README

This command (static-cache:clean) removes orphaned static cache files that remain on disk after Redis has forgotten about them — a common edge case in high-traffic Statamic sites using full static caching.

🧨 Why this is needed

Statamic’s FileCacher stores all cached URLs for a site in a single array, saved to Redis under a single key (statamic:static-cache:urls:<domain>). When multiple requests (e.g. page renders, cache warmers, or invalidations) happen concurrently, they may:

• Read the current cache index
• Modify it (add/remove one URL)
• Write the entire array back to Redis

Because this process is not atomic, changes from one process can silently overwrite others. This race condition causes certain pages to disappear from Redis, even though their static .html files remain on disk.

🔒 What about Statamic’s built-in locks?

Statamic’s static cache middleware does apply a per-URL file lock (via Symfony's LockFactory) during a web request. This prevents the same URL from being rendered by multiple requests at the same time.

However, it does not protect the shared Redis key that holds the list of all cached URLs. This Redis write is global and unguarded, which means:

• Concurrent requests to different URLs (/about, /products) are not locked against each other.
• Console commands (e.g. static:warm, invalidateUrl, or queue jobs) bypass middleware entirely, skipping locks.
• Any of these can cause unintentional overwrites of Redis state.

🔁 Statamic has mutex guard rails per page, but not for the global Redis index.

❗ The risk

• Redis says a page is not cached.
• But the static .html file still exists.
• The web server continues serving the stale page, bypassing Statamic completely.

This issue is most common during:

• Bulk imports that trigger repeated invalidations (e.g. EntrySaved events)
• Cache warming across many URLs in parallel
• Sites using query string variations (ignore_query_strings => false)

✅ What this command does

This command:

1. Uses Statamic’s public APIs (getUrls(), getFilePath(), etc.) to collect all known cache entries from Redis.
2. Walks the static cache directories to find all existing .html files.
3. Deletes any file that is no longer referenced in Redis.
4. Optionally removes empty parent directories — with safeguards to never delete above the static cache root.

💡 Benefits

• Keeps your static cache clean and trustworthy
• Ensures Redis remains the source of truth
• Fixes stale page issues without requiring upstream changes
• Safe to run on production — includes --dry-run mode