Deployment troubleshooting: port conflicts, config automation, and safe validation

Two recurring deployment failures are stale listeners on app ports and unsafe manual config edits. This is the production workflow I use to solve both.

Incident pattern

Error: listen EADDRINUSE: address already in use :::3000

Diagnose the listener first

ss -lntp | grep ':3000'
lsof -i :3000
ps -fp <PID>

Controlled fix

kill -15 <PID>
sleep 2
kill -9 <PID> 2>/dev/null || true

fuser -k 3000/tcp
ss -lntp | grep ':3000' || echo 'port free'

Prevent recurrence

Use pm2 or systemd instead of background shell starts.

pm2 start ecosystem.config.js --env production
pm2 save
pm2 startup

Safe config automation

Wrong quoting:

sed -i 's/$OLD/$NEW/g' wp-config.php

Correct quoting:

sed -i "s/$OLD/$NEW/g" wp-config.php

With dry-run:

OLD_VAL=$(grep DB_NAME wp-config.php | cut -d"'" -f4 | cut -d_ -f1)
NEW_VAL=$(pwd | cut -d/ -f3)

sed "s/$OLD_VAL/$NEW_VAL/g" wp-config.php | head -n 20
sed -i "s/$OLD_VAL/$NEW_VAL/g" wp-config.php

Safety checklist

cp wp-config.php wp-config.php.bak.$(date +%F-%H%M)
php -l wp-config.php

Rollback

cp wp-config.php.bak.YYYY-MM-DD-HHMM wp-config.php
systemctl restart php-fpm

Final takeaway

Stable deployment comes from repeatable operations: precise port diagnosis, controlled process cleanup, validated automation, and immediate rollback.

This post is licensed under CC BY-NC.