Adjusting server logging level at runtime

log4j is an awesome logging library for Java. One thing I found out today is that you can change the logging level (debug, warn, info) at runtime, with a few lines of code.

static
{
       // This monitors the log4j for changes over a specified period of milliseconds
       PropertyConfigurator.configureAndWatch("./resources/log4j.properties", 60000);
}

What this function does is to tell log4j to check the configuration file every minute (60000 ms) for changes. So you can deploy a production server with only warn and error logging, but if something goes wrong you can enable debug and info logging (for example) simply by editing the configuration file and waiting a minute.

(source code formatted by this handy website http://formatmysourcecode.blogspot.com)

Growling Mac

(pic from flickr)

Last year I wrote a blog post about how to make growl notifications (little pop ups driven by the application Growl) appear from emacs in windows. Since then I've become a total Mac-head. Here's an update on how you get that working from emacs on a mac...

Everything is the same except you need to install Growl for Mac (obv) and also make sure you copy the folder called Extras somewhere.

Open up a terminal and then cd to where you put the Extras directory. Find a sub-directory called growlnotify. Enter that and run 'sudo ./install.sh' which will prompt you for your admin password, and then go ahead and install growlnotify.

You can now run growlnotify from a terminal window. Try it out, if it doesn't work you won't be able to growl from emacs.

Now follow the instructions in my blog post above but ignore all the windows parts. You'll need to customize the growl command (M-x customize group, todochiku) and replace the full path I used for windows with just 'growlnotify'.

If everything is cool you can now do M-x todochiku-in and type a message, a time in minutes (zero for now), and pop, you should have a notification.

Using tar and gzip

Picture from flickr by Windwirral

Here are the basic steps to use tar and gzip to package up some files in a compressed archive, so that you can move them to another computer or back them up.

First you want to create and add files to the archive. Go to the target directory and enter individual filenames and directories you want to have included.

tar -vzcf mystuff.tar.gz file1.txt file2.doc mypictures/

Now you can test if that worked by listing the files in the archive.

tar -tvf mystuff.tar.gz

When you want to extract the files, go to the folder where you want to extract them, in this case 'targetfolder'.

cd targetfolder

tar -vxf mystuff.tar.gz

Nice Code

Just wanted to point out a couple of very nice pieces of source code I came across recently.

ipin.py

Firstly check out this Python code ipin.py, by Axel E. Brzostowski, which converts png files from an iphone application into a format that you can read on any computer.

The python code is clean, easy to read, almost like literate programming. Very useful if you want to write a similar application that processes every file in a directory.

transmission

Another really nice program, is really a website. When you run transmission-daemon (linux bittorrent program), you can connect to it with a command line app called transmission-remote. In addition you can connect via a built in web server, which is a really excellently designed application. The html and javascript is so well documented and tidy it's as beautiful as the actual web page. Check it out here.

http://trac.transmissionbt.com/browser/trunk/web

Securing an ssh server with fail2ban

There are all kinds of ways to secure an ssh server, with varying degrees of increasing security and decreasing flexibility. For example by limiting your server to only accept connections from certain known IP's, you are secure from random hackers on the internet, but you lose the ability to connect to your machine from anywhere you want to. Perhaps while travelling, for example.

Port knocking and listening on a high numbered non-standard port. make it harder for an attacker to even start trying to hack your connection. But this also requires you connect with a machine that you have the knock program installed on. Again, less convenient, more secure.

However, once an attacker does find your port there's nothing to stop brute force password hacking. If you look in your log file, you should see people connecting to your ssh port quite frequently and trying password attacks.

cat /var/log/auth.log

If you have a secure password then it would require days of brute force hacking to gain access to your ssh account, but even so, if you don't watch your logs then it's perfectly possible somebody will gain access eventually.

Brute force attacks can be limited using fail2ban. There's a great article on setting it up here. This program will scan your auth.log for you, using a regular expression to find failed password attempts. On a specified number of failures from a given IP, it will then modify the iptables on your machine (the firewall), to lock that IP out for a specified time.

Now instead of watching your auth.log fill up with reams of failed passwords, you'll see a greatly reduced amount of brute force attacks, and you can watch your fail2ban log file fill up with the IP addresses of hackers.

Using screen

When running a linux app, perhaps on a remote box, you don't want it to terminate when you close the terminal window. You can run some applications as daemons, or run them with the nohup command (no hang up).

For example:

nohup ./myapp > output.txt

will run the program myapp and send the output to output.txt.

You can then follow the output.txt with a command

tail -f output.txt

but if you hit Ctrl-C or close the terminal (or have a power cut), the application will still be running on the remote machine.

A more powerful solution is the Gnu screen application. This lets you run multiple shell sessions and switch between them. Using screen you can be logged into a server in your office, then detach from the screen session, go home and reattach to it there (assuming you have network access to the computer).

Here's a cheat sheet for using screen.

Running it:

screen

Learning to use it:

Ctrl-a help

Important commands:

Ctrl-a c (open a new window)

Ctrl-a p (prev window)

Ctrl-a n (next window)

exit or C-a q to exit

How to reattach to the screen session:

screen -ls

That shows screen sessions on the machine your logged on to, and then you reattach using:

screen -r name

That's all folks!

Transparent emacs on windows

Here's a handy function which lets you choose a transparent level (0 is fully transparent and 100 is opaque), for the main emacs frame in both focused and unfocused state.

(defun transparent(alpha-level no-focus-alpha-level)
 "Let's you make the window transparent"
 (interactive "nAlpha level (0-100): \nnNo focus alpha level (0-100): ")
 (set-frame-parameter (selected-frame) 'alpha (list alpha-level no-focus-alpha-level))
 (add-to-list 'default-frame-alist `(alpha ,alpha-level)))

To run M-x transparent. 

Enter the values you want for when the window has focus and when it does not. In the background you can see system status information on the desktop, which is another excellent utility from sysinternals called BgInfo. It let's you display useful info about your system right on your desktop, similar to tools you may find on linux.