15.1 Common Linux Commands¶

This chapter summarizes common Linux commands for daily work on the computing cluster. In command examples, $ represents a normal user shell prompt. Do not copy $ when copying commands.

Warning

Login nodes are mainly used for logging in, editing files, compiling small programs, submitting jobs, and checking results. Do not run long-running, high-CPU, high-memory, or heavy I/O computing tasks on login nodes. Submit these tasks to the HTCondor or Slurm job systems.

1. Get Current Environment Information¶

Task	Command Example
Show current user	`whoami`
Show user ID and groups	`id`
Show user groups	`groups`
Show current host name	`hostname`
Show current directory	`pwd`
Show current time	`date`
Show system uptime and load	`uptime`
Show kernel and architecture	`uname -a`
Show operating system version	`cat /etc/os-release`

Examples:

$ whoami
$ id
$ hostname
$ pwd
$ uptime
$ uname -a
$ cat /etc/os-release

2. Directory and Path Operations¶

Task	Command Example
Change to a directory	`cd /path/to/dir`
Return to home directory	`cd` or `cd ~`
Go to the parent directory	`cd ..`
Go back to the previous working directory	`cd -`
Show the absolute path of the current directory	`pwd`
Create a directory	`mkdir dirname`
Create nested directories	`mkdir -p project/logs`
View a directory tree	`tree`

If tree is not installed, use find to view directory levels:

$ find . -maxdepth 2 -type d

3. View Files and Directories¶

Task	Command Example
List files in the current directory	`ls`
Show detailed information	`ls -l`
Show hidden files	`ls -a`
Sort by time, newest first	`ls -lt`
Sort by file size	`ls -lhS`
Show human-readable sizes	`ls -lh`
Show file type	`file filename`
Show the first lines of a file	`head filename`
Show the last lines of a file	`tail filename`
Follow log output in real time	`tail -f job.out`
Page through a file	`less filename`
Print a file directly	`cat filename`

Common combinations:

$ ls -lah
$ ls -lt | head
$ tail -n 100 job.out
$ tail -f slurm-12345.out
$ less large.log

4. Create, Copy, Move, and Remove Files or Directories¶

Task	Command Example
Create an empty file	`touch file.txt`
Copy a file	`cp source.txt target.txt`
Copy a directory	`cp -r source_dir target_dir`
Preserve attributes while copying	`cp -a source_dir target_dir`
Move a file or directory	`mv old_path new_path`
Rename a file	`mv old.txt new.txt`
Remove a file	`rm file.txt`
Remove a file interactively	`rm -i file.txt`
Remove an empty directory	`rmdir empty_dir`
Remove a non-empty directory	`rm -r dir`

Warning

Files removed with rm usually cannot be restored directly. Before running rm -r or rm -rf, carefully confirm the path and variable expansion. Do not use recursive deletion when a variable may be empty or a path is uncertain.

Safer deletion examples:

$ rm -i result.txt
$ rm -r old_output/

5. Search Files and Text¶

Task	Command Example
Find files by name	`find . -name "*.txt"`
Find directories	`find . -type d -name "logs"`
Find regular files	`find . -type f`
Find large files	`find . -type f -size +1G`
Find recently modified files	`find . -type f -mtime -1`
Search text in a file	`grep "keyword" file.txt`
Search text recursively	`grep -R "keyword" .`
Show matching line numbers	`grep -n "keyword" file.txt`
Search case-insensitively	`grep -i "keyword" file.txt`
Count matching lines	`grep -c "keyword" file.txt`

If rg is installed, prefer it for faster recursive searches:

$ rg "keyword"
$ rg -n "error|failed" logs/
$ rg --files

Common examples:

