The Philosophy Behind unicorn
Application_Timeouts CONTRIBUTORS DESIGN FAQ HACKING ISSUES KNOWN_ISSUES LICENSE Links NEWS PHILOSOPHY README SIGNALS Sandbox TUNING unicorn_1 Unicorn Configurator HttpServer OobGC PrereadInput StreamInput TeeInput Util Worker The Philosophy Behind unicorn Being a server that only runs on Unix-like platforms, unicorn is strongly tied to the Unix philosophy of doing one thing and (hopefully) doing
Debugging memory leaks in Ruby
Sam Saffron Programming, Technology and the Art of Hacking At some point in the life of every Rails developer you are bound to hit a memory leak. It may be tiny amount of constant memory growth, or a spurt of growth that hits you on the job queue when certain jobs run. Sadly, most Ruby devs out there simply employ monit , inspeqtor or unicorn worker killers. This allows you to move along and do mo
workerã®æž¯æ¸‡çŠ¶æ³ãŒåˆ†ã‹ã‚‹rack-server_statusã¨ã„ã†gemを書ã„ã¦ã„ã¾ã™ - CubicLouve
unicornã®wokerã®busyæ•°ã¨idleæ•°ã®æ¯”率ãŒã‚ã‹ã‚‰ãªãã¦ã€ã†ã£ã‹ã‚ŠworkerãŒæž¯æ¸‡ã—ã¦å¤§å¤‰ãªã“ã¨ã«ãªã£ãŸã®ã§ã€workerã®çŠ¶æ³ã‚’è¿”ã™ã‚¨ãƒ³ãƒ‰ãƒã‚¤ãƒ³ãƒˆã‚’è¿½åŠ ã™ã‚‹rackミドルウェアを書ã„ã¦ã„ã¾ã™ã€‚ github.com ã¶ã£ã¡ã‚ƒã‘kazeburoã•ã‚“ã®Plack-Middleware-ServerStatus-Liteã®ruby版ã§ã™ã€‚ github.com ã“ã®ä¸ã§ä½¿ã£ã¦ã„ã‚‹kazuhoã•ã‚“製ã®Parallel-Scoreboardも移æ¤ã—ã¾ã—ãŸã€‚(ã¾ã READMEãŒãªã„。。。) github.com å…ƒã¯ä¸‹è¨˜ã§ã™ã€‚ github.com Kazuho@Cybozu Labs: Parallel::Scoreboard ã§ãƒ¯ãƒ¼ã‚«ãƒ¼ãƒ—ãƒã‚»ã‚¹ã‚’モニタリングã™ã‚‹æ–¹æ³• 使ã„æ–¹ config.ruã«ä¸‹è¨˜ã®ç”¨ã«æ›¸ã。 # In config.ru use Rack::ServerSta
Unicorn: Rack HTTP server for Unix and fast clients
../ 2024-08-19 20:26 - Unicorn/ 2023-10-03 17:15 - examples/ 2023-10-03 17:15 - files/ 2023-12-30 01:23 - public/ 2021-03-12 20:34 - Application_Timeouts 2023-10-03 17:15 3K Application_Timeouts.html 2023-10-03 17:15 6K CONTRIBUTORS 2023-10-03 17:15 1K CONTRIBUTORS.html 2023-10-03 17:15 5K COPYING 2023-10-03 17:15 34K ChangeLog 2020-04-22 02:00 42 ChangeLog.html 2020-04-22 02:00 230 DESIGN 2023-10
