Simple command to find out the directory disk usage
ls -F | grep “/” | xargs du -hs
For admin to snoop everybody, try
ls -F | grep “/” | sudo xargs du -hs
ssh public/private key using ssh-keygen
me@connect-from-machine:~> ssh-keygen -t rsa
Generating public/private rsa key pair.
Enter file in which to save the key (/home/me/.ssh/id_rsa):
Enter passphrase (empty for no passphrase):
Enter same passphrase again:
Your identification has been saved in /home/me/.ssh/id_rsa.
Your public key has been saved in /home/me/.ssh/id_rsa.pub.
The key fingerprint is:
blah blah ….
The command ssh-keygen -t rsa initiated the creation of the key pair.
The private key was saved in .ssh/id_rsa. The file is read only by you and you can see why since if someone else get it he can try to use it to connect to the same machine. If you happen to make it passphrase free, it will be even easier for the other person.
The public key is save in .ssh/id_rsa.pub.
Its content is then copied in file .ssh/authorized_keys of the system you wish to SSH to without being prompted for a password. Using scp, this can be done by scp .ssh/id_rsa.pub username@remote-machine:/.ssh/authorized_keys .
Once the file is copied you are on ssh private key login.
Use “rm” to delete files, a lot of files
The other day when I use Xenu to check the broken links the /tmp directory was quickly filled up. When I tried to remove the temporary files using “rm *” command I was told “bach:/bin/rm Argument list too long”. Apparently there is a max number of files that rm can handle.
A little bit search on the web yielded this following solution:
use the find command and xargs, the file can be removed one by one. Like this in my case:
find . -name ’sess_*’ | xargs rm
Since find command can also look for files with certain age, it is a great way to clean up files in this way.
A good find about “find”
“find” command, like a lot of other UNIX/LINUX commands, can be really powerful if used with the right options. This is an article about using “find” in some different and very useful ways.
cron job reference
The crontab is a scheduled job utility on *nix server.
To edit a cron schedule:
crontab -e
This will open the schedule file based on the EDITOR setting in your shell. VI is the most commonly used editor.
A cron schedule file can look like this:
MAILTO="support@phinesolutions.com"
*/10 * * * * php -f Ping.php > ping.out
MAILTO sets the address where the output of the cron goes. If you don't even want to receive it, set MAILTO="".
The first part of the second line is the schedule, and the second part is the command.
The schedule
The job schedule is set in the following sequence:
- minute of the hour
- hour of the day
- day of the month
- month of the year
- day of the week
So in the example above:
- */10 means every 10 minute in an hour
- 2nd * means every hour of the day
- 3rd * means every day of the month
- 4th * means every month of the year
- 5th * means every day of the week
*/10 is the abbreviated way of saying "every 10 minutes in an hour".
Two other commonly used forms are comma separated values and a range using "-" (dash). For example:
0 9,12 * * * sets the job to run on top of the hour at 9am and 12 am, every day.
0 9,12 * * 1-5 sets the job to run on top of the hour at 9am and 12 am, every weekday.
The command
The command can be anything executable, either be a unix command or a script. A good practice is to test the command from the console and make sure everything is setup environmentally and then put them into the crontab. Since I want to track the output the above example writes the output from the script to a file.
Sometimes I prefer the logging implemented in the script and do not want to receive it from cron email, I can set the it like this:
*/10 * * * * php -f Ping.php > /dev/null
/dev/null is the almighty black hole of a Unix system and it will swallow everything without a trace.
If I really don't want to be bothered with ANYTHING, I can also do this:
*/10 * * * * php -f Ping.php > /dev/null 2>&1
This may look weird the first time you saw it. In a nutshell, the Unix shell assigns different handlers to STDOUT and STDERR, and "2>&1" means all the errors are re-directed to standard output.
We can also create a separate cron file that has all the jobs and use the crontab command to load them up in once:
crontab filename
Some other notes:
Some hosting does not encourage the use of the crontab from a console. In most cases, a control panel like Cpanel provides a nice interface to set cron jobs.
Once I tried to set up cron job and it kept give me some weird error like "command not found", which made no sense since I have tested everything from the script to the PATH. It turned out the cron file I was trying to load is in DOS format and the carriage return messed it up. I ran "dos2unix" on the cron file and fixed it.