# Find all .out files under the current directory
$ find . -type f -name "*.out"

# Find files larger than 1 GB
$ find . -type f -size +1G -lh

# Search logs for errors
$ grep -Rni "error" logs/

6. Check File Size, Disk Usage, and Quotas¶

Task	Command Example
Show file system usage	`df -h`
Show total size of current directory	`du -sh .`
Show sizes of subdirectories	`du -h --max-depth=1`
Sort directories by size	`du -h --max-depth=1 \| sort -h`
Sort files by size	`ls -lhS`
Check AFS quota	`fs listquota /afs/ihep.ac.cn/users/...`
Check Lustre user quota	`lfs quota -u username /scratchfs`
Check Lustre file system space	`lfs df -h /scratchfs`

Examples:

$ df -h
$ du -sh *
$ du -h --max-depth=1 | sort -h
$ fs listquota /afs/ihep.ac.cn/users/a/username
$ lfs quota -u username /scratchfs
$ lfs quota -u username /workfs

When disk space is insufficient, first locate large directories and files:

$ du -h --max-depth=1 ~ | sort -h
$ find . -type f -size +500M -lh

7. File Permissions and ACLs¶

Task	Command Example
View permissions	`ls -l file.sh`
Add execute permission	`chmod +x file.sh`
Set file permissions to owner read/write only	`chmod 600 file.txt`
Set directory permissions to owner access only	`chmod 700 private_dir`
Set file permissions to owner read/write and others read-only	`chmod 644 file.txt`
Set common directory access permissions	`chmod 755 dir`
View ACL permissions	`getfacl file_or_dir`
Set ACL permissions	`setfacl -m u:username:rwx dir`

Common numeric permissions:

Permission	Meaning
`7`	read, write, and execute
`6`	read and write
`5`	read and execute
`4`	read only
`0`	no permission

Examples:

$ chmod +x run.sh
$ chmod 600 id_rsa
$ chmod 700 ~/.ssh
$ chmod 644 ~/.ssh/authorized_keys
$ getfacl /workfs/group/project

8. Text Editing and Text Processing¶

Task	Command Example
Edit with vim	`vim file.txt`
Edit with nano	`nano file.txt`
Count lines, words, and bytes	`wc file.txt`
Count lines only	`wc -l file.txt`
Sort lines	`sort file.txt`
Remove adjacent duplicate lines	`uniq file.txt`
Sort and remove duplicates	`sort file.txt \| uniq`
Extract columns	`awk '{print $1}' file.txt`
Replace text	`sed 's/old/new/g' file.txt`
Compare two files	`diff old.txt new.txt`
Inspect binary or hidden characters	`od -c file.txt`

Examples:

# Show the first 20 lines
$ head -n 20 data.txt

# Show lines 10 to 20
$ sed -n '10,20p' data.txt

# Extract the first column and remove duplicates
$ awk '{print $1}' data.txt | sort | uniq

# Count occurrences of error in a log
$ grep -i "error" job.log | wc -l

9. Compression and Extraction¶

File Type	Extraction Command
`.tar`	`tar -xvf file.tar`
`.tar.gz` or `.tgz`	`tar -zxvf file.tar.gz`
`.tar.bz2`	`tar -jxvf file.tar.bz2`
`.tar.xz`	`tar -Jxvf file.tar.xz`
`.gz`	`gunzip file.gz`
`.zip`	`unzip file.zip`

Create archives:

# Archive a directory without compression
$ tar -cvf project.tar project/

# Create a gzip-compressed archive
$ tar -zcvf project.tar.gz project/

# Exclude selected files or directories
$ tar --exclude="*.tmp" --exclude="output/" -zcvf project.tar.gz project/

# Create a zip archive of a directory
$ zip -r project.zip project/

View archive contents:

$ tar -tf project.tar
$ tar -ztf project.tar.gz
$ unzip -l project.zip

Note

Large numbers of small files can affect shared file system performance. Before backup or transfer, consider packaging them with tar.