Configurator.html
Application_Timeouts CONTRIBUTORS DESIGN FAQ HACKING ISSUES KNOWN_ISSUES LICENSE Links NEWS PHILOSOPHY README SIGNALS Sandbox TUNING unicorn_1 Unicorn Configurator HttpServer OobGC PrereadInput StreamInput TeeInput Util Worker Methods #after_fork #after_worker_exit #after_worker_ready #before_exec #before_fork #check_client_connection #client_body_buffer_size #default_middleware #early_hints #list
Unicornã®Workerプãƒã‚»ã‚¹ã‚’リサイクルã™ã‚‹ - Qiita
Web サーãƒã‚„アプリケーションサーãƒã¯ã€ãƒªã‚¯ã‚¨ã‚¹ãƒˆæ•°ã‚„メモリ使用é‡ãŒã‚る閾値を超ãˆãŸã‚‰ãƒ—ãƒã‚»ã‚¹ã®å†èµ·å‹•ã‚’è¡Œã†ä»•çµ„ã¿ãŒç”¨æ„ã•ã‚Œã¦ã„ã¾ã™ã€‚ Apache ã«ã‚‚ IIS ã«ã‚‚ãã®ãŸã‚ã®è¨å®šé …ç›®ãŒç”¨æ„ã•ã‚Œã¦ã„ã¦ã€ã“ã®ä»•çµ„ã¿ã‚’一般ã«ãƒªã‚µã‚¤ã‚¯ãƒ«ã¨å‘¼ã¶ã‚ˆã†ã§ã™ã€‚ UA ã‹ã‚‰ã®ãƒªã‚¯ã‚¨ã‚¹ãƒˆã‚’ã•ã°ãワーカープãƒã‚»ã‚¹ãŒç–²å¼Šã›ãšã€ãƒ‘フォーマンスをç¶æŒã™ã‚‹ãŸã‚ã®ä»•çµ„ã¿ã§ã™ã。 Ruby ã§äººæ°—ã®ã‚¢ãƒ—リケーションサーãƒã§ã‚ã‚‹ Unicorn ã§ã™ãŒã€æ®‹å¿µãªãŒã‚‰ã“ã®ãƒªã‚µã‚¤ã‚¯ãƒ«ã®æ©Ÿèƒ½ã‚’æŒã£ã¦ã„ã¾ã›ã‚“。 ã©ã†ã«ã‹ã§ããªã„ã‹ãªãƒ¼ã¨æ€ã£ã¦èª¿ã¹ã¦ã„ãŸã‚‰ã€ã“ã®çŸæ‰€ã‚’補完ã™ã‚‹ Gem ãŒãƒªãƒªãƒ¼ã‚¹ã•ã‚Œã¦ã„ã¾ã—ãŸã€‚ Unicorn Worker Killer | github ãã®ã¾ã‚“ã¾ã§ã™ã。 ã“ã® Gem ã§ã§ãã‚‹ã®ã¯ã€ä»¥ä¸‹ã®ã‚ˆã†ãªãƒªã‚µã‚¤ã‚¯ãƒ«ã§ã™ã€‚ ・一定ã®ãƒªã‚¯ã‚¨ã‚¹ãƒˆæ•°ã‚’超ãˆãŸã‚‰ãƒ©ãƒ³ãƒ€ãƒ ã§å†èµ·å‹•ã™ã‚‹ï¼ˆä¸Šé™ã«é”ã—ãŸã‚‰å¿…ãšå†èµ·å‹•ï¼‰ã€‚
rubyã§å‹•ç”»ãªã©ã®å¤§ããªãƒ•ã‚¡ã‚¤ãƒ«ã‚’アップãƒãƒ¼ãƒ‰ã™ã‚‹å ´åˆã®ãƒ¡ãƒ¢ãƒªã®ä½¿ã„æ–¹ã«ã¤ã„㦠- Qiita
動画ファイルãªã©ã®ã€ãƒ¡ãƒ¢ãƒªé‡ã«å¯¾ã—ã¦å·¨å¤§ãªãƒ•ã‚¡ã‚¤ãƒ«ã‚’アップãƒãƒ¼ãƒ‰ã™ã‚‹å ´åˆã€ãƒ•ã‚¡ã‚¤ãƒ«ã‚’å…¨ã¦èªã¿è¾¼ã‚“ã§é€ä¿¡ï¼å—ä¿¡ã‚’è¡Œã†ã¨ãƒ¡ãƒ¢ãƒªã‚’圧迫ã—ã¦ã—ã¾ã†ãŸã‚å•é¡ŒãŒã‚ã‚‹å ´åˆãŒã‚ã‚Šã¾ã™ã€‚ ã“ã†ã„ã†å ´åˆã¯ãƒ¡ãƒ¢ãƒªã®ä½¿ç”¨é‡ã¯ä¸€å®šç¯„囲ã«æŠ‘ãˆã‚‹ãŸã‚ã«ã€ãƒãƒƒãƒ•ã‚¡ã«å°‘ã—ã¥ã¤èªã¿è¾¼ã‚“ã§ã¯é€ä¿¡ã—ã€é€ä¿¡ã§ããŸã‚‰æ¬¡ã‚’é€ä¿¡ã—ã€ã‚ã‚‹ã„ã¯å°‘ã—å—ã‘å–ã£ãŸã‚‰ãƒ•ã‚¡ã‚¤ãƒ«ã‚·ã‚¹ãƒ†ãƒ ã«æ›¸ã出ã—ã€æ›¸ã出ã—ãŸã‚‰æ¬¡ã‚’èªã¿è¾¼ã‚€ã¨ã„ã£ãŸã“ã¨ã‚’ã‚„ã‚‹ã“ã¨ã«ãªã‚Šã¾ã™ã€‚ rubyã§ã“れを行ã†å ´åˆã«ã©ã†ã„ã†æ–¹æ³•ã‚’ã¨ã‚Œã°ã„ã„ã®ã‹ã«ã¤ã„ã¦ç¢ºèªã—ãŸã®ã§ã¾ã¨ã‚ã¦ã¿ã¾ã—ãŸã€‚ çµè«– é€ä¿¡ï¼ˆã‚¯ãƒ©ã‚¤ã‚¢ãƒ³ãƒˆå´ï¼‰ net/http を使ã†å ´åˆã¯ Net::HTTPGenericRequest#body_stream= ã§èªã¿å‡ºã—å¯èƒ½ãª File ã‚„ IO を渡ã™ã€‚ HTTPClient を使ã†å ´åˆã¯ãƒ‡ãƒ¼ã‚¿ã¨ã—ã¦èªã¿å‡ºã—å¯èƒ½ãª File ã‚„ IO を渡ã™ã€‚ RestClient ã¯1.7.
Unicornã®graceful restartã§å°‘ã—ãƒãƒžã£ãŸä»¶ - hack in 3 minutes
Railsサーãƒãƒ¼ã®Unicornã¯masterプãƒã‚»ã‚¹ã«USR2シグナルをé€ã‚‹ã¨ã€æ–°ã—ã„è¨å®šãƒ»ã‚¢ãƒ—リã®ãƒªãƒãƒ¼ãƒ‰ã‚’ç„¡åœæ¢ã§è¡Œã†graceful restartãªå‹•ãã‚’ã—ã¦ãã‚Œã¾ã™ã€‚ ã“ã®ä»•çµ„ã‚’ç†è§£ã—ã¦ãªã‹ã£ãŸã®ã§ãã‚Œã®ãƒ¡ãƒ¢ã€‚ ドã‚ュメントも何も見ã¦ãªã„時ã®èªè˜ ç¾è¡Œã®masterãŒæ–°ã—ã„è¨å®šãƒ»ã‚¢ãƒ—リをãƒãƒ¼ãƒ‰ã—ãŸæ–°masterプãƒã‚»ã‚¹ã‚’forkã™ã‚‹ æ–°masterã¨ãã“ã‹ã‚‰ç”Ÿãˆã‚‹workerãŒreadyã«ãªã£ã¦ã€æ–°æ—§ã®UnicornãŒå‡¦ç†ã‚’è¡Œã† é ƒåˆã„を見ã¦æ—§masterã¯å‹æ‰‹ã«è½ã¡ã‚‹ ã¨æ€ã£ã¦ãŸã‚“ã ã‘ã©ã€å®Ÿéš›è©¦ã—ã¦ã¿ã¦ã‚‚3ã®ã‚¿ã‚¤ãƒŸãƒ³ã‚°ã«ãªã£ã¦ã‚‚æ—§masterãŒãªã‹ãªã‹è½ã¡ã¦ãã‚Œãªã„。 æ—§masterã®CPU利用率ã¯ã ã‚“ã ん下ãŒã£ã¦ãã¦ã„ã‚‹ã®ã§ã€ãã‚ŒãŒãªããªã£ãŸã‚‰è½ã¡ã‚‹ã®ã‹ã¨å‹æ‰‹ã«æ€ã£ã¦ãŸã‘ã©ãã†ã§ã‚‚ãªã„。ウーンã¨ãªã£ã¦ã¦ãƒ‰ã‚ュメント見ãŸã‚‰æ™®é€šã«æ›¸ã„ã¦ã‚ã‚Šã¾ã—ãŸã€‚ USR2ã‚’é€ã£ãŸæ™‚ã®å‹•ã
Unicorn ã® graceful restart ã¨ç’°å¢ƒå¤‰æ•° - eagletmt's blog
Unicorn ã® graceful restart ã¯ç„¡åœæ¢ã§ã®ãƒ‡ãƒ—ãƒã‚¤ã‚’å¯èƒ½ã«ã—ã¦éžå¸¸ã«ä¾¿åˆ©ã ãŒã€fork を用ã„ã¦å®Ÿè£…ã•ã‚Œã¦ã„る都åˆã§å¤ã„プãƒã‚»ã‚¹ã‹ã‚‰æ–°ã—ã„プãƒã‚»ã‚¹ã«ç’°å¢ƒå¤‰æ•°ãŒå¼•ã継ãŒã‚Œã‚‹ãŸã‚ã€ãã®ã“ã¨ã«èµ·å› ã™ã‚‹ãƒˆãƒ©ãƒ–ルãŒã„ãã¤ã‹ã‚る。 dotenv ã®è¨å®šãŒæ›¸ã変ã‚らãªã„ è¨å®šæƒ…å ±ã‚’ dotenv ã§ç®¡ç†ã—ã¦ã„る人も多ã„ã¨æ€ã†ã‘ã©ã€ç’°å¢ƒå¤‰æ•°ã‚’使ã£ã¦ã„ã‚‹ã®ã§ç½ ãŒã‚る。 例ãˆã°æœ€åˆã« .env ã« MEMCACHE_SERVERS=memcache-server-001:11211 ã¨æ›¸ã„ã¦ã‚ã£ãŸã¨ã™ã‚‹ã€‚ ã“ã®ã¨ã Unicorn ã‚’èµ·å‹•ã™ã‚‹ã¨ã€dotenv ã«ã‚ˆã£ã¦ MEMCACHE_SERVERS=memcache-server-001:11211 ãŒç’°å¢ƒå¤‰æ•°ã«è¿½åŠ ã•ã‚Œã‚‹ã€‚ ãã®å¾Œã€æŽ¥ç¶šå…ˆã¨ã—㦠memcache-server-002:11211 ã‚’è¿½åŠ ã—ãŸããªã£ã¦ .env ã‚’ç·¨
How To Optimize Unicorn Workers in a Ruby on Rails App | DigitalOcean
