Find multiple strings in a file and email on occurence of any defined string

Question

I want to write a shell script which will find the occurence of multiple strings like "Errors|Notice|Warnings" from a given log file, such as /var/log/messages. If any string matches it should send a mail notification to specified mail ID.

I can use:

grep -i -E '^Errors|Notice|Warnings' /var/log/messages

But my main problem is, the log file always growing, and if I want to add this script in cron, how can I record the file line or contents that I had already checked on the last execution of my script?

For example, if the log file is 100 lines and I have read the file using cat or anything similar, then before second execution, the file becomes 300 lines, then now i want to read from 101 line number to 300.

Can anyone please suggest how I can record this?

Answer 1

You can use following script to do that:

start=0

[[ -f last-processed ]] && start=$(<last-processed)

start=$((start+1))

tail +$start /var/log/messages | grep -i -E 'Errors|Notice|Warnings' &&\
wc -l /var/log/messages | awk '{print $1}' > last-processed

btw you have a problem in your regx, it should be 'Errors|Notice|Warnings' instead of '^Errors|Notice|Warnings'

Answer 2

Rotating your log file could be the best solution.

But if you want to grep file from line_first to line_last you can use sed:

For example, get line from 100 to 110 from input stream:

$> line_first=100; line_last=110
$> seq 1 1000 | sed -n "${line_first},${line_last}p"
100
101
102
103
104
105
106
107
108
109
110