10. File Transfer¶

Task	Command Example
Copy from local machine to cluster	`scp local.txt username@lxlogin.ihep.ac.cn:/path/`
Copy from cluster to local machine	`scp username@lxlogin.ihep.ac.cn:/path/file.txt ./`
Copy a directory	`scp -r local_dir username@lxlogin.ihep.ac.cn:/path/`
Resume or synchronize a directory	`rsync -avP source/ target/`
Use sftp	`sftp username@lxlogin.ihep.ac.cn`
Download a file from a URL	`wget URL`
Download a file with curl	`curl -O URL`

Examples:

# Upload a file
$ scp input.dat username@lxlogin.ihep.ac.cn:/workfs/group/username/

# Download results
$ scp username@lxlogin.ihep.ac.cn:/workfs/group/username/result.out ./

# Synchronize a directory, useful for large directories or unstable networks
$ rsync -avP ./project/ username@lxlogin.ihep.ac.cn:/workfs/group/username/project/

# Preview what rsync would synchronize
$ rsync -avP --dry-run ./project/ username@lxlogin.ihep.ac.cn:/workfs/group/username/project/

11. View and Manage Processes¶

Task	Command Example
Show processes in current terminal	`ps`
Show all processes	`ps -ef`
Show your own processes	`ps -u username`
Search processes by keyword	`pgrep -af python`
View processes dynamically	`top`
View processes dynamically with htop	`htop`
Terminate a process	`kill PID`
Force-terminate a process	`kill -9 PID`
Run a command in the background	`command &`
Show background jobs	`jobs`
Bring a background job to the foreground	`fg %1`
Continue a stopped job in the background	`bg %1`

Examples:

$ ps -u $USER
$ pgrep -af python
$ kill 12345
$ top -u $USER

Warning

Do not use kill -9 as the default option. A normal kill PID gives the program a chance to exit cleanly. Use kill -9 PID only when the process cannot terminate normally.

12. View System Resources¶

Task	Command Example
View CPU and memory usage	`top`
View memory usage	`free -h`
View process, memory, and system load	`vmstat 1`
View disk I/O	`iostat -x 1`
View GPU status	`nvidia-smi`
Refresh GPU status continuously	`watch -n 1 nvidia-smi`
Show logged-in users	`who`
Show current terminal and load information	`w`

Examples:

$ free -h
$ top
$ nvidia-smi
$ watch -n 1 nvidia-smi

If iostat or htop is unavailable, the corresponding tool is not installed on the current system. Use basic commands such as top and free instead.

13. Network and Connection Checks¶

Task	Command Example
Test host reachability	`ping hostname`
Show network route path	`traceroute hostname`
Test whether a port is reachable	`nc -vz hostname port`
Show network addresses	`ip addr`
Show routes	`ip route`
Show listening ports	`ss -lntp`
Show all TCP connections	`ss -ant`

Examples:

$ ping lxlogin.ihep.ac.cn
$ nc -vz lxlogin.ihep.ac.cn 22
$ ss -lntp

14. SSH and Remote Sessions¶

Task	Command Example
Log in with SSH	`ssh username@lxlogin.ihep.ac.cn`
Log in with a specified port	`ssh -p 2222 username@hostname`
Log in with a private key	`ssh -i ~/.ssh/id_rsa username@hostname`
Enable X11 forwarding	`ssh -XYC username@hostname`
Generate an SSH key	`ssh-keygen -t ed25519`
Copy a public key to a remote host	`ssh-copy-id username@hostname`

Examples:

$ ssh username@lxlogin.ihep.ac.cn
$ ssh -XYC username@lxlogin.ihep.ac.cn
$ ssh-keygen -t ed25519

For long interactive sessions, use tmux or screen to avoid losing the session when the network disconnects:

# Create a new tmux session
$ tmux new -s work

# Reattach to an existing tmux session
$ tmux attach -t work

# List sessions
$ tmux ls

Note

tmux and screen are suitable for editing, compiling, and small interactive tests. Formal computing tasks should still be submitted to the job system.

15. Software Environment and Command Lookup¶

Task	Command Example
Show environment variables	`env`
Show one environment variable	`echo $PATH`
Set an environment variable temporarily	`export VAR=value`
Locate a command	`which python`
Show how a command is resolved	`type module`
Show dynamic library dependencies	`ldd program`
Show current shell	`echo $SHELL`
Show command help	`command --help`
Show manual page	`man command`

The cluster commonly uses module to manage software environments:

# List available modules
$ module avail

# Load a module
$ module load python/3.7

# List loaded modules
$ module list

# Unload a module
$ module unload python/3.7

# Clear all loaded modules
$ module purge

# Show module help
$ module help python/3.7

Locate programs and dynamic library issues:

$ which python
$ python --version
$ ldd ./my_program
$ echo $LD_LIBRARY_PATH

16. AFS and Kerberos Authentication¶

When reading or writing files under AFS directories, a valid token is often required.

Task	Command Example
Obtain a Kerberos ticket	`kinit username`
Obtain an AFS token	`aklog`
View Kerberos tickets	`klist`
View AFS tokens	`tokens`
Remove AFS tokens	`unlog`
Change Kerberos password	`kpasswd`

Examples:

$ kinit username
$ aklog
$ klist
$ tokens

If you cannot write files under /afs, check these first:

$ tokens
$ fs listquota /afs/ihep.ac.cn/users/a/username
$ ls -ld /afs/ihep.ac.cn/users/a/username

17. Shell History, Aliases, and Scripts¶

Task	Command Example
View command history	`history`
Search command history	`history \| grep keyword`
View current aliases	`alias`
Set a temporary alias	`alias ll='ls -lh'`
Remove an alias	`unalias ll`
Trace script execution	`bash -x script.sh`
Check script syntax	`bash -n script.sh`

Common script header:

#!/bin/bash
set -e
set -u

Notes:

set -e: exit the script when a command fails.
set -u: exit the script when an undefined variable is used.

18. Input/Output Redirection and Pipes¶

Task	Command Example
Write standard output to a file and overwrite it	`command > output.txt`
Append standard output to a file	`command >> output.txt`
Write standard error to a file	`command 2> error.txt`
Write standard output and error to the same file	`command > all.log 2>&1`
Pipe output to another command	`command1 \| command2`
Show output and save it to a file	`command \| tee output.txt`
Append output while showing it	`command \| tee -a output.txt`

Examples:

$ ./run.sh > run.log 2>&1
$ grep -i error run.log | tail
$ ./run.sh | tee run.log

19. Common Log Troubleshooting Commands¶

Task	Command Example
View the end of a log	`tail -n 100 job.out`
Follow a log in real time	`tail -f job.out`
Search for errors	`grep -i "error" job.out`
Search for failure messages	`grep -Ei "error\|failed\|fatal\|killed" job.out`
View output file sizes	`ls -lh job.out`
View recently generated files	`ls -lt`
View the exit code of the previous command	`echo $?`

Examples:

$ tail -n 100 slurm-12345.out
$ grep -Ei "error|failed|fatal|killed|oom" slurm-12345.out
$ ls -lt | head

Common keywords:

Keyword	Possible Meaning
`No such file or directory`	The path is wrong or the file does not exist
`Permission denied`	Permission is insufficient or the script is not executable
`command not found`	The command is not loaded, or the command path is not in `PATH`
`Killed`	The process may have been terminated because it used too much memory
`out of memory` / `OOM`	Insufficient memory
`Disk quota exceeded`	Storage quota exceeded

20. Common Checks Before and After Jobs¶

Before submitting a job, check:

# Which directory am I in?
$ pwd

# Does the input file exist?
$ ls -lh input.dat

# Is the script executable?
$ ls -l run.sh
$ chmod +x run.sh

# Is the software environment correct?
$ module list
$ which python
$ python --version

# Is there enough storage space?
$ du -sh .
$ df -h .

After a job finishes, check:

# View output files
$ ls -lh
$ tail -n 100 job.out

# Search for errors
$ grep -Ei "error|failed|fatal|killed|oom" job.out

# Check output directory size
$ du -sh output/

21. Get Help¶

Task	Command Example
View brief command help	`command --help`
View the manual page	`man command`
Show command path	`which command`
View help for a shell built-in command	`help cd`
View module help	`module help module_name`

Examples:

$ ls --help
$ man find
$ help cd
$ module help python/3.7

🤖 AI 助手