Vendix EC2 Maintenance & Troubleshooting

Description

Guidelines for troubleshooting and maintaining EC2 instances, specifically for deployment failures related to disk space or Docker issues.

Trigger

• •"no space left on device" error in deployment logs.
• •"failed to register layer" error in Docker build/pull.
• •Deployment hangs or fails mysteriously on EC2.
• •User asks to check or clean the server.

Protocols

1. 🛑 Security First

• •NEVER assume the location or name of the SSH private key (.pem).
• •ALWAYS ask the user to provide the path to the SSH key for the specific environment (dev/prod).
• •DO NOT search for keys in .ssh/ automatically unless explicitly instructed by the user in the current session.
• •Example: "Para proceder, necesito que me indiques la ruta de la llave .pem para conectarme al servidor."

2. 🔍 Diagnosis

Connect to the server and check disk usage:

ssh -i <USER_PROVIDED_KEY> ec2-user@<SERVER_IP> "df -h && docker system df"

3. 🧹 Cleanup (Disk Space)

If disk usage is high (>85%) or Docker cache is large:

Prune Unused Docker Objects:

docker system prune -a -f

Note: This removes all stopped containers and unused images. It is usually safe for CI/CD servers that pull fresh images on deploy.

Check Specific Volumes: If docker prune isn't enough, check for large log files or orphan volumes:

du -sh /var/log/*
docker volume ls

4. 🔄 Common Issues

• •Docker Layer Registration Failed: Indicates disk full. Run cleanup immediately.
• •npm install fails: May indicate OOM (Out of Memory) or disk full. Check free -m and df -h.