272 lines
8.5 KiB
Bash
272 lines
8.5 KiB
Bash
#!/bin/bash
|
||
|
||
# 增强版进程资源监控脚本（支持多同名进程）
|
||
# 用法: ./proc_monitor_enhanced.sh <进程名> <监控时长(秒)>
|
||
|
||
# 检查参数是否正确
|
||
if [ $# -ne 2 ]; then
|
||
echo "错误: 参数不正确!"
|
||
echo "用法: $0 <进程名> <监控时长(秒)>"
|
||
exit 1
|
||
fi
|
||
|
||
PROCESS_NAME=$1
|
||
DURATION=$2
|
||
BASE_LOG_PREFIX="process_monitor_${PROCESS_NAME}"
|
||
TIMESTAMP=$(date +%Y%m%d_%H%M%S)
|
||
SUMMARY_LOG="${BASE_LOG_PREFIX}_summary_${TIMESTAMP}.log"
|
||
|
||
# 函数：获取所有匹配进程的PID
|
||
get_all_pids() {
|
||
local name=$1
|
||
# 方法1: 使用pgrep（推荐，更简洁）
|
||
pgrep -f "$name" 2>/dev/null
|
||
|
||
# 方法2: 使用ps和awk（备选，更精确）
|
||
# ps -ef | awk -v proc="$name" '$0 ~ proc && $0 !~ /grep|proc_monitor/ {print $2}'
|
||
}
|
||
|
||
echo "正在查找进程: $PROCESS_NAME"
|
||
PIDS=$(get_all_pids "$PROCESS_NAME")
|
||
|
||
if [ -z "$PIDS" ]; then
|
||
echo "错误: 未找到进程 '$PROCESS_NAME'!"
|
||
exit 1
|
||
fi
|
||
|
||
# 将PID转换为数组
|
||
PID_ARRAY=($PIDS)
|
||
PID_COUNT=${#PID_ARRAY[@]}
|
||
|
||
echo "找到 $PID_COUNT 个同名进程，PID: $PIDS"
|
||
echo "监控时长: $DURATION 秒"
|
||
echo "汇总日志: $SUMMARY_LOG"
|
||
echo "开始时间: $(date)"
|
||
|
||
# 为每个PID创建独立的日志文件
|
||
declare -A LOG_FILES
|
||
for pid in ${PID_ARRAY[@]}; do
|
||
LOG_FILES[$pid]="${BASE_LOG_PREFIX}_pid${pid}_${TIMESTAMP}.log"
|
||
|
||
# 初始化独立日志文件
|
||
echo "进程监控日志 - $PROCESS_NAME (PID: $pid)" > ${LOG_FILES[$pid]}
|
||
echo "开始时间: $(date)" >> ${LOG_FILES[$pid]}
|
||
echo "监控间隔: 2秒" >> ${LOG_FILES[$pid]}
|
||
echo "==============================================" >> ${LOG_FILES[$pid]}
|
||
printf "%-20s %-8s %-8s %-12s %-12s\n" "时间戳" "CPU(%)" "内存(%)" "虚拟内存" "物理内存" >> ${LOG_FILES[$pid]}
|
||
done
|
||
|
||
# 初始化统计数组
|
||
declare -A CPU_SUM
|
||
declare -A MEM_SUM
|
||
declare -A CPU_MAX
|
||
declare -A MEM_MAX
|
||
declare -A COUNT
|
||
|
||
# 初始化每个PID的统计变量
|
||
for pid in ${PID_ARRAY[@]}; do
|
||
CPU_SUM[$pid]=0
|
||
MEM_SUM[$pid]=0
|
||
CPU_MAX[$pid]=0
|
||
MEM_MAX[$pid]=0
|
||
COUNT[$pid]=0
|
||
done
|
||
|
||
# 全局统计变量
|
||
GLOBAL_CPU_SUM=0
|
||
GLOBAL_MEM_SUM=0
|
||
GLOBAL_CPU_MAX=0
|
||
GLOBAL_MEM_MAX=0
|
||
GLOBAL_COUNT=0
|
||
|
||
# 函数：监控单个进程
|
||
monitor_single_process() {
|
||
local pid=$1
|
||
local log_file=$2
|
||
|
||
if [ ! -d /proc/$pid ]; then
|
||
echo "进程 $pid 已退出"
|
||
return 1
|
||
fi
|
||
|
||
# 使用top命令获取进程资源使用情况
|
||
local PROCESS_INFO=$(top -bn1 -p $pid 2>/dev/null | grep -w $pid)
|
||
|
||
if [ -z "$PROCESS_INFO" ]; then
|
||
return 1
|
||
fi
|
||
|
||
# 解析CPU和内存使用率
|
||
local CPU_USAGE=$(echo $PROCESS_INFO | awk '{print $9}')
|
||
local MEM_USAGE=$(echo $PROCESS_INFO | awk '{print $10}')
|
||
|
||
# 从/proc获取详细内存信息(单位:MB)
|
||
local VMRSS=0
|
||
local VMSIZE=0
|
||
if [ -f /proc/$pid/status ]; then
|
||
VMRSS=$(grep VmRSS /proc/$pid/status | awk '{print $2}')
|
||
VMSIZE=$(grep VmSize /proc/$pid/status | awk '{print $2}')
|
||
|
||
