2022 Getting started at TACC

Getting to a remote computer

The Terminal window

• Macs and Linux have a Terminal program built-in – find it now on your computer
• Windows 10 or later has ssh and scp in Command Prompt or PowerShell (may require latest Windows updates)
  • Open the Start menu → Search for Command
    

From now on, when we refer to "Terminal", it is either the Mac/Linux Terminal program, Windows Command Prompt or PowerShell, or the PuTTY program.

SSH

ssh is an executable program that runs on your local computer and allows you to connect securely to a remote computer. We're going to use ssh to access the Lonestar6 compute cluster at TACC, where the remote host name is ls6.tacc.utexas.edu.

In your local Terminal window:

ssh <your_TACC_userID>@ls6.tacc.utexas.edu

# For example:
ssh abattenh@ls6.tacc.utexas.edu

• Answer yes to the SSH security question prompt
• Enter the password associated with your TACC account
• Wait for your 2-factor authentication code to arrive via SMS or app, then type it in

The bash shell

You're now at a command line! It looks as if you're running directly on the remote computer, but really there are two programs communicating:

1. your local Terminal
2. the remote Shell

There are many shell programs available in Linux, but the default is bash (Bourne-again shell).

The Terminal is pretty "dumb" – just sending what you type over its secure sockets layer (SSL) connection to TACC, then displaying the text sent back by the shell. The real work is being done on the remote computer, by programs called by the bash shell.

Setting up your environment

Create some symbolic links and directories

First we will create a few directories and links we will use (more on these later).

Create some symbolic links that will come in handy later:

cd  # makes your Home directory the "current directory"
ln -s -f $SCRATCH scratch
ln -s -f $WORK work
ln -s -f /work/projects/BioITeam/projects/courses/Core_NGS_Tools CoreNGS

Symbolic links (a.k.a. symlinks) are "pointers" to files or directories elsewhere in the file system hierarchy. You can almost always treat a symlink as if it is the actual file or directory.

echo $SCRATCH

ls -l

Set up a $HOME/local/bin directory and link several scripts there that we will use in the class.

mkdir -p ~/local/bin
cd ~/local/bin
ln -s -f /work/projects/BioITeam/common/bin/launcher_creator.py

Setup your login profile (~/.bashrc)

Now execute the lines below to set up a login script, called .bashrc

When you login via an interactive shell, a well-known script is executed to establish your favorite environment settings. We've set up a common login script for you to start with that will help you know where you are in the file system and make it easier to access some of our shared resources. To set it up, perform the steps below:

cd
cp .bashrc .bashrc.beforeNGS

cd
cp /work/projects/BioITeam/projects/courses/Core_NGS_Tools/tacc/bashrc.corengs.ls6 .bashrc
chmod 600 .bashrc

Since .bashrc is executed when you login, to ensure it is set up properly you should first log off ls6 like this:

exit

Then log back in to ls6.tacc.utexas.edu. This time your .bashrc will be executed and you should see a new shell prompt:

ls6:~$

The great thing about this prompt is that it always tells you where you are, which avoids you having to execute the pwd (present working directory) command every time you want to know where you are. Execute these commands to see how the prompt reflects your current directory.

mkdir -p ~/tmp/a/b/c
cd ~/tmp/a/b/c

# Your prompt should look like this:
ls6:~/tmp/a/b/c$ 

The prompt now tells you you are in the c sub-directory of the b sub-directory of the a sub-directory of the tmp sub-directory of your Home directory ( ~ ).

So why don't you see the .bashrc file you copied to your home directory? Because all files starting with a period (dot files) are hidden by default. To see them add the -a (all) option to ls:

cd
ls -a

To see even more detail, including file type and permissions and symbolic link targets, add the -l (long listing) option:

ls -la

Details about your login script

We list the contents of your .bashrc login script to the Terminal with the cat (concatenate files) command. cat simply reads a file and writes each line of content to standard output (here, your Terminal):

cd
cat .bashrc

# or for larger files...
more .bashrc

You'll see the following (you may need to scroll up a bit to see the beginning):