Introducing Unicorn If you are a Rails developer, you’ve probably heard of Unicorn, a HTTP server that can handle multiple requests concurrently. Unicorn uses forked processes to achieve concurrency. Since forked processes are essentially copies of each other, this means that the Rails application need not be thread safe. This is great because it is difficult to ensure that our own code is thread
Debugging, Unicorns, Rails and Timeouts
The goal of this post is to describe a debugging session, and different possible approaches for debugging in general. It also contains some tips and tricks to use with unicorn and Rails to debug request timeouts. I usually enjoy debugging because of the unknown thing to fix, like a pirate finding a hidden box of gold :). Seriously, I think one of the most important things in writing software is de
Mac + unicorn + nginx + Rails　 - Giworld
環境 Mac OSX 10.8.3 Rails 3.2.13 Ruby 2.0.0-p0 Homebew ã¯å°Žå…¥æ¸ˆã¿ã¨ã™ã‚‹ Unicorn Install gem install unicorn è¨å®š /config/unicorn.rb ã‚’ä½œæˆ # ワーカーã®æ•° worker_processes 2 # ソケット経由ã§é€šä¿¡ã™ã‚‹ listen File.expand_path('tmp/sockets/unicorn.sock', ENV['RAILS_ROOT']) # ãƒã‚° stderr_path File.expand_path('log/unicorn.log', ENV['RAILS_ROOT']) stdout_path File.expand_path('log/unicorn.log', ENV['RAILS_ROOT']) # ダウンタイムãªãã™ preload_app
Ruby 2.0 Garbage Collection
There’s been a lot of talk about the improvements in Ruby 2.0’s Garbage Collection performance. That’s a great thing, because GC in older versions of Ruby wasn’t very performant. Most of the conversation about generational GC (or other strategies) revolves around making Ruby GC more efficient. This has a great impact on your users, because less GC time means that their pages return faster. But wha
Tuning unicorn
