Linux

Stuff I forget again and again:

Rename files according to their timestamp

This one-liner renames all images in a directory according to their file modification date. The resulting file names have 8 digits with leading zeroes. Note that it does not distinguish file types - it works just for jpg as it is printed here.

i=0; for f in $(ls -lrt | cut -c 52-65); do newname=$(printf %08d $i); echo "mv $f $newname.jpg"; mv $f $newname.jpg; let i=i+1; done

Send mail from command line

echo "mail content" | mail -s "subject" mail@domain.tld
(echo "mail content"; cat somefile.txt} | mail -s "subject" mail@domain.tld

All clear?

What was this || and && again?

if exp; then T; else F; fi
is equal to
exp && T || F
other way round doesn't work :-)
$ [ X = xxX ] || echo F && echo T
F
T 
}}}

All clear?

Set editing mode

For me the "intiutive" editing mode is as emacs behaves:

set -o emacs

Filename of a path only

basename /usb/bin/su

yields: su

Find biggest subdirectories

Find all subdirectories of the current dir and get size of each of this directories (including children):

find . -type d -maxdepth 1 -exec du -h -s {} ;

Mount ISO image

mount -o loop,ro -t iso9660 image.iso /mnt/iso

Rsync using SSH

rsync -rtvz -e 'ssh -l user' src-dir/ user@host:dest-dir/

• a for archive (=rlptgoD)
• D Devices
• g groups
• l symlinks
• n dry run
• o owners
• p permissions
• r recurse
• t times
• v verbose

unison

Nice mirroring tool available at http://www.cis.upenn.edu/~bcpierce/unison/ It provides either text or graphical interface for full two-way synchronization of directory trees including conflict resolution and remote operation. And this even without server operation, SSH does the job, very nice!

Spyder whole Webpage

Some people like to point and click, but is that nice or what:

wget -m http://foo/[bar/]
a bit more advanced, but doesn't really do the trick, because it follows all calendar links, need to disable this.
wget --convert-link --mirror --no-parent --relative --page-requisites --exclude-directories=exec http://www.stefanrufer.ch/snipsnap/space/start

Get server fingerprint

If I am asked by ssh whether a server fingerprint is correct - I first need to have the correct fingerprint. Given the situation I can get hold to a console of the machine in question I really trust - here's how to get the fingerprint (linux machine, pretty much standard config I would say):

ssh-keygen -l -f /etc/ssh/ssh_host_key.pub
ssh-keygen -l -f /etc/ssh/ssh_host_dsa_key.pub
ssh-keygen -l -f /etc/ssh/ssh_host_rsa_key.pub

HTTP/1.1

How to get a page from a webserver if connected by telnet? I always forget this HTTP/1.1 syntax:

foo@bar > telnet somehost 8080
GET /index.html HTTP/1.1
Host: somehost

(or use HEAD, OPTIONS HTTP commands instead)

Cron

Fields: minute(s) hour(s) day(s) month(s) weekday(s) command(s)

The fields are separated by spaces or tabs. The first five are integer patterns and the sixth is the command to be executed. The following table briefly describes each of the fields.

The integer fields can contain patterns like */5 (every five minutes) or 0,8,16 to run at hours 0, 8 and 16.

You may want to specify MAILTO=foo@bar.com as first line in the crontab file to have the output of a cron run mailed to this address.

Output redirection

In order to get all output (stdout and stderr) to a logfile (or /dev/null), redirect e.g. stderr to stdout:

foo-command.sh >> foo-command.log 2>&1

Backup

• (+) Snapshots http://www.mikerubel.org/computers/rsync_snapshots/
• (+) MySQL http://www.linux-fuer-alle.de/doc_show.php?docid=234&catid=4 http://sourceforge.net/projects/automysqlbackup/
• misc http://www.linux-backup.net/Full_Inc/
• simple http://www.xs4all.nl/~voorburg/backup.html
• some theory http://kmself.home.netcom.com/Linux/FAQs/backups.html

wget

Most useful: continue canceled downloads:

wget -c URL