#!/bin/bash
# TACC startup script: ~/.bashrc version 2.1 -- 12/17/2013
#   This file is NOT automatically sourced for login shells.
# Your ~/.profile can and should "source" this file.
# Note neither ~/.profile nor ~/.bashrc are sourced automatically
# by bash scripts.
#   In a parallel mpi job, this file (~/.bashrc) is sourced on every
# node so it is important that actions here not tax the file system.
# Each nodes' environment during an MPI job has ENVIRONMENT set to
# "BATCH" and the prompt variable PS1 empty.
#################################################################
# Optional Startup Script tracking. Normally DBG_ECHO does nothing
if [ -n "$SHELL_STARTUP_DEBUG" ]; then DBG_ECHO "${DBG_INDENT}~/.bashrc{"; fi
##########
# SECTION 1 -- modules
if [ -z "$__BASHRC_SOURCED__" -a "$ENVIRONMENT" != BATCH ]; then
  export __BASHRC_SOURCED__=1
  module load launcher
fi
############
# SECTION 2 -- environment variables
if [ -z "$__PERSONAL_PATH__" ]; then
  export __PERSONAL_PATH__=1
  export PATH=.:$HOME/local/bin:$PATH
fi
# For better colors using a dark background terminal, un-comment this line:
#export LS_COLORS=$LS_COLORS:'di=1;33:fi=01:ln=01;36:'
# For better colors using a white background terminal, un-comment this line:
#export LS_COLORS=$LS_COLORS:'di=1;34:fi=01:ln=01;36:'
export LANG="C"  # avoid the annoying Perl locale warnings 
export BIWORK=/work/projects/BioITeam
export CORENGS=$BIWORK/projects/courses/Core_NGS_Tools
export BI=/corral-repl/utexas/BioITeam
export ALLOCATION=OTH21164        # For ls6        Group is G-824651
##export ALLOCATION=UT-2015-05-18 # For stampede2  Group is G-816696

##########
# SECTION 3 -- controlling the prompt
if [ -n "$PS1" ]; then PS1='ls6:\w$ '; fi
##########
# SECTION 4 -- Umask and aliases
#alias ls="ls --color=always"
alias ll="ls -la"
alias lah="ls -lah"
alias lc="wc -l"
alias hexdump='od -A x -t x1z -v'
umask 002
##########
# Optional Startup Script tracking
if [ -n "$SHELL_STARTUP_DEBUG" ]; then DBG_ECHO "${DBG_INDENT}}"; fi

So what does this login script do? A lot! Let's look at just a few of them.

the "she-bang"

The first line is the she-bang. Even though the expression is inside a shell comment (denoted by the # character), it tells the shell (bash) what program should execute this file – in this case, bash itself.

#!/bin/bash

environment variables

The login script also sets an environment variable $BIWORK to point to the shared directory /work/projects/BioITeam, and another environment variable $CORENGS to point to the specific sub-directory for our class.

export BIWORK=/work/projects/BioITeam
export CORENGS=$BIWORK/projects/courses/Core_NGS_Tools

Environment variables are like variables in a programming language like python or perl (in fact bash is a complete programming language). They have a name (like BIWORK above) and a value (the value of $BIWORK is the pathname /work/projects/BioITeam). Read more about environment variables here: More on environment variables.

shell completion

You can use these environment variables to shorten typing, for example, to look at the contents of the shared /work/projects/BioITeam directory as shown below, using the magic Tab key to perform shell completion.

# hit Tab once after typing $BIWORK/ to expand the environment variable
ls $BIWORK/

# now hit Tab twice to see the contents of the directory
ls /work/projects/BioITeam/

# type "pr" and hit Tab again
ls /work/projects/BioITeam/pr

# type "co" and hit Tab again
ls /work/projects/BioITeam/projects/co

# type "Co" and hit Tab again
ls /work/projects/BioITeam/projects/courses/Co

# your command line should now look like this
ls /work/projects/BioITeam/projects/courses/Core_NGS_Tools/

# now type "mi" and one Tab
ls /work/projects/BioITeam/projects/courses/Core_NGS_Tools/mi

# your command line should now look like this
ls /work/projects/BioITeam/projects/courses/Core_NGS_Tools/misc/

# now hit Tab once
# the shell expands as far as it can unambiguously,
# so your command line should look like this
ls /work/projects/BioITeam/projects/courses/Core_NGS_Tools/misc/small

# now hit Tab twice
# You should see 3 filenames, all starting with "small"
# small.bam  small.fq   small2.fq

# type a period (".") then hit Tab twice again
# You're narrowing down the choices -- you should see two filenames
ls /work/projects/BioITeam/projects/courses/Core_NGS_Tools/misc/small
# small.bam  small.fq

# finally, type "f" then hit Tab again. It should complete to this:
ls /work/projects/BioITeam/projects/courses/Core_NGS_Tools/misc/small.fq

extending the $PATH

When you type a command name the shell has to have some way of finding what program to run. The list of places (directories) where the shell looks is stored in the $PATH environment variable. You can see the entire list of locations by doing this:

echo $PATH

As you can see, there are a lot of locations on the $PATH. That's because when you load modules at TACC (such as the module load lines in the common login script), that makes the programs available to you by putting their installation directories on your $PATH. We'll learn more about modules later.

Here's how the common login script adds your $HOME/local/bin directory to the location list – recall that's where we linked several useful scripts – along with a special dot character ( . ) that means "here", or "whatever the current directory is". In the statement below, colon ( : ) separates directories in the list.

export PATH=.:$HOME/local/bin:$PATH

setting up the friendly command prompt

The complicated looking if statement in SECTION 3 of your .bashrc sets up a friendly shell prompt that shows the current working directory. This is done by setting the special PS1 environment variable and including a special \w directive that the shell knows means "current directory".

##########
# SECTION 3 -- controlling the prompt
if [ -n "$PS1" ]; then PS1='ls6:\w$ '; fi