Northeast PHP 2012 - Varnish

Transcript

1. Varnish, The Good, The Awesome, and the Downright Crazy By Mike Willbanks Sr. Web Architect Manager NOOK Developer Northeast PHP August 12, 2012
2. 2 • Talk  Slides will be online later! • Me  Sr. Web Architect Manager at NOOK Developer  Prior MNPHP Organizer  Open Source Contributor (Zend Framework and various others)  Where you can find me: • Twitter: mwillbanks G+: Mike Willbanks • IRC (freenode): mwillbanks Blog: http://blog.digitalstruct.com • GitHub: https://github.com/mwillbanks Housekeeping…
3. 3 • What is Varnish • The Good : Why…   The quick, easy and hardly informed way… • The Awesome : How…   VCL’s, Directors and more… • The Crazy : Go…   ESI, Purging, VCL C, and VMOD… • Varnish Command Line Apps   varnishtop, varnishstat, etc. Agenda
4. What is Varnish? Official Statement What the hell it means Graphs, oh my!
5. 5 “Varnish is a web application accelerator. You install it in front of your web application and it will speed it up significantly.” Official Statement
6. 6 • Varnish allow you to accelerate your website  By using memory and keeping in mind cookies, request headers and more… • It caches pages so that your web server can RELAX!  What about my apache, tomcat, nginx and (mongrel|thin| goliath….)  Generally caching by TTL + HTTP Headers (cookies too!) • A load banancer, proxy and more…  What? …. Yes, it can do that! What The Hell? Tell me!
7. 7 • CaringBridge Status Server  Getting a message to mobile users.  The system is down, or we want to be able to communicate a message to them about some subject… maybe a campaign.  The apps and mobile site rely on an API • Trouble in paradise? Few and far in between.  Let an API talk to a server…  A story on crashing and burning before varnish. A General Use Case
8. 8 The Graph - AWS 0 10000 20000 30000 40000 50000 60000 70000 80000 Small X-Large Small Varnish Requests Requests 0 50 100 150 200 250 300 350 400 450 500 Small X-Large Small Varnish Time Time 0 100 200 300 400 500 600 700 Small X-Large Small Varnish Req/s Req/s 0 2 4 6 8 10 12 14 Small X-Large Small Varnish Peak Load Peak Load
9. 9 The Raw Data Small   X-­‐Large   Small  Varnish   Concurrency   10   150   150   Requests   5000   55558   75000   Time   438   347   36   Req/s   11.42   58   585   Peak  Load   11.91   8.44   0.35   Comments   19,442   failed  reqs  
10. The Good – Listen Up! Installment Documentation Finding Existing VCL’s
11. 11 • RTM : http://goo.gl/hl4Tt  Debian: sudo apt-get install varnish  EPEL: yum install varnish • only 6.x otherwise you’ll be out of date!  WOOT Compiling #git • git clone git://git.varnish-cache.org/varnish-cache • cd varnish-cache • sh autogen.sh • ./configure • make && make install Installment
12. 12 Varnish Daemon • varnishd  -a address[:port] listen for client  -b address[:port] backend requests  -T address[:port] administration http  -s type[,options] storage type (malloc, file, persistence)  -P /path/to/file PID file  Many others; these are generally the most important. Generally the defaults will do with just modification of the default VCL (more on it later).
13. 13 • Reference Manual  https://www.varnish-cache.org/docs/3.0/reference/index.html • Tutorial – more like a book version of the reference manual  https://www.varnish-cache.org/docs/3.0/tutorial/index.html • Knock yourselves out! There is a ton of documentation • Yes, this makes happy developers.  Documentation is very accurate, read carefully.  Focus heavily on VCL’s, that is generally what you need.  I’m attempting to show you some of how this works but you will require the documentation to assist you. Documentation
14. 14 • VCL’s are available for common open source projects  Hi wordpress and drupal! • https://www.varnish-cache.org/trac/wiki/VarnishAndWordpress • https://www.varnish-cache.org/trac/wiki/VarnishAndDrupal  Examples of all sorts of crazy • https://www.varnish-cache.org/trac/wiki/VCLExamples Existing VCL’s – The truly lazy…
15. 15 backend default { .host = "127.0.0.1“; .port = "8080"; } sub vcl_recv { if (!(req.url ~ "wp-(login|admin)")) { unset req.http.cookie; } } sub vcl_fetch { if (!(req.url ~ "wp-(login|admin)")) { unset beresp.http.set-cookie; } } Wordpress = Bad Slashdot Bad!!!
