This guide covers essential bash commands and concepts for new users of the Ruqola server.
# Print current directory
pwd
# List files and directories
ls # basic listing
ls -la # detailed listing with hidden files
ls -lh # human-readable file sizes
# Change directory
cd /path/to/directory
cd ~ # go to home directory
cd .. # go up one directory
cd - # go to previous directory
# Create and remove
mkdir myproject # create directory
mkdir -p deep/nested/dirs # create nested directories
rmdir myproject # remove empty directory
rm file.txt # remove file
rm -rf directory/ # remove directory and all contents (BE CAREFUL!)
# Copy and move
cp file.txt backup.txt # copy file
cp -r directory/ backup/ # copy directory recursively
mv oldname.txt newname.txt # rename/move file
mv file.txt /path/to/directory/ # move file to directory
# View file contents
cat file.txt # display entire file
head file.txt # first 10 lines
tail file.txt # last 10 lines
tail -f logfile.txt # follow file changes (useful for logs)
less file.txt # paginated view (press q to quit)
# File information
file somefile # determine file type
du -sh directory/ # directory size
df -h # disk space usage
# Find files by name
find /path/to/search -name "*.py"
find . -name "train.py"
# Find files by type
find . -type f -name "*.log" # files only
find . -type d -name "models" # directories only
# Find recent files
find . -mtime -1 # files modified in last day
find . -mtime +7 # files older than 7 days
# Locate command (faster, uses database)
locate filename
updatedb # update locate database (as admin)
# View permissions
ls -l file.txt
# Change permissions
chmod +x script.py # make executable
chmod 644 file.txt # rw-r--r--
chmod 755 script.py # rwxr-xr-x
chmod -R 755 directory/ # recursive
# Change ownership (usually requires sudo)
chown user:group file.txt
chown -R user:group directory/
# Quick text editors
nano file.txt # beginner-friendly editor
vim file.txt # powerful editor (learning curve)
# Text processing
grep "pattern" file.txt # search for pattern
grep -r "function" . # recursive search in current directory
grep -i "error" logfile.txt # case-insensitive search
grep -n "TODO" *.py # show line numbers
# Sorting and filtering
sort file.txt # sort lines
sort -n numbers.txt # numeric sort
uniq file.txt # remove duplicate lines
sort file.txt | uniq # sort and remove duplicates
# Counting
wc file.txt # word count (lines, words, characters)
wc -l file.txt # line count only
# Cut and paste
cut -d',' -f1 data.csv # extract first column from CSV
cut -c1-10 file.txt # extract characters 1-10
# Sed (stream editor)
sed 's/old/new/g' file.txt # replace all occurrences
sed -i 's/old/new/g' file.txt # edit file in-place
# Awk (pattern processing)
awk '{print $1}' file.txt # print first column
awk -F',' '{print $2}' data.csv # use comma as delimiter
# Run commands
command # run in foreground
command & # run in background
nohup command & # run in background, ignore hangup
# Process monitoring
ps aux # list all processes
ps aux | grep python # find python processes
top # real-time process monitor
htop # better version of top (if available)
# Process control
jobs # list background jobs
fg %1 # bring job 1 to foreground
bg %1 # send job 1 to background
kill PID # terminate process by PID
kill -9 PID # force kill process
killall python # kill all python processes
# Screen
screen # start new screen session
screen -S session_name # start named session
screen -ls # list sessions
screen -r session_name # attach to session
# Inside screen: Ctrl+A, D to detach
# Tmux (if available)
tmux # start new session
tmux new -s session_name # start named session
tmux ls # list sessions
tmux attach -t session_name # attach to session
# Inside tmux: Ctrl+B, D to detach
# View environment variables
env # all environment variables
echo $HOME # home directory
echo $USER # current user
echo $PATH # executable search path
# Set variables
export MY_VAR="value" # set environment variable
export PATH=$PATH:/new/path # add to PATH
# Persistent variables (add to ~/.bashrc or ~/.bash_profile)
echo 'export CUDA_VISIBLE_DEVICES=0' >> ~/.bashrc
source ~/.bashrc # reload configuration
# Important paths
echo $HOME # /home/username
echo $PWD # current directory
echo $OLDPWD # previous directory
# CUDA-related (for GPU computing)
export CUDA_VISIBLE_DEVICES=0,1 # use GPUs 0 and 1
export CUDA_DEVICE_ORDER=PCI_BUS_ID
# Connect to server
ssh username@server.com
ssh -p 2222 username@server.com # custom port
# SSH with key
ssh-keygen -t rsa # generate SSH key pair
ssh-copy-id username@server.com # copy public key to server
# File transfer
scp file.txt username@server.com:/path/to/destination/
scp -r directory/ username@server.com:/path/to/destination/
rsync -avz directory/ username@server.com:/path/to/destination/
# Forward remote port to local
ssh -L 8888:localhost:8888 username@server.com
# Background tunnel
ssh -f -N -L 8888:localhost:8888 username@server.com
# Navigation
Ctrl+A # beginning of line
Ctrl+E # end of line
Ctrl+U # clear line before cursor
Ctrl+K # clear line after cursor
# History
Ctrl+R # search command history
!! # repeat last command
!string # repeat last command starting with 'string'
history # show command history
# Completion
Tab # auto-complete
Tab Tab # show all possibilities
Add to ~/.bashrc for convenience:
# Useful aliases
alias ll='ls -la'
alias la='ls -A'
alias l='ls -CF'
alias ..='cd ..'
alias ...='cd ../..'
alias grep='grep --color=auto'
# GPU-related aliases
alias gpu='nvidia-smi'
alias gpuwatch='watch -n 1 nvidia-smi'
alias q='gpuq status'
# Quick functions
mkcd() { mkdir -p "$1" && cd "$1"; } # make directory and cd into it
# NVIDIA commands
nvidia-smi # GPU status
nvidia-smi -l 1 # continuous monitoring (1-second intervals)
nvidia-smi -q # detailed GPU info
# Our custom queue system
gpuq status # check queue status
gpuq submit --command "python train.py" --gpus 1
gpuq kill --job-id 12345 # kill specific job
# Hardware info
lscpu # CPU information
lsmem # memory information
lspci | grep -i nvidia # GPU information
df -h # disk usage
free -h # memory usage
uptime # system uptime and load
# Ubuntu/Debian
apt update # update package list
apt install package-name # install package
apt search keyword # search packages
# Python packages
pip install package-name # install Python package
pip install -r requirements.txt # install from requirements file
pip list # list installed packages
pip show package-name # show package info
experiment_2024_transformer.py not exp1.pyls before rmgit commandsgpuq status before submitting jobs# Most used commands for server work
pwd # where am I?
ls -la # what's here?
cd directory # go somewhere
cp file.txt backup.txt # make a copy
nvidia-smi # check GPUs
gpuq status # check queue
gpuq submit --command "python train.py" --gpus 1
top # what's running?
kill PID # stop a process
Next Steps: Once comfortable with these basics, check out the GPU Queue Guide to start running your first jobs on the server.