Skip to content

Node.js passthrough HTTP mirror with caching for large files.

Notifications You must be signed in to change notification settings

oxygen/CachedPassthroughMirror

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

67 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

FOSSA Status

Cached passthrough HTTP mirror

Very simple HTTP passthrough mirror (proxy) with an aggresive cache for large files (ignores cache headers and caches everything larger than a specified minimum size). Originally written to cache large files to accelerate some repetitive operations.

Writes to the cache and serves the large files at the same time (stream-copy).

Uses http-proxy for the HTTP proxy part.

This cache saves bandwidth (or speeds up transfers) and makes some files highly available locally (prefetch).

It replicates the directory structure of the target server, only for the cached files (which have at least nBytesMinimumFileSize).

The Content-length and Last-modified headers (from a HEAD request to the target server) are used to determine if the cached file is to be invalided. The HEAD request is always made and so is depended upon (to proxy updated headers and for immediate cache invalidation).

To prevent cache invalidation (deletion of files which respond with 404 on HTTP or updating of files which changes) place an empty file next to the file that needs to persist, sufixed with .keep, like this: [original path name].keep.

There are no plans to add non-ASCII characters support or saving of headers, or 100% independent mirror capabilities.

Place a file containing whitespace separated file paths (URL encoded), named cache_prefetch.txt, in the root of the target URL base path. cache_prefetch.txt is ignored unless you periodically call the .sync() method of the HTTPProxyCache class to prefetch or update the list of files.

Security WARNING: HTTP authorization skip: when the HEAD request fails either with a non-200 HTTP status code or at network level (or something else), and the file is served directly from the cache storage, there will be no HTTP authorization. This might be fixed in the future, but at the present time, it presents a risk where security matters.

@TODO: write examples, CLI endpoint, etc.

Usage

	const HTTPProxyCache = require("http-proxy-cache-lf");
	const http = require("http");

	const httpServer = http.createServer();

	const httpProxyCache = new HTTPProxyCache(
		/*strTargetURLBasePath*/ "http://kakao.go.ro/", 
		/*nBytesMinimumFileSize*/ 4 * 1024 * 1024 /*4 MB*/, 
		/*strCacheDirectoryRootPath*/ "/tmp/repo-cache"
	);

	httpServer.on(
		"request",
		async (incomingMessage, serverResponse) => {
			await httpProxyCache.processHTTPRequest(incomingMessage, serverResponse);
		}
	);

	httpServer.listen(8008, "127.0.0.1");

License

FOSSA Status

About

Node.js passthrough HTTP mirror with caching for large files.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published