# 转换为MB
|
||
VMRSS=$((VMRSS / 1024))
|
||
VMSIZE=$((VMSIZE / 1024))
|
||
fi
|
||
|
||
local TIMESTAMP=$(date +"%Y-%m-%d %H:%M:%S")
|
||
|
||
# 记录到独立日志文件
|
||
printf "%-20s %-8.1f %-8.1f %-12s %-12s\n" "$TIMESTAMP" $CPU_USAGE $MEM_USAGE "${VMSIZE}MB" "${VMRSS}MB" >> $log_file
|
||
|
||
# 更新单个进程的统计信息
|
||
CPU_SUM[$pid]=$(echo "scale=2; ${CPU_SUM[$pid]} + $CPU_USAGE" | bc)
|
||
MEM_SUM[$pid]=$(echo "scale=2; ${MEM_SUM[$pid]} + $MEM_USAGE" | bc)
|
||
|
||
if [ $(echo "$CPU_USAGE > ${CPU_MAX[$pid]}" | bc) -eq 1 ]; then
|
||
CPU_MAX[$pid]=$CPU_USAGE
|
||
fi
|
||
|
||
if [ $(echo "$MEM_USAGE > ${MEM_MAX[$pid]}" | bc) -eq 1 ]; then
|
||
MEM_MAX[$pid]=$MEM_USAGE
|
||
fi
|
||
|
||
COUNT[$pid]=$((COUNT[$pid] + 1))
|
||
|
||
# 更新全局统计
|
||
GLOBAL_CPU_SUM=$(echo "scale=2; $GLOBAL_CPU_SUM + $CPU_USAGE" | bc)
|
||
GLOBAL_MEM_SUM=$(echo "scale=2; $GLOBAL_MEM_SUM + $MEM_USAGE" | bc)
|
||
|
||
if [ $(echo "$CPU_USAGE > $GLOBAL_CPU_MAX" | bc) -eq 1 ]; then
|
||
GLOBAL_CPU_MAX=$CPU_USAGE
|
||
fi
|
||
|
||
if [ $(echo "$MEM_USAGE > $GLOBAL_MEM_MAX" | bc) -eq 1 ]; then
|
||
GLOBAL_MEM_MAX=$MEM_USAGE
|
||
fi
|
||
|
||
GLOBAL_COUNT=$((GLOBAL_COUNT + 1))
|
||
|
||
echo "$TIMESTAMP - PID $pid: CPU=${CPU_USAGE}%, 内存=${MEM_USAGE}%, 物理内存=${VMRSS}MB"
|
||
return 0
|
||
}
|
||
|
||
# 主监控循环
|
||
START_TIME=$(date +%s)
|
||
END_TIME=$((START_TIME + DURATION))
|
||
|
||
echo "开始监控进程..."
|
||
echo "进程列表: ${PID_ARRAY[*]}"
|
||
|
||
while [ $(date +%s) -lt $END_TIME ]; do
|
||
CURRENT_TIME=$(date +%s)
|
||
ELAPSED=$((CURRENT_TIME - START_TIME))
|
||
REMAINING=$((DURATION - ELAPSED))
|
||
|
||
if [ $REMAINING -le 0 ]; then
|
||
break
|
||
fi
|
||
|
||
echo "======= 监控进度: $ELAPSED/$DURATION 秒, 剩余: $REMAINING 秒 ======="
|
||
|
||
# 检查是否有新的同名进程启动
|
||
CURRENT_PIDS=$(get_all_pids "$PROCESS_NAME")
|
||
NEW_PIDS=$(echo "$CURRENT_PIDS" | grep -v -w "$(echo ${PID_ARRAY[@]} | sed 's/ /\\|/g')" || true)
|
||
|
||
if [ -n "$NEW_PIDS" ]; then
|
||
echo "发现新进程: $NEW_PIDS，已加入监控"
|
||
for new_pid in $NEW_PIDS; do
|
||
PID_ARRAY+=($new_pid)
|
||
LOG_FILES[$new_pid]="${BASE_LOG_PREFIX}_pid${new_pid}_${TIMESTAMP}.log"
|
||
|
||
# 初始化新进程的日志和统计
|
||
echo "进程监控日志 - $PROCESS_NAME (PID: $new_pid)" > ${LOG_FILES[$new_pid]}
|
||
echo "开始时间: $(date)" >> ${LOG_FILES[$new_pid]}
|
||
printf "%-20s %-8s %-8s %-12s %-12s\n" "时间戳" "CPU(%)" "内存(%)" "虚拟内存" "物理内存" >> ${LOG_FILES[$new_pid]}
|
||
|
||
CPU_SUM[$new_pid]=0
|
||
MEM_SUM[$new_pid]=0
|
||
CPU_MAX[$new_pid]=0
|
||
MEM_MAX[$new_pid]=0
|
||
COUNT[$new_pid]=0
|
||
done
|
||
fi
|
||
|
||
# 监控每个进程
|
||
ACTIVE_PIDS=()
|
||
for pid in ${PID_ARRAY[@]}; do
|
||
if monitor_single_process $pid ${LOG_FILES[$pid]}; then
|
||
ACTIVE_PIDS+=($pid)
|
||
fi
|
||
done
|
||
|
||
# 更新活跃PID数组
|
||
PID_ARRAY=(${ACTIVE_PIDS[@]})
|
||
PID_COUNT=${#PID_ARRAY[@]}
|
||
|
||
if [ $PID_COUNT -eq 0 ]; then
|
||
echo "所有监控的进程都已退出!"
