rc.d/init.d/functions [ -f /lib/lsb/init-functions ] && . /lib/lsb/init-functions ################################################ # # returns the pid for a submodule # getpidfile() { mod="$1" modPIDVAR=$(echo $mod | tr ‘a-z‘ ‘A-Z‘)"PID" pidfile=$(echo $(eval echo \${$modPIDVAR})) if test "$pidfile" then echo "$pidfile" else echo "$RUN/${mod}d.pid" fi } getmodpid() { mod=$1 pidfile=$(getpidfile "$mod") if [ -s $pidfile ]; then cat $pidfile fi } # 分析: 生成调试模式下文件/tmp/bad_start_for_${mod} getdebugfile() { mod="$1" modDEBUG=$(echo $mod | tr ‘a-z‘ ‘A-Z‘)"DEBUGFILE" debugfile=$(echo $(eval echo \${$modDEBUG})) if test "$debugfile" then echo "$debugfile" else echo "${VAR}/${mod}-debug.log" fi } # # Display status # do_status() { mod=$1 pidfile=$(getpidfile "$mod") [ -e "$pidfile" ] || { echo "$mod NOT RUNNING (pidfile ($pidfile) not exist)" return 3 } [ -r "$pidfile" ] || { echo "$mod NOT RUNNING (pidfile ($pidfile) unreadable)" return 3 } pid=$(cat "$pidfile") if [ -z "$pid" ]; then echo "$mod NOT RUNNING (pid file empty)" return 4 fi ps -p "$pid" >/dev/null 2>&1 rc=$? if [ $rc != 0 ]; then log_failure_msg "$mod NOT RUNNING (process $pid doesn‘t exist?)" return 1 fi echo "$mod RUNNING (pid $pid)" return 0 } # 分析: 启动/bin/shinken-${mod}单个组件 # # starts our modules # do_start() { mod=$1 modfilepath="$BIN/shinken-${mod}" [ -e "$modfilepath" ] || { log_failure_msg "FAILED: did not find $mod file ($modfilepath) ; are you sure shinken-$mod is installed?" return 5 } # 分析: 启动调试后生成调试文件/tmp/bad_start_for_${mod} [ "$DEBUG" = 1 ] && DEBUGCMD="--debug "$(getdebugfile "$mod") # Arbiter shinken.cfg, and the other OTHERd.ini if [ "$mod" != "arbiter" ]; then modINI=$(echo "$"${mod}CFG | tr ‘[:lower:]‘ ‘[:upper:]‘) modinifile=$(eval echo ${modINI}) output=$($PYTHON "$modfilepath" -d -c "${modinifile}" $DEBUGCMD 2>&1) rc=$? else # 分析: $SHINKENSPECIFICCFG的存在主要是为了类Centreon监控的可能需要独立配置 if ! test "$SHINKENSPECIFICCFG" then output=$($PYTHON "$modfilepath" -d -c "$SHINKENCFG" $DEBUGCMD 2>&1) else output=$($PYTHON "$modfilepath" -d -c "$SHINKENCFG" -c "$SHINKENSPECIFICCFG" $DEBUGCMD 2>&1) fi rc=$? fi # debug: #resfile="/tmp/bad_start_for_$mod" #echo "$output" > "$resfile" || true if [ $rc != 0 ]; then resfile="/tmp/bad_start_for_$mod" echo "$output" > "$resfile" || true output=$(echo "$output" | tail -1) echo "FAILED: $output (full output is in $resfile)" return 1 fi echo "OK" return 0 } # 分析: 关闭/bin/shinken-${mod}单个组件 # # stops modules # do_stop() { mod=$1 pid=$(getmodpid "$mod") statusoutput=$(do_status "$mod") [ $? -ne 0 ] && { echo "$statusoutput" return 0 } if [ ! -z "$pid" ]; then kill "$pid" sleep 1 ## TODO: instead of ‘sleep 1‘: wait up to when pid file is removed (with timeout)? for i in 1 2 3 do # TODO: use a better way to get the children pids.. allpids="$(ps -aef | grep "$pid" | grep "shinken-$mod" | awk ‘{print $2}‘)" if [ -z "$allpids" ]; then echo "OK" return 0 fi sleep 1 done echo "there are still remaining processes to $mod running.. ; trying to kill them (SIGTERM).." allpids="$(ps -aef | grep "$pid" | grep "shinken-$mod" | awk ‘{print $2}‘)" for cpid in $(ps -aef | grep "$pid" | grep "shinken-$mod" | awk ‘{print $2}‘); do kill $cpid > /dev/null 2>&1 done for i in 1 2 3 do # TODO: eventually use a better way to get the children pids.. allpids="$(ps -aef | grep "$pid" | grep "shinken-$mod" | awk ‘{print $2}‘)" if [ -z "$allpids" ]; then echo "OK" return 0 fi sleep 1 done echo "there are still remaining processes to $mod running.. ; trying to kill -9 them.." allpids="$(ps -aef | grep "$pid" | grep "shinken-$mod" | awk ‘{print $2}‘)" for cpid in $(ps -aef | grep "$pid" | grep "shinken-$mod" | awk ‘{print $2}‘); do kill -9 $cpid > /dev/null 2>&1 done sleep 1 allpids="$(ps -aef | grep "$pid" | grep "shinken-$mod" | awk ‘{print $2}‘)" if [ ! -z "$allpids" ]; then echo "FAILED: one or more process for $mod are still running after kill -9!" echo "Remaining processes are (pids="$allpids"):" ps -lf $(for p in $allpids ; do echo -n "-p$p " ; done) echo "You should check this." return 1 fi echo "OK" else echo "NOT RUNNING" fi return 0 } # 分析: 调用python shinken-arbiter -v -c /etc/shinken/shinken.cfg检测配置是否有问题 # # does the config check # do_check() { [ "$DEBUG" = 1 ] && DEBUGCMD="--debug $VAR/${mod}-debug.log" if ! test "$SHINKENSPECIFICCFG" then $PYTHON "$BIN/shinken-arbiter" -v -c "$SHINKENCFG" $DEBUGCMD 2>&1 else $PYTHON "$BIN/shinken-arbiter" -v -c "$SHINKENCFG" -c "$SHINKENSPECIFICCFG" $DEBUGCMD 2>&1 fi return $? } ############################ # 分析: 默认在/var/lib/shinken下启动组件但是arbiter比较特殊没有声明所以就到/var/下面去启动 do_start_() { echo "Starting $1: " status=$(do_status "$1") rc=$? if [ $rc -eq 0 ]; then log_warning_msg "Already running" return fi if test "$1" = "arbiter" then # arbiter is special: # it doesn‘t actually declare a "workdir" properties in its config # so we have explicitely to cd to the "VAR" directory. # so that the default pidfile ( == nagios lock_file) which is now "arbiterd.pid" # will be created at the correct place. cd "$VAR" # TODO: check if other possibility wouldn‘t be better: # declare a "workdir" properties for the arbiter module definition.. in shinken-specific.cfg. # but if the lock_file path is absolute then this ‘cd‘ isn‘t required. fi # 分析: 调用do_start启动对应组件 startoutput=$(do_start "$1") rc=$? if [ $rc -eq 0 ]; then echo_success else echo "$startoutput" echo_failure fi return $rc } # 分析: 关闭所有组件 do_stop_() { echo "Stopping $1" statusoutput=$(do_status "$1") rc=$? if [ $rc -ne 0 ]; then failuremsg="Couldn‘t get status of $1: $statusoutput" else stopoutput=$(do_stop "$1" 2>&1) rc=$? [ $rc -ne 0 ] && failuremsg="Couldn‘t stop $1: $stopoutput" fi if [ $rc -ne 0 ]; then log_failure_msg "$failuremsg" echo_failure else echo_success fi return $rc } # 分析: 重启所有组件 do_restart_() { mod="$1" echo "Restarting $mod" if [ "$mod" = "arbiter" ]; then do_check_ "$mod" checkrc=$? if [ $checkrc -ne 0 ]; then return 1 fi fi stopoutput=$(do_stop "$mod") startoutput=$(do_start "$mod") rc=$? if [ $rc -eq 0 ]; then echo_success else log_failure_msg "$startoutput" echo_failure fi return $rc } # 分析: 强制重新加载组件 do_force_reload_() { do_restart_ $1 } # 分析: 重新加载所有组件 do_reload_() { mod="$1" echo "Reloading $mod" if [ "$mod" = "arbiter" ]; then do_check_ "$mod" checkrc=$? if [ $checkrc -ne 0 ]; then return 1 fi fi stopoutput=$(do_stop "$mod") startoutput=$(do_start "$mod") rc=$? if [ $rc -eq 0 ]; then echo_success else log_failure_msg "$startoutput" echo_failure fi return $rc } # 分析: 获取所有组件状态 do_status_() { mod=$1 echo "Checking status of $mod" do_status "$1" rc=$? if [ $rc -eq 0 ]; then echo_success else echo_failure fi } # 分析: 检查/etc/shinken/shinken.cfg以及其包含配置文件配置,并将检测结果写在/tmp/shinken_checkconfig_result文件 do_check_() { echo "Doing config check" output=$(do_check "$1" 2>&1) rc=$? check_res_file=$(mktemp /tmp/shinken_checkconfig_resultXXXXXXXX) echo "$output" > "$check_res_file" mv $check_res_file /tmp/shinken_checkconfig_result check_res_file="/tmp/shinken_checkconfig_result" if [ $rc -eq 0 ]; then echo_success else output=$(echo "$output" | tail -1) log_warning_msg "full result is in ${check_res_file}" log_failure_msg "ConfigCheck failed: $output" echo_failure fi return $rc } do_checkconfig_() { do_check_ "$1" ; } ############################ # 分析: 遍历组件每个组件调用对应处理函数do_${action}_ "${mod}" do_cmd_on() { # 分析: action就是start stop status restart reload force-reload check action=$1 # 分析: mods其实就是scheduler poller reactionner broker receiver arbiter单个组件或是组件组合 mods=$2 local return_value return_value=0 # 分析: 遍历组件组合 for mod in $mods do # If at least one action fails, the return value is 1. # 调用do_${action}_ "${mode}"来处理 do_${action}_ "$mod" || return_value=1 done return $return_value } # 分析: 代码入口,根据$CMD来调用do_cmd_on函数处理 ############################ ## Main: case "$CMD" in start|stop|restart|status|force-reload) do_cmd_on "$CMD" "$SUBMODULES" ;; force-reload) do_cmd_on "force_reload" "$SUBMODULES" ;; check|checkconfig|reload) do_cmd_on "$CMD" "arbiter" ;; *) usage >&2 exit 2 ;; esac
最终目标::
说明: 由于目前市面上监控系统Zabbix/Nagios/Ganglia/OneAPM/Cacti/监控宝/Open-falcon/OWL/Zenoss/Hyperic HQ/OpenNMS/360网站服务监控/阿里云监控/百度云观测/小蜜蜂网站监测,几乎很少用PY作为后端监控,之前自己也尝试编写过一套基于REDIS的全自动插件式监控系统xmzoomeye,但是自从看到Shinken纯PY实现的完整的监控系统解决方案,我觉得这正是我需要的,接下来我会和大家一起来分析Shinken整个监控框架的源码,最后我们自己实现一套监控框架~
本文出自 “@湖北@白头发” 博客，请务必保留此出处http://xmdevops.blog.51cto.com/11144840/1867125
以上是关于源码分析_Shinken-2.4.0001.启动脚本/etc/init.d/shinken源码分析?的主要内容，如果未能解决你的问题，请参考以下文章
从头用脚分析FFmpeg源码 - avformat_write_header