UNIX - How to

External References

Writing scripts for cd

edit .bashrc

alias cdr="cd /cul/web/elr37-dev.library.cornell.edu/htdocs/rails/;pwd"

Reference: .bash_profile vs .bashrc by Josh Staiger

NOTE: Requires .bash_profile to have the following code.

edit .bash_profile

if [ -f ~/.bashrc ]; then
  source ~/.bashrc
fi

Updating $PATH

# view PATH
echo $PATH

# update PATH
PATH=$PATH\:/dir/path ; export PATH

To make this change persistent, add the following to .bash_profile

edit .bash_profile

export PATH="$PATH:~elr37/bin"

NOTE: cd scripts won't work correctly from bin. Need to add them as an alias in .bashrc. See notes under Writing scripts for cd.

Append a line to a text file

echo "line to append" >> path/to/file

echo "gem: --no-document" >> ~/.gemrc

Check file's md5

CYGWIN

md5sum FILE_NAME

Mac terminal

openssl md5 FILE_NAME

sudo, type, whereis, whoami, scp, echo, grep

Command	Example	Action	Comments
`sudo`	`sudo mkdir test`	run command as root
`sudo -s`		open a root shell
`sudo -s -u` `user`	`sudo -s -u` net33	open a shell as user
`type` `cmdname`		shows you the type of a command which is the executable location for programs
`whereis` `filename`	`whereis mysql`	shows location of all files with that name
`whoami`	`whoami`	shows the currently logged in user
`scp -r` `localfilename` `remoteuser`:`remotedir` scp -r elr37@*remote-host:/path/to/directory /local/path/to/where/you/want/to/put/it*	`scp -r elr37@elr37-dev.library.cornell.edu://cul/web/elr37-dev.library.cornell.edu/htdocs/rails/questioning_authority ~/questioning_authority_LD_BRANCH`	secure copy from local to remote machine	NOTE: -r is only needed for recursive copy of an entire directory
`echo $SHELL`	`echo $SHELL`	show which shell is running
`grep -r` `pattern` `path`	grep -r foo *	Starting at current directory, look for 'foo' in all files including subdirs.	NOTE: -r is only needed for recursive search through subdirs
`grep pattern path \| wc -l`	grep "<record>" bib.xml \| wc -1	Count the number of matches for grep.	In the example, it will count the number of times that <record> appears in file bib.xml.

Working with Services

Action	Command	Comments
start	`/etc/inid.d/service_name start`	ex. sudo /etc/init.d/tomcat-sf-lib-app-018.serverfarm.cornell.edu start
stop	`/etc/inid.d/service_name stop`
restart	`/etc/inid.d/service_name restart`
status	`/etc/inid.d/service_name status`

OR

Action	Command	Comments
start	sudo service `service_name` start	ex. sudo service mysqld start
stop	sudo service `service_name` stop
restart	sudo service `service_name` restart
status	sudo service `service_name` status

Find Command

(reference)

find where-to-look criteria what-to-do

# prints names of files that directly match foo  (e.g. /usr/foo, /usr/bin/foo)
find / -name foo

# prints names of files that contain foo (e.g. /usr/myfoo, /usr/bin/foobar)
find / -name foo

# suppress permission denied statements
find /. -name *foo* 2>/dev/null

Tree command

$ tree
.
├── contrib
│   ├── analysis-extras
│   │   ├── README.txt
│   │   ├── lib
│   │   │   ├── icu4j-49.1.jar
│   │   │   ├── morfologik-fsa-1.5.5.jar
│   │   │   ├── morfologik-polish-1.5.5.jar
│   │   │   └── morfologik-stemming-1.5.5.jar
│   │   └── lucene-libs
│   │       ├── lucene-analyzers-icu-4.3.0.jar
│   │       ├── lucene-analyzers-morfologik-4.3.0.jar
│   │       ├── lucene-analyzers-smartcn-4.3.0.jar
│   │       └── lucene-analyzers-stempel-4.3.0.jar
│   └── extraction
│       └── lib
└── solr-analysis-extras-4.3.0.jar

NOTE: If you are on the mac and it says tree is not a command, run the following to install.

brew install tree

Create all directories on a path

# creates directories /my    /my/path    /my/path/here   if they don't already exist
mkdir -p /my/path/here/

pgrep, pidof, top, ps, kill

Reference: http://www.cyberciti.biz/faq/show-all-running-processes-in-linux/

pgrep mysqld                lists PID of all running processes that have the pattern (mysqld in the example) in the COMMAND; matches mysqld and mysqld_safe
pidof mysqld                lists PID of the specific process (mysqld in the example); matches mysqld only; will not match mysqld_safe

top                         shows top running processes (according to CPU usage); show enough to fill the size of the window; continues to update
q                           quit viewing top

ps                          show my processes
ps u                        show userid that started each process
ps a                        show processes for all users
ps au                       show processes for all users and userid that started each process
ps aux                      show processes for all users and userid that started each process and x adds processes started on different machines (tty) including services  

ps -U root -u root          show all processes started under root user
ps -U root -u root -N       show all processes started by users other than root
ps u -U root -u root -N     show all processes started by users other than root and userid that started each process

# see all the processes I started
ps -U elr37

Reference: http://www.cyberciti.biz/faq/kill-process-in-linux-or-terminate-a-process-in-unix-or-linux-systems/

Be very very careful what you kill.

$ pidof mysqld
4135
$ kill  4135

You will have to use sudo if you didn't start the process.

Resuming a paused job after <ctrl><z>

$ jobs
[1]+  Stopped                 vi test
$ fg 1

The #1 is identifying the job to resume in the case where there is more than one job paused. If there is only one paused job, then the number is optional.

Ref: how-can-i-resume-a-stopped-job-in-linux

Symbolic Links

ln -s {real-filename} {symbolic-filename}

Compressing and Extracting tar.gz and zip files

# extract to current directory
tar -zxvf yourfile.tar.gz

# extract to specified directory
tar -C /myfolder -zxvf yourfile.tar.gz

# zip file
zip filename.zip file1 file2 file3

# zip directory
zip -r filename.zip /myfolder

# unzip
unzip filename.zip

recursive diff

diff -bur folder1/ folder2/

tail

# show last few lines and exit
tail /path/to/file

# show last few lines as they update
tail -f /path/to/file

# show last n lines as they update
tail -n 60 -f /path/to/file

du

# how much space is used in the current directory recursively
du -h . 2>/dev/null

This will list all directories and subdirectories and show how many kB, MB, GB are used by each dir and it's subdirs.

# how much space is used in the current directory summed
du --max-depth=1 -c -h 2>/dev/null

This will list all directories, but not subdirs. Totals will include space used by each directory + their subdirs.

Debugging full memory (RAM)

List processes that are using the most memory...

$ top
<shift><m>

NOTE: <shift><m> sorts the top results by memory usage (i.e., %MEM)

$ free -m
             total       used       free     shared    buffers     cached
Mem:         16041       1298      14742          0         36        131
-/+ buffers/cache:       1130      14911
Swap:            0          0          0

Debugging full memory (disk space)