1

Calling all bash gurus, please give me a shove in the right direction. The following does what I need for only 1 host unreliably and logs too many lines. How do I make it more elegant and functional? Ideally, I would like it to ping with several packets and log ONLY CHANGES. Also, ping SEVERAL hosts. Only log one line when a host is not reachable/unreachable for several packets in a row.

#!/bin/bash

while [ 1 ]; do
        ping -c 1 -w 2 $1
        if [[ $? == 0 ]]; then
                spd-say "up"
                notify-send "up"
                echo "up $1 $(date +%Y%m%d-%H%M)" >> /tmp/ping.log
        else
                echo "down $1 $(date +%Y%m%d-%H%M)" >> /tmp/ping.log
        fi
        sleep 2m
done;
Improve this question

2 Answers 2

Reset to default
2

Like Romeo I recommend using cron instead of an infinite loop, but to solve your logging issue I came up with:

#!/usr/bin/env bash

hosts=("$@")

log=~/tmp/ping.log

[[ ! -f "$log" ]] && touch "$log"

check_log () {
    local h=$1
    local s
    s=$(awk -v h="$h" '$2 == h {print $1}' "$log" | tail -1)
    printf '%s' "$s"
}

for host in "${hosts[@]}"; do
    ping -qc 5 "$host" >/dev/null 2>&1 &
    pids+=([$!]="$host")
done
for pid in "${!pids[@]}"; do
    host=${pids[$pid]}
    s=$(check_log "$host")
    if wait "$pid"; then
        if [[ "$s" == down || -z "$s" ]]; then
            printf '%s\n' "up $host $(date +%Y%m%d-%H%M)" >> "$log"
        fi
    else
        if [[ "$s" == up || -z "$s" ]]; then
            printf '%s\n' "down $host $(date +%Y%m%d-%H%M)" >> "$log"
        fi
    fi
done

The check_log function will search the log file for entries for the given host, specifically the very last entry for said host and whether it was up or down.

The script will loop through each host and ping them with a packet count of 5. It pings them as a background process to try and speed things up a bit. Then it loops through the PIDs of the background ping commands and waits for them to complete. If the ping succeeded it checks the log to see if the last entry for that host was a failure, if so it will log an up entry for that host, if not it does nothing. If the ping command failed it will check the log to see if the last entry for that host was a success, if so it will log a down entry for that host, if not it does nothing.

Improve this answer
4
  • 1) thank you 2) since I had to reactivate my account to post this question, I guess I don't have the reputation necessary to upvote you, sorry 3) my initial impression is that you gave an excellent answer to my question. however, I am clueless as to how to employ an associative array. Could you possibly elaborate for minions like myself? Commented Dec 19, 2019 at 16:37
  • @bumblingfool: The declare -A pids creates an associative array named pids, really it's not even necessary since I'm using the PID instead of the hostname as the index but I initially was going to do it backwards. Anyway it allows you to set the index of the array element as a string rather than a number. So in the second loop we can loop through the PIDs and then have a hostname/IP address associated with each. Commented Dec 19, 2019 at 16:46
  • 1) thanks again 2) I was thinking in terms of feeding it a list of IPs/hostnames from a file but find that your offering accepts multiple command line arguments. 3) however, I need it to log something when hosts are down as well. 4) I'm sure I'm missing something... Commented Dec 19, 2019 at 17:00
  • @bumblingfool: Sorry the code as originally written wouldn't log down entries unless the host was previously up due to a flaw in the way I wrote it. I have modified it so that it should work either way now. Commented Dec 19, 2019 at 17:53
1

I would use an associative array to hold the status and only log status changes. Something like this:

#!/usr/bin/env bash

## This will let us use the host names as keys in the associative array 'status'
declare -A status

while :; do
  for host in "$@"; do
    ## Ping the server and, if the ping is successful, set $isUp to "Up", 
    ## if the ping fails, set $isUp to "Down".
    ping -c 1 -w 2 "$host" &>/dev/null &&
      isUp="Up" || isUp="Down"
    ## If the current value of $isUp isn't the same as what is stored in the
    ## status array for this server, we should report it. 
    if [[ ${status[$host]} != $isUp ]]; then
      spd-say "$host is $isUp"
      notify-send "$host is $isUp"
      printf "%s is %s\n" "$host" "$isUp" >> /tmp/ping.log
      ## save the current status in the array for this server.
      status[$host]=$isUp
    fi
  done
  sleep 2s;
done

You can then run it with the host names as arguments:

checkHost.sh hostname1.foo.com hostanme2.bar.com

If you cannot use associative arrays (if you're running an old bash version) you can use two separate arrays instead:

hosts=("$@");
while :; do
  for ((i=0;i<${#hosts[@]}; i++)); do
    host=${hosts[i]};
    ping -c 1 -w 2 "$host" &>/dev/null &&
      isUp="Up" || isUp="Down"
    if [[ ${status[i]} != $isUp ]]; then
      spd-say "$host is $isUp"
      notify-send "$host is $isUp"
      printf "%s is %s\n" "$host" "$isUp" >> /tmp/ping.log
      status[i]=$isUp
    fi
  done
  sleep 2s;
done
Improve this answer
0

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.