====== Connection ====== This page explains how to connect to the VU MIF High Performance Computing (HPC) system. If this is your first time using HPC, don't worry. The goal is simply to log in successfully. Once connected, you can continue with the [[en:hpc-new:getting_started|First Successful Job]] guide and run your first computation. ===== Before You Start ===== Before connecting, make sure that: * You have completed [[en:hpc-new:registration|Registration]] * Your HPC account has been activated * You know your HPC username * You have an SSH client installed ===== Quick Start ===== Most users can connect using a single command: ssh username@hpc.mif.vu.lt Replace ''username'' with your HPC username. If this command works, you're ready to continue with the rest of the guide. ===== What Happens When I Connect? ===== When you connect to HPC, you are placed on a ''login node''. Think of the login node as your workspace where you can: * Edit files * Transfer data * Prepare applications * Compile code * Submit jobs Your actual calculations are performed on separate ''compute nodes'' managed by the SLURM scheduler. Unlike a personal computer, computational workloads are not normally executed directly on the login node. Instead, jobs are submitted to SLURM, which schedules them on available compute nodes. Please use the login node for preparing work and submitting jobs, not for running computationally intensive calculations. ===== Authentication Methods ===== There are two common ways to connect: * **Kerberos authentication** - convenient for VU MIF users * **SSH key authentication** - recommended for frequent access, file transfers, and development tools ===== Kerberos Authentication ===== VU MIF users can authenticate using Kerberos credentials. ==== Check Existing Kerberos Credentials ==== Verify that you have a valid Kerberos ticket: klist If no valid ticket exists, create one: kinit You will be prompted for your VU MIF password. ==== Connect to HPC ==== After obtaining a valid Kerberos ticket: ssh username@hpc.mif.vu.lt or, if supported by your local environment: ssh hpc ===== SSH Key Authentication ===== SSH keys are recommended if you: * Connect frequently * Transfer large files * Use WinSCP * Use Visual Studio Code Remote SSH * Run automated workflows ==== Generate an SSH Key ==== Linux and macOS: ssh-keygen -t ed25519 Windows PowerShell: ssh-keygen -t ed25519 Accept the default location and optionally protect the key with a passphrase. ==== Configure SSH Access ==== Before using SSH keys for the first time: - Log in using Kerberos authentication. - Create the SSH directory if it does not already exist: mkdir -p ~/.ssh chmod 700 ~/.ssh - Add your public key to: ~/.ssh/authorized_keys - Protect the file: chmod 600 ~/.ssh/authorized_keys After that, you can connect using your SSH private key without entering your HPC password each time. ===== Connecting from Windows ===== Windows users can connect with: * Windows Terminal (recommended) * PowerShell * PuTTY * MobaXterm ==== Windows Terminal or PowerShell ==== Open Windows Terminal or PowerShell and run: ssh username@hpc.mif.vu.lt ==== PuTTY ==== If you prefer a graphical application: - Download PuTTY from https://www.putty.org/ - Configure: * Host Name: ''hpc.mif.vu.lt'' * Port: ''22'' * Connection Type: ''SSH'' - Click **Open** ===== Connecting from Linux ===== Open a terminal and run: ssh username@hpc.mif.vu.lt ===== Connecting from macOS ===== Open Terminal.app and run: ssh username@hpc.mif.vu.lt ===== Connecting with Visual Studio Code ===== Visual Studio Code can connect directly to HPC using the Remote SSH extension. ==== Install the Extension ==== Install: * Remote - SSH ==== Configure SSH ==== Add the following entry to your local SSH configuration: Host hpc HostName hpc.mif.vu.lt User username Replace ''username'' with your HPC username. ==== Connect ==== Open: * Remote Explorer * SSH Targets * hpc Visual Studio Code will open a remote session on the HPC login node. ===== Transferring Files ===== You will often need to move files between your computer and the HPC system. Choose whichever method is most convenient: * WinSCP (Windows, graphical) * SCP (command line) * Visual Studio Code Remote SSH * Any SFTP-compatible application Most Windows users find WinSCP the easiest option. ==== WinSCP ==== Download: https://winscp.net/ Connection settings: * File protocol: ''SFTP'' * Host name: ''hpc.mif.vu.lt'' * Port number: ''22'' * Username: your HPC username ==== Upload Files Using SCP ==== Upload a file: scp myfile.txt username@hpc.mif.vu.lt:~ Upload a directory: scp -r myproject username@hpc.mif.vu.lt:~ ==== Download Files Using SCP ==== Download a file: scp username@hpc.mif.vu.lt:~/myfile.txt . Download a directory: scp -r username@hpc.mif.vu.lt:~/myproject . ===== First Login ===== The first time you connect, your HPC environment is prepared automatically. Your SLURM account is created automatically after your first successful login. This process may take up to 5 minutes. During this time, job submission commands such as ''sbatch'' may return account-related errors. No action is required. Simply wait a few minutes and try again. ===== Verify Everything Works ===== After connecting, run: whoami hostname pwd These commands show: * Your username * Which login node you are connected to * Your current directory If all commands work, your HPC account is ready to use. ===== Verify SLURM Access ===== Once your SLURM account has been created: sinfo Example output: PARTITION AVAIL TIMELIMIT NODES STATE NODELIST main* up ... gpu up ... ===== Available Partitions ===== The cluster currently provides two main partitions: ^ Partition ^ Typical Usage ^ | ''main'' | General-purpose CPU workloads | | ''gpu'' | Machine learning, AI, and GPU-accelerated applications | Most users should start with the ''main'' partition. Only use the ''gpu'' partition when your application requires NVIDIA GPUs. ===== Verify Storage Access ===== List files in your home directory: ls -la Display the current location: pwd Shared storage is described in: * [[en:hpc-new:lustre|Shared File System (Lustre)]] ===== Common Problems ===== ==== Permission denied ==== Possible causes: * Incorrect username * Incorrect password * SSH key misconfiguration * Account has not been activated ==== Connection timed out ==== Possible causes: * No Internet connectivity * Firewall restrictions * Incorrect hostname Verify: ping hpc.mif.vu.lt ==== Host key verification failed ==== Remove the previous SSH host entry: ssh-keygen -R hpc.mif.vu.lt Then reconnect. ==== SLURM commands fail after first login ==== Wait up to 5 minutes for automatic SLURM account creation. ==== File transfer fails ==== Verify that: * The destination path exists * You have sufficient storage space * SSH access works correctly ===== Frequently Asked Questions ===== ==== Do I need Kerberos every time? ==== No. Once SSH keys are configured, you can connect directly using SSH key authentication. ==== Can I connect from home? ==== Yes. The HPC system is accessible over the Internet using SSH. ==== Can I use Visual Studio Code? ==== Yes. Use the Remote SSH extension. ==== Can I transfer files graphically? ==== Yes. Use WinSCP or any other SFTP-compatible application. ===== Next Step ===== Congratulations, you have successfully connected to the HPC system. Continue with: * [[en:hpc-new:getting_started|First Successful Job]] * [[en:hpc-new:lustre|Shared File System (Lustre)]] * [[en:hpc-new:slurm|Batch Job Execution (SLURM)]] In the next guide, you'll submit your first job and see how computations are executed on the cluster. ---- This guide was prepared and updated by the VU MIF HPC team with the assistance of Microsoft Copilot. All procedures and examples should be validated against the current HPC environment, software stack, and operational policies.