[SOLVED] Output IP address and exec command after counting occurrences of string using bash

Issue

I have the following sample data from the log file:

[10:35:52] world_log_data: user (starrr)(NormieBL)@Arca is already connected on srv 21
[10:35:53] world_log_data: user (starrr)(NormieBL)@Arca is already connected on srv 21
[10:35:54] world_log_data: user (starrr)(NormieBL)@Arca is already connected on srv 21
[10:35:54] world_log_data: user (starrr)(NormieBL)@Arca is already connected on srv 21
[10:35:54] world_log_data: user (starrr)(NormieBL)@Arca is already connected on srv 21
[10:35:56] world_log_event: kick (starrr)(NormieBL)@Arca from srv 192.168.1.6(21)  
[10:39:17] world_log_data: user (chrisxJ02)(Delaon)@Arca is already connected on srv 7
[10:39:19] world_log_data: user (chrisxJ02)(Delaon)@Arca is already connected on srv 7
[10:39:23] world_log_event: kick (chrisxJ02)(Delaon)@Arca from srv 192.168.1.39(7)
[10:39:17] world_log_data: user (test01)(testDW)@Arca is already connected on srv 39
[10:39:19] world_log_data: user (test01)(testDW)@Arca is already connected on srv 39
[10:39:23] world_log_event: kick (test01)(testDW)@Arca from srv 192.168.1.100(39)

I need to count the number of occurrences of a string using bash and execute a command once the string repeats itself more than 5 times.

• the string to search for in the log file is is already connected on srv
• the number that follows the string (eg, 21) can change after the process is restarted, so the complete string will be something like this is already connected on srv # where # is the ID number of a PC and can randomly be anywhere from 1 to 100
• a given number cannot have more than one IP address assigned (eg, 21 can only have IP_A(21) and it cannot have IP_A(21) and IP_B(21)
• the IP address can change after process restart, so the log file may register a different IP address for number 21 next time it restarts (eg, 192.168.1.6(21) can become 192.168.0.72(21)
• the code awk or grep I guess would the best should find the lines containing the string + number and count them; if more than 5 occurrences are the same it should execute the command
• if multiple entries show up 5+ times the code should flag all of them, however, this shouldn't be necessary as I plan to truncate the .log file right after detection
• the point of interest is parsing the IP address after the script detects multiple occurences of the mentioned string and then telnet into that IP address.
• both the IP address and the ID number associated with it will change so two variables will be needed, the only static string the script should look for is is already connected on srv and followed by the ID number

Some examples of how the script should behave:

if string "is already connected on srv 21" count is =>5 times then "exec command telnet 192.168.1.6"
if string "is already connected on srv 7" count is =>5 times then "exec command telnet 192.168.1.39"

Here is an attempt using awk but the IP address is not correctly displayed under the regular format containing dots such as 127.0.0.1. Probably grep would be a better choice.

#!/bin/bash

search_string='is already connected on srv'
threshold=5


awk -v ss="${search_string}" -v threshold="${threshold}" '

$0 ~ ss    { counter[$NF]++ }                    # counter[##]++

/ IPaddr_/ { for (i=2; i<=NF; i++)               # loop through fields ...
                 if ($(i) ~ "IPaddr_") {         # looking for string "IPaddr_"
                    split($(i),arr,"[()]")       # split "IPaddr_?(##)" on parents
                    ip[arr[2]]=arr[1]            # ip[##]=IPaddr_?
                    next}                        # skip to next input line
           }

END        { for (i in counter)                  # for every "##" encountered ...
                 if (counter[i] >= threshold)    # if the count is greater than threshold then ...
                    print i,ip[i]                # print "## IPaddr_?"
           }
' world.log

Solution

Assumptions:

• we only have to deal with IPv4 formats
• the only strings like <space><IPv4-address>( in the file are the ones we're interested in

One idea is to modify the current awk code to look for <space><IPv4-address>(:

awk -v ss="${search_string}" -v threshold="${threshold}" '

$0 ~ ss    { counter[$NF]++ }

/ [0-9]+.[0-9]+.[0-9]+.[0-9]+[(]/ {

             for (i=2; i<=NF; i++)
                 if ( $(i) ~ "[0-9]+.[0-9]+.[0-9]+.[0-9]+[(]" ) {
                    split($(i),arr,"[()]")
                    ip[arr[2]]=arr[1]
                    next}
           }

END        { for (i in counter)
                 if (counter[i] >= threshold)
                    print i,ip[i]
           }
' world.log

Run against the sample input with threshold=5 generates:

21 192.168.1.6

With threshold=2 this generates:

7 192.168.1.39
21 192.168.1.6
39 192.168.1.100

---

At this point OP can manipulate the awk output as appropriate (eg, load into array(s), load each row into while loop variables, etc).

For while loop processing it will be a bit cleaner if we move the awk code into a function, eg:

parse_srv_ip ()
{
    awk -v ss="${search_string}" -v threshold="${threshold}" '
    ... snip ...
    ' "$1"              # name of log file passed as only arg to function
}

We can then process the awk output via a while loop like such:

logfile='world.log'

while read -r srv ip
do
    echo "srv = $srv : ip = $ip"
done < <(parse_srv_ip "${logfile}")

This generates:

srv = 7 : ip = 192.168.1.39
srv = 21 : ip = 192.168.1.6
srv = 39 : ip = 192.168.1.100

NOTE: OP would obviously replace the echo "srv ..." with the desired code (eg, "exec command telnet...", etc).

Answered By - markp-fuso