16. The Awesome – Going Places VCL Directors A Few Examples
17. 17 VCL’s by Diagram…
18. 18 • VCL State Engine  Each Request is Processed Separately & Independently  States are Isolated but are Related  Return statements exit one state and start another  VCL defaults are ALWAYS appended below your own VCL • VCL can be complex, but…  Two main subroutines; vcl_recv and vcl_fetch  Common actions: pass, hit_for_pass, lookup, pipe, deliver  Common variables: req, beresp and obj  More subroutines, functions and complexity can arise dependent on condition. VCL – Varnish Configuration Language
19. 19 • vcl_init – VCL is loaded, no request yet; VMOD initialization • vcl_recv – Beginning of request, req is in scope • vcl_pipe – Client & backend data passed unaltered • vcl_pass – Request goes to backend and not cached • vcl_hash – call hash_data to add to the hash • vcl_hit – called on request found in the cache • vcl_miss – called on request not found in the cache • vcl_fetch – called on document retrieved from backend • vcl_deliver – called prior to delivery of cached object • vcl_error – called on errors • vcl_fini – all requests have exited VCL, cleanup of VMOD’s VCL – Subroutines – breaking it down.
20. 20 • Always Available   now – epoch time • Backend Declarations   .host – hostname / IP   .port – port number • Request Processing   client – ip & identity   server – ip & port   req – request information VCL - Variables • Backend Req Prepartion   bereq – backend request • Retrieved Backend Request   beresp – backend response • Cached Object   obj – Cached object, can only change .ttl • Response Preparation   resp – http stuff
21. 21 • hash_data(string) – adds a string to the hash input.   Request host and URL is default from the default vcl. • regsub(string, regex, sub) – substitution on first occurance   sub can contain numbers 0-n to inject matches from the regex. • regsuball(string, regex, sub) – substitution on all occurances • ban(expression) – Ban all objects in cache that match • ban(regex) – Ban all objects in cache that have a URL match VCL - Functions
22. 22 • Directors allow you to talk to the backend servers • Directors are a glorified reverse proxy   Allows for certain types of load balancing   Allows for talking to a cluster “A director is a logical group of backend servers clustered together for redundancy. The basic role of the director is to let Varnish choose a backend server amongst several so if one is down another can be used.” Directors
23. 23 • Random Director – picks a backend by random number • Client Director – picks a backend by client identity • Hash Director – picks a backend by URL hash value • Round-Robin Director – picks a backend in order • DNS Director – picks a backend by means of DNS  Random OR Round-Robin • Fallback – picks the first “healthy” backend Directors – The Types
24. 24 • To ensure healthy backends, you need to use probing.  It really sounds like a colonoscopy for servers; which it is. • Variables  .url  .request  .window  .threshold  .intial  .expected_response  .interval  .timeout Director - Probing
25. 25 Example VCL Configuration
26. The Crazy ESI – Edge-Side Includes Purging VMOD
27. 27 • ESI is a small markup language much like SSI (server side includes) to include fragments (or dynamic content for that matter). • Think of it as replacing regions inside of a page as if you were using XHR (AJAX) but single threaded. • Three Statements can be utilized.  esi:include – Include a page  esi:remove – Remove content  <!-- esi --> - ESI disabled, execute normally ESI – Edge Side Includes
28. 28 ESI – By Diagram
29. 29 • In vcl_fetch, you must set ESI to be on  set beresp.do_esi = true;  By default, ESI will still cache, so add an exclusion if you need it • if (req.url == “/show_username.php”) { return (pass); } • This is a good thing, you may want to cache user information to the right people (aka by cookie value) so that you don’t reload it on every request.  Varnish refuses to parse content for ESI if it does not look like XML • This is by default; so check varnishstat and varnishlog to ensure that it is functioning like normal. Using ESI
