rootninja

Cron is so simple yet so useful I think it’s often unappreciated! Being able to toss a command into a cron, or execute a script at specific times of the day, days of the week, etc. That’s good stuff!

You could call a script that checks your mail once an hour, cleans out a temporary directory, or archives a special directory where you dump files to get saved without ever having to visit the server manually and tar/gzip the files yourself.

My latest use for cron is a nightly build script for building code checked out of subversion.  I’m doing this on my system for personal use, but also at work along with a python script i’m writing that also runs from a cron a few hours later.  It parses the build log looking for errors.  If it sees errors, it sends them out in email.  Python is wonderful… /off topic.

# +———— minute (0 - 59)
# | +———- hour (0 - 23)
# | | +——– day of month (1 - 31)
# | | | +—— month (1 - 12)
# | | | | +—- day of week (0 - 6) (Sunday=0 or 7)
# | | | | |
* * * * * command to be executed

If you want to run a script called check-for-rootkits.sh every 10 minutes, its as easy as:

*/10 * * * * /root/check-for-rootkits.sh

You could also get fancy and use “run as” to run crons as specific users.  Users can do this on their own (”crontab -e”), but you could force things to happen for them with your root crons if you wanted.  Fedora makes use of this for running hourly and nightly cron directories.  If you want anything to run along with the standard Fedora cron jobs, you just drop your script into those directories and they fire off alongside the defaults.

Linux cron, fedora, Linux, subversion

FSVS is the abbreviation for “Fast System Versioning” and is pronounced [fisvis].

It is a complete backup/restore/versioning tool for all files in a directory tree or whole filesystems, with a subversion repository as the backend. 

Using this application, all files(almost) in /etc/ are maintained in a subversion repository. As files get created in /etc/ they are added to the repository automatically. commit’s to the repository are executed on a daily basis with a cron job in /etc/cron.daily.  Configuration creation, updates, and deletions are easy to see and track down long after the fact, eliminating questions like, “When did this get changed?? and what the heck did it look like before!?”

• fsvs does not pollute directories with .svn folders.
• It keeps track of everything with file lists and hashes in /etc/fsvs/ and /var/spool/fsvs/
• All fsvs repositories can be stored in one place to make it easy to back up.

I think of it as a safety parachute, incase the regular one doesn’t work. (my regular method of backing up a config is to copy the original file with the extension “.old”, “.bak”, or “.orig”)

How to install it

Commands to run from the subversion server, “TheServer”

# yum install fsvs
# mkdir /RAID/fsvs_repos
# svnadmin create /RAID/fsvs_repos/TargetHost

Commands run from the client machine, “TargetHost”

# yum install fsvs
# cd /etc   ($pwd must be the base directory you want to use to populate the repository)
# fsvs urls svn://theserver/fsvs_repos/TargetHostName  (this points to the berkley db you just created on “theserver”)
# fsvs ignore ./fsvs
# fsvs ignore ./ld.so.cache
# fsvs ignore ./prelink.cache
# fsvs ignore ./lvm
# fsvs ignore ./localtime
# fsvs ignore ./.pwd.lock
# fsvs ignore ./selinux
# fsvs commit -m “Initial fsvs commit”

Daily cron commit setup

A one liner cron in /etc/cron.weekly/fsvs.cron should be all you need.

/usr/bin/fsvs commit -m “cron commit at `date +%R_%d-%m-%Y`”

Linux backups, cron, etc, fsvs, Linux, subversion, svn

You can take several regular pictures or other image files and put them together into one image that looks like they’re all polaroids laid out on a desk using ImageMagick’s montage tool.  If you have a dual monitor setup you can easily fit 8 across your screens at once without even overlapping much.  Or you could lay them out half-covering each other to get a whole album on the screen at once.  Don’t worry about the size of each image either because you’ll take thumbnails of each of them to make the final larger image.

[user@hostname ~]$ /usr/bin/montage -size 3200×1200 ~/Pictures/*.jpg -thumbnail 722×594 -bordercolor Lavender -background black +polaroid -background DarkGray -geometry -30-47 -tile x2 ~/Pictures/montage.gif

You could rotate the pictures in a script and run it in a crontab every hour to rotate your pictures for the next time you make the montage pic.  If you’re not constantly downloading a new image (such as a satelite weather map or something), you could bump each file’s name along, saving the last one in a temporary position before rotating it back to the beginning of your stack.  The simplest way to do this would be just moving the file names one by one.  Think of it as files “new”, “old”, “older”, and “oldest”.   Use “temp” as a placeholder while you bump the names along in the chain.

[user@host ~]$ mv -f oldest.jpg temp.jpg; mv -f older.jpg oldest.jpg; mv -f old.jpg older.jpg; mv -f new.jpg old.jpg; mv -f temp.jpg new.jpg

And if you’re using Gnome you can run gconftool-2 to swap out your background image on the fly everytime the cron runs.

[user@host ~]$ /usr/bin/gconftool-2 –type string –set /desktop/gnome/background/picture_filename “$HOME/Pictures/montage.gif”

Linux, ooo! Shiny... backgrounds, cron, ImageMagick, montage, polaroid