Question

我在bash练习但是我在这个练习中遇到了很大的问题，你能帮助我吗？

这是练习：

创建一个shell脚本，该脚本将列出所有用户数量少于root且％mem的用户数至少大于根进程。对于这种类型的每个用户，脚本将创建一个文件.txt，其中列出了此类进程的PID，PPID，％MEM

可悲的是，我的尝试根本不起作用，它在第7行报告问题＆＃34; ps：等待整数表达＆＃34;。

#! /bin/bash

N_root=$(ps hax -o user | sort | uniq -c | grep root | awk '{print $1}')     # here I save how many processes root has
mem_root=$(ps -u root -o %mem | sort | head -1)                                # I save the lowest %mem of root
for i in $(ps hax -o user | sort | uniq -c | sed '/root/d')             # listing all processes
do
    if [ $(echo $i | awk '{print $1}') -lt $N_root ]                     # i check if this user has less processes than root
    then
        user=$(echo $i | awk '{print $2}')                         # i save the name of the user
        if [ $(ps -u $user -o %mem | sort -r | head -1) -gt $mem_root ]   # check if the max %mem is greather than the lowest %mem of root
        then
            echo $user
            touch <$user>.txt                                  # creating file "<user>.txt"
            for p in $(ps -u $user -o %mem,pid,ppid | sed '/PID/d' | sort -r)  # listing all processis of current user ($n)
            do
                if [ $(echo $p | awk '{print $1}') -gt $mem_root ]             # check if the process has more %mem than root
                then
                    echo $p >> "<$user>.txt"                           # copying %mem,pid,ppid in the file
                fi
            done
        fi
    fi
done

Answer 1

有很多方法可以解决这个问题。对于多重比较，我通常喜欢将值索引到数组中，然后通过值（如果可能）通过单个循环执行比较。适用于您的问题，它可能如下所示：

#!/bin/bash

## set output filename
ofn=./userpctmem.txt

## arrays to hold procs, users, and %mem
declare -a procarray
declare -a userarray
declare -a memarray

## read values into arrays
while read -r pval uval || [ -n "$uval" ]; do
    procarray+=( "$pval" )
    userarray+=( "$uval" )
    memarray+=( $(ps -u "$uval" -o %mem | grep -v '%' | sort -r | head -n 1) )
done <<<"$(ps hax -o user | sort | uniq -c)"

## save root procs and %mem in separate variables for convenience
rproc=$(ps -u root | tail -n +2 | wc -l)
rmem=$(ps -u root -o %mem | grep -v '%' | sort -r | head -n 1)

## get the number of users found
nusers=${#userarray[@]}

## print the captured data (for verification)
printf "\n%-12s  %-6s  %s\n\n" "user" "nprocs" "maxmem"
for ((i=0; i<$nusers; i++)); do
    printf " %-12s %5s   %s \n" "${userarray[i]}" "${procarray[i]}" "${memarray[i]}"
done

## test user process less than root and %mem >= root (max %mem process)
#  and write matching users to output
printf "\nUsers with nprocs less than root and %%mem at least as great as root\n\n"
for ((i=0; i<$nusers; i++)); do
    [ "${userarray[i]}" == 'root' ] && continue
    umem="${memarray[i]}"
    if [ "${umem//.*/}" -ge "${rmem//.*/}" -a "${procarray[i]}" -lt "$rproc" ]; then
        printf " %s has process with %%mem: %s, root %%mem: %s\n" "${userarray[i]}" "$umem" "$rmem" | tee -a "$ofn"
    fi
done

printf "\n"

exit 0

<强>输出：

alchemy:~/scr/tmp/stack> bash procmem.sh

user          nprocs  maxmem

 avahi            1   0.0
 david           68   18.7
 man              2   0.0
 messagebus       1   0.0
 nscd             1   0.0
 ntp              1   0.0
 polkitd          1   0.5
 postfix          2   0.0
 root           105   2.8

Users with nprocs less than root and %mem at least as great as root

 david has process with %mem: 18.7, root %mem:  2.8

输出文件：

alchemy:~/scr/tmp/stack> cat userpctmem.txt
 david has process with %mem: 18.7, root %mem:  2.8

注意：检索具有最大％MEM的流程的PID和PPID留待练习。如果您在尝试后遇到困难，请发表评论，然后我会告诉您一种捕捉它的方法。

在进程中进行bash练习

1 个答案: