Using 'find' and 'grep' in a for loop

linuxlastslonge · 03-23-2011, 01:31 PM

I have a server hosting 100+ websites. I need to quickly identify which websites are configured with a database. There are way too many to manually check every website for a PHP file with a database name. So, I created a list of all databases from MySQL and put them in a text file. I then exported the text file to a shell variable and used it in a for loop.

bash variable

Code:

DBLIST=`cat dblist.txt`

Example of $DBLIST

Code:

db1 db_testing2 database_clientname production words4cheap

for loop

Code:

for db in $DBLIST; do find . -type "f" -iname "*.php" -exec grep -i $db '{}' \; -print; done

Note: my find statement starts searching at . which is the directory that contains all of my websites and their data, each website is setup in a sub directory, identified by it's domain name.

Example: I'm in /var/www. Beneath /var/www are a list of directories:

blah.com
foobar.com
blahfoo.com

However, this is taking too long (it's been running most of the day) and I was wondering if there wasn't an easier way to accomplish what I'm trying to achieve?

Thanks!

corp769 · 03-23-2011, 01:59 PM

What distro are you running? If you have the resources installed, you could use locate, since updatedb is ran as a scheduled cron job and creates a index for using locate. It is very fast, and I use it all the time. Hope that helps,

Josh

linuxlastslonge · 03-23-2011, 02:14 PM

Good idea, but the real lag is coming from looping through 100+ databases and grepping every php file in /var/www for a match. I guess I'm just stuck playing the waiting game ... :-\

corp769 · 03-23-2011, 02:19 PM

I know that, but using locate is usually faster than find.

MTK358 · 03-23-2011, 04:16 PM

Quote:

Originally Posted by linuxlastslonge

Code:

DBLIST=`cat dblist.txt`

Use $(command) instead of `command`. It can't be confused with commas and is more easily nested.