Starlet - a simple, high-performance PSGI/Plack HTTP server
% start_server --port=80 -- plackup -s Starlet [options] your-app.psgi or if you do not need hot deploy, % plackup -s Starlet --port=80 [options] your-app.psgi
Starlet is a standalone HTTP/1.0 server with keep-alive support, formerly known as Plack::Server::Standalone::Prefork and Plack::Server::Standalone::Prefork::Server::Starter.
The server supports following features, and is suitable for running HTTP application servers behind a reverse proxy.
- prefork and graceful shutdown using Parallel::Prefork
- hot deploy using Server::Starter
- fast HTTP processing using HTTP::Parser::XS (optional)
In addition to the options supported by plackup, Starlet accepts following options(s).
number of worker processes (default: 10)
seconds until timeout (default: 300)
timeout for persistent connections (default: 2)
max. number of requests allowed per single persistent connection. If set to one, persistent connections are disabled (default: 1)
max. number of requests to be handled before a worker process exits (default: 100)
if set, randomizes the number of requests handled by a single worker process between the value and that supplied by
--max-reqs-per-chlid (default: none)
if set, worker processes will not be spawned more than once than every given seconds. Also, when SIGHUP is being received, no more than one worker processes will be collected every given seconds. This feature is useful for doing a "slow-restart". See http://blog.kazuhooku.com/2011/04/web-serverstarter-parallelprefork.html for more information. (dedault: none)
Starlet is designed and implemented to be simple, secure and fast, especially for running as a HTTP application server running behind a reverse proxy. It only depends on a minimal number of well-designed (and well-focused) modules.
On the other hand if you are looking for a standalone preforking HTTP server that receives HTTP requests directly from the Internet, then you should look at Starman.