|
||
break
|
||
fi
|
||
|
||
sleep 2
|
||
done
|
||
|
||
# 生成统计报告
|
||
echo "==============================================" > $SUMMARY_LOG
|
||
echo "多进程监控统计报告" >> $SUMMARY_LOG
|
||
echo "==============================================" >> $SUMMARY_LOG
|
||
echo "进程名称: $PROCESS_NAME" >> $SUMMARY_LOG
|
||
echo "监控时长: $DURATION 秒" >> $SUMMARY_LOG
|
||
echo "开始时间: $(date -d @$START_TIME)" >> $SUMMARY_LOG
|
||
echo "结束时间: $(date)" >> $SUMMARY_LOG
|
||
echo "" >> $SUMMARY_LOG
|
||
|
||
# 单个进程统计
|
||
for pid in ${!LOG_FILES[@]}; do
|
||
if [ ${COUNT[$pid]} -gt 0 ]; then
|
||
CPU_AVG=$(echo "scale=2; ${CPU_SUM[$pid]} / ${COUNT[$pid]}" | bc)
|
||
MEM_AVG=$(echo "scale=2; ${MEM_SUM[$pid]} / ${COUNT[$pid]}" | bc)
|
||
|
||
echo "进程 PID: $pid" >> ${LOG_FILES[$pid]}
|
||
echo "数据点数: ${COUNT[$pid]} 个" >> ${LOG_FILES[$pid]}
|
||
echo "CPU平均使用率: $CPU_AVG%" >> ${LOG_FILES[$pid]}
|
||
echo "CPU最大使用率: ${CPU_MAX[$pid]}%" >> ${LOG_FILES[$pid]}
|
||
echo "内存平均使用率: $MEM_AVG%" >> ${LOG_FILES[$pid]}
|
||
echo "内存最大使用率: ${MEM_MAX[$pid]}%" >> ${LOG_FILES[$pid]}
|
||
|
||
echo "PID $pid - 平均CPU: $CPU_AVG%, 最大CPU: ${CPU_MAX[$pid]}%, 平均内存: $MEM_AVG%, 最大内存: ${MEM_MAX[$pid]}%" >> $SUMMARY_LOG
|
||
fi
|
||
done
|
||
|
||
# 全局统计
|
||
if [ $GLOBAL_COUNT -gt 0 ]; then
|
||
GLOBAL_CPU_AVG=$(echo "scale=2; $GLOBAL_CPU_SUM / $GLOBAL_COUNT" | bc)
|
||
GLOBAL_MEM_AVG=$(echo "scale=2; $GLOBAL_MEM_SUM / $GLOBAL_COUNT" | bc)
|
||
else
|
||
GLOBAL_CPU_AVG=0
|
||
GLOBAL_MEM_AVG=0
|
||
fi
|
||
|
||
echo "" >> $SUMMARY_LOG
|
||
echo "===== 全局统计（所有进程聚合） =====" >> $SUMMARY_LOG
|
||
echo "总监控进程数: ${#LOG_FILES[@]}" >> $SUMMARY_LOG
|
||
echo "总数据点数: $GLOBAL_COUNT" >> $SUMMARY_LOG
|
||
echo "全局CPU平均使用率: $GLOBAL_CPU_AVG%" >> $SUMMARY_LOG
|
||
echo "全局CPU最大使用率: $GLOBAL_CPU_MAX%" >> $SUMMARY_LOG
|
||
echo "全局内存平均使用率: $GLOBAL_MEM_AVG%" >> $SUMMARY_LOG
|
||
echo "全局内存最大使用率: $GLOBAL_MEM_MAX%" >> $SUMMARY_LOG
|
||
|
||
# 屏幕输出总结
|
||
echo ""
|
||
echo "===== 监控完成 ====="
|
||
echo "详细日志文件:"
|
||
for pid in ${!LOG_FILES[@]}; do
|
||
echo " - PID $pid: ${LOG_FILES[$pid]}"
|
||
done
|
||
echo "汇总报告: $SUMMARY_LOG"
|
||
echo ""
|
||
echo "全局统计结果:"
|
||
echo " - 总监控进程数: ${#LOG_FILES[@]}"
|
||
echo " - 全局CPU平均使用率: $GLOBAL_CPU_AVG%"
|
||
echo " - 全局CPU最大使用率: $GLOBAL_CPU_MAX%"
|
||
echo " - 全局内存平均使用率: $GLOBAL_MEM_AVG%"
|
||
echo " - 全局内存最大使用率: $GLOBAL_MEM_MAX%" |