Application_Timeouts CONTRIBUTORS DESIGN FAQ HACKING ISSUES KNOWN_ISSUES LICENSE Links NEWS PHILOSOPHY README SIGNALS Sandbox TUNING unicorn_1 Unicorn Configurator HttpServer OobGC PrereadInput StreamInput TeeInput Util Worker Tuning unicorn unicorn performance is generally as good as a (mostly) Ruby web server can provide. Most often the performance bottleneck is in the web application running on
Heroku+Railsã®é«˜é€ŸåŒ– - æ€ã£ãŸã“ã¨
ã“ã‚“ã°ã‚“ã‚。 Herokuã«ä¹—ã£ã‘ã¦ã‚‹è‡ªåˆ†ã®ã‚µãƒ¼ãƒ“スを3Gã§è¦‹ãŸã¨ã〠èªã¿è¾¼ã¿é…ã™ãŽã¦ç™ºç‹‚ã—ã¾ã—ãŸã€‚ 対ç–ã€ä½•å€‹ã‹ã‚„ã£ã¦ã¿ãŸãƒ¡ãƒ¢ã€‚ Unicornå°Žå…¥ Gemfileã«unicornã‚’è¿½åŠ group :production gem 'unicorn' end config/unicorn.rbã‚’ä½œæˆ worker_processes Integer(ENV["WEB_CONCURRENCY"] || 3) timeout 15 preload_app true before_fork do |server, worker| Signal.trap 'TERM' do puts 'Unicorn master intercepting TERM and sending myself QUIT instead' Process.kill 'QUIT', Process.pid end de
logentriesã§nginxã®ã‚¨ãƒ©ãƒ¼ã‚’監視ã™ã‚‹ - komagataã®ãƒ–ãƒã‚°
å°ã•ã„railsサービスã‚ã‚‹ã‚ã‚‹ railsã§ä½•ã‹ã‚µãƒ¼ãƒ“スを作ã£ãŸã¨ã™ã‚‹ã€‚ExceptionNotificationも入れãŸã€‚NewRelicã§5分毎ã«æ»æ´»ç›£è¦–ã‚‚ã—ã¦ã‚‹ã€‚ãªã®ã«ã€Œè½ã¡ã¦ã‚‹ã¿ãŸã„ã§ã™ã€ã¨Twitterã§è¨€ã‚れる。 nginx + unicornã®çµ„ã¿åˆã‚ã›ã§ã‚¢ã‚¯ã‚»ã‚¹ãŒå¢—ãˆãŸæ™‚ã€å¤§æŠµçœŸã£å…ˆã«èµ·ã“ã‚‹ã®ãŒunicornã®time out。 2013/02/08 18:36:10 [error] 20932#0: *3506622 upstream timed out (110: Connection timed out) while reading response header from upstream, client: 〜ã£ã¦ã‚„ã¤ã€‚ é‡ã„DBã®ã‚¯ã‚¨ãƒªã¨ã‹ã§ã‚¢ã‚¯ã‚»ã‚¹ã‚’ã•ã°ãスピードより接続数ãŒå¢—ãˆã¦ãã¦èµ·ã“る。railsアプリã¾ã§å‡¦ç†ãŒè¡Œã‹ãªã„ã®ã§Exceptionã¯èµ·ã“らãªã„
リリースã€éšœå®³æƒ…å ±ãªã©ã®ã‚µãƒ¼ãƒ“スã®ãŠçŸ¥ã‚‰ã›
最新ã®äººæ°—エントリーã®é…ä¿¡
j次ã®ãƒ–ックマーク
kå‰ã®ãƒ–ックマーク
lã‚ã¨ã§èªã‚€
eコメント一覧を開ã
oページを開ã
GitHub - kzk/unicorn-worker-killer: Automatically restart Unicorn workers based on 1) max number of requests and 2) max memory
Server::Starter 㧠Unicorn ã‚’èµ·å‹•ã™ã‚‹å ´åˆã® Slow Restart : sonots:blog
「Server::Starterã«å¯¾å¿œã™ã‚‹ã¨ã¯ã©ã†ã„ã†ã“ã¨ã‹ã€ã®è£œè¶³ : sonots:blog
Ruby Threads, Unicorn, and You
{{#tags}}- {{label}}
{{/tags}}