代码之家 › 专栏 › 技术社区 › Bogdan Stoica

Bash循环通过多行netstat命令输出

shell bash

0

Bogdan Stoica · 技术社区 · 5 年前

我正在运行此命令(将输出保存到文本文件:

netstat -ntp | grep tcp | grep EST | grep 34341

tcp      593      0 10.10.1.11:43856       10.10.2.14:3434      ESTABLISHED     146597/daemon-
tcp      417      0 10.10.1.11:43859       10.10.2.15:3434      ESTABLISHED     146567/daemon-
tcp      317      0 10.10.1.11:43121       10.10.2.16:3434      ESTABLISHED     146582/daemon-

#! /bin/bash

SLEEP=5
COUNTER=0

recvq()
{
    while read -r proto recvq x local remote state x
    do
        if [[ "$proto" == tcp && "$state" == ESTABLISHED && "$remote" =~ .*:3434 ]]
        then
            printf "%d\n" "$recvq"
        fi
    done < "$1"
}

while true; do

        (( COUNTER++ ))
        # measure recvq value
        declare -A first
        while read -r socket recvq
        do
            first[$socket]=$recvq
        done < <(recvq netstat1.txt)

        # sleep
        sleep "$SLEEP"

        # measure recvq value after sleep
        declare -A second
        while read -r socket recvq
        do
            second[$socket]=$recvq
        done < <(recvq netstat2.txt)

        [ ${#first[*]} != ${#second[*]} ] && { echo "Arrays are different size"; }

        for ii in ${!first[*]}; do
            [ "${first[$ii]}" == "${second[$ii]}" ] || { echo different element $ii; exit 1; }
        done
        echo "Arrays are identical"

done

现在我需要比较文件中每一行的recvq(睡眠前)和recvq(睡眠后)的值。如果任何初始recvq值与最终recvq值相同,则采取措施。

0 回复 | 直到 5 年前

1

ceving 5 年前

我会这样解析:

#! /bin/bash

while read proto recvq x x port state x
do
  if [[ "$proto" == tcp && "$state" == ESTABLISHED && "$port" =~ .*:3434$ ]]
  then
    printf "%d\n" "$recvq"
  fi
done < <(netstat -ntp)

你不需要 grep , cat 或 awk

请勿进行复制粘贴编程。如果你有代码,你想重用,把它放在一个函数里。

Bash具有关联数组,您可以使用它来存储每个套接字接收的数据。

#! /bin/bash

recvq()
{
  while read proto recvq x local remote state x
  do
    if [[ "$proto" == tcp && "$state" == ESTABLISHED && "$remote" =~ .*:3434$ ]]
    then
      printf "%s/%s %d\n" "$local" "$remote" "$recvq"
    fi
  done < "$1"
}

# First measure

declare -A first
while read socket recvq
do
  first[$socket]=$recvq
done < <(recvq netstat1.txt)

# Wait

sleep 10

# Second measure

declare -A second
while read socket recvq
do
  second[$socket]=$recvq
done < <(recvq netstat2.txt)

# Compare measures

for socket in "${!first[@]}"
do
  if [[ "${first[$socket]}" == "${second[$socket]}" ]]
  then
    printf "match for %s: %d\n" "$socket" "${first[$socket]}"
  fi
done

2

0

Bogdan Stoica 5 年前

我终于明白了。也许它可以像@ceving所建议的那样,很好地处理关联数组,并提供了一个答案,但它并没有像预期的那样工作。

不过,我不会把自己的答案标为正确答案。我刚把它贴出来,以防有人觉得有用。

#/bin/bash

n1="$(cat netstat1.txt | awk '{split($5,a,":"); print a[1]"="$2}')"

sleep 5

n2="$(cat netstat2.txt | awk '{split($5,a,":"); print a[1]"="$2}')"

# check
for i in $n1; do
    ip1=`echo $i | awk '{split($i,a,"="); print a[1]}'`
    val1=`echo $i | awk '{split($i,a,"="); print a[2]}'`
    #echo $ip1 $val1

    for j in $n2; do

        ip2=`echo $j | awk '{split($j,a,"="); print a[1]}'`
        val2=`echo $j | awk '{split($j,a,"="); print a[2]}'`
        #echo $ip2 $val2

        if [ $ip1 == $ip2 ]; then
            if [ $val1 == $val2 ]; then
                echo "values: $val1 vs $val2 equal for remote ip: $ip1 => NOK! Do whatever!"
            else
                echo "values: $val1 vs $val2 dfferent for remote ip: $ip1 => OK"
            fi
        fi
    done
done