30. 30 <html> <head><title>Rock it with ESI</title></head> <body> <header> <esi:include src="/user_header.php" /> <!-- Don't do this as you'd lose the advantage of varnish --> <!--esi <?php include 'user_header.php'; ?> --> </header> <section id="main"></section <footer></footer> </body> </html> ESI – By Example
31. 31 • The various ways of purging  varnishadm – command line utility • It’s the ole finger in the back of the throat  Sockets (port 6082) – everyone likes a good socket wrench • Sure, Ipecac is likely overkill.  HTTP – now that is the sexiness • A few headers, nothing forced. Purging
32. 32 varnishadm -T 127.0.0.1:6082 purge req.url == "/foo/bar“ telnet localhost 6082 purge req.url == "/foo/bar telnet localhost 80 Response: Trying 127.0.0.1... Connected to localhost. Escape character is '^]'. PURGE /foo/bar HTTP/1.0 Host: bacon.org Purging Examples
33. 33 • Distributed Purging… like a sorority party.  Use a message queue (or gearman job server)  Have a worker that knows about the varnish servers  Submit the request to clear the cache in the asynchronously or synchronously depending on your use case. • Have enough workers to make this effective at purging the cache quickly.  This will make it far easier to scale; you can either store the servers in a config file, database or anything else you think is relevant. Distributed Purging
34. 34 • Before getting into VMOD; did you know you can embed C into the VCL for varnish? • Want to do something crazy fast or leverage a C library for pre or post processing? • I know… you’re thinking that’s useless..  On to the example; and a good one from the Varnish WIKI! Embedding C in VCL – you must be crazy
35. 35 C{ #include <syslog.h> }C sub vcl_something { C{ syslog(LOG_INFO, "Something happened at VCL line XX."); }C } # Example with using varnish variables C{ syslog(LOG_ERR, "Spurious response from backend: xid %s request %s %s \"%s\" %d \"%s\" \"%s\"", VRT_r_req_xid(sp), VRT_r_req_request(sp), VRT_GetHdr(sp, HDR_REQ, "\005host:"), VRT_r_req_url(sp), VRT_r_obj_status(sp), VRT_r_obj_response(sp), VRT_GetHdr(sp, HDR_OBJ, "\011Location:")); }C VCL - Embedded C for syslog – uber sexy
36. 36 • Taking VCL embedded C to the next level • Allows you to extend varnish and create new functions • Now, if you are writing modules for varnish you have a specialty use case!  Go read up on it!  https://www.varnish-cache.org/docs/trunk/reference/vmod.html VMOD – Varnish Modules / Extensions
37. 37 • The VMOD std is shipped with varnish; it provides some useful commands  toupper  tolower  set_up_tos  random  log VMOD - std  syslog  fileread  duration  integer  collect
38. Varnish Command Line Apps varnish varnishadm varnishhist varnishlog varnishncsa varnishreplay varnishsizes varnishstat varnishtest varnishtop
39. 39 • What is varnish doing right now? • How do I debug what is happening?  varnishtop What is Varnish doing…
40. 40 What is Varnish doing…
41. 41 • Many times people want to log the requests to a file  By default Varnish only stores these in shared memory.  Apache Style Logs • varnishncsa –D –a –w log.txt  This will run as a daemon to log all of your requests on a separate thread. Logging
42. 42 Logging
43. 43 • Need to warm up your cache before putting a sever in the queue or load test an environment?  varnishreplay –r log.txt • Replaying logs can allow you to do this. This is great for when you are going to be deploying code to check for performance issues.  Although… be careful so that you don’t POST data or create data on peoples accounts. Maybe cat the file and remove anything that executes on data. Cache Warmup
44. 44 • How to see your cache hit ratios…  varnishstat • Want to parse them from XML so you can create a sexy administration panel?  varnishstat –x Cache Hit Ratios? No Problem
45. 45 Cache Hit Ratios? No Problem
46. Questions? These slides will be posted to SlideShare & SpeakerDeck.  SpeakerDeck: http://speakerdeck.com/u/mwillbanks  Slideshare: http://www.slideshare.net/mwillbanks  Twitter: mwillbanks  G+: Mike Willbanks  IRC (freenode): mwillbanks  Blog: http://blog.digitalstruct.com  GitHub: https://github.com/mwillbanks