Writing a Zero Downtime Deployment Script

The post describes building a shell script called FlipIt™ to deploy a site with zero downtime. It creates a timestamped deployment directory, clones the Git repository into it, and then switches the server’s public directory to the new release with a symbolic link.

It also shows how to turn those commands into a reusable script with arguments, run it locally, and have it execute on a remote server over SSH. The article notes that extra commands like composer install can run before the link switch, and it mentions possible future additions like rollback, init, and cleanup.

Reading notes

#

• The author writes the deployment flow as a shell script after becoming more comfortable with terminal commands.
• Zero downtime comes from keeping the previous site active while the new version is prepared.
• The deployment directory uses Unix time so each release folder has a unique, ordered name.
• The repository is cloned into the new deployment directory instead of using git pull on the live copy.
• The public server directory is linked to the deployed release with ln -s -n -f.
• The script accepts arguments for the server public directory, repository URL, and repository public directory.
• A shebang line identifies the file as a shell script.
• The script can be wrapped in an SSH heredoc so it runs on a remote server.
• The post says this setup works for the author’s own environment but may differ on other servers.
• The notes mention that old deploys are not cleaned up automatically.
• The update recommends git clone –depth 1 to download only the latest version of the files.