getting is in checking Log files

swapnil titar · 03-07-2016, 01:22 AM

Hi,

I have one requirement in my project, we need to clean log file location which is having plenty of log files.
but i need to find one string "Messages is InvalidXXXX:" some number like this.

can you tell me how can i do it ?

i am thinking of grep or fgrep, help me out.

Regards,
Swapnil Titar

TenTenths · 03-07-2016, 05:19 AM

grep to find the string, awk to slice the file name out of the resulting line, rm -f to force the deletion.

What have you done so far?

swapnil titar · 03-07-2016, 05:56 AM

Hi TenTenths,

I have created normal logic right now.here is code stuf. i used only one hard coded log file but i want to check all file at this location. which contains some message like "Invalidxxxxx". I want to find 1st how many file contains that message and the want to place some other loc. then clean up .

1 #!/bin/bash
2 #sceipt for checking file names uniquly and checking unique string in each file
3
4 String path="/home/tmadmin/SwapnilT"
5
6 read -p "Enter the path :" $path
7 if [ -e $path ]
8 then
9 echo " path is present / Exits"
10 fi
11
12
13 cat t2.log | grep "invalid*" | sort | uniq
14
15

If you tell me atleast flow will also help me to proceed further coding ???

Regards,
Swapnil Titar

pan64 · 03-07-2016, 06:10 AM

please use [code]here comes your script[/code] to keep formatting
Also

Code:

#instead of:
cat t2.log | grep "invalid*"
#you can use
grep invalid t2.log

Code:

sort | uniq
#is equal to
sort -u

I do not know what is that path good for, but you did not exit if path was present, just message is printed.

TenTenths · 03-07-2016, 06:12 AM

In this line:

Code:

cat t2.log | grep "invalid*" | sort | uniq

you don't need to cat in to grep, you should just use

Code:

grep "invalid" *

to check all the files in whatever folder you're in.

So the flow would be:

Enter the path you want to check
Use cd to change to that path
grep "invalid" * to find all the files that contain your search
Pass the results of grep in to awk to slice out the filename
Pass that through uniq and wc -l to find the number of files

Repeat the above but use rm -f to remove the file