ADN Columbus 2013 - DevOps with Visual Studio 2012

Transcript

1. DevOps  with  Visual  Studio  2012   Developer  and  Operations  Interactions   Josh  Harrison   Improving  Enterprises  
2. Agile  Development  Methodology  
3. Deployment!  
5. Panic!  
6. Developer’s  Response  
7. Operation’s  Response   •  You  gave  us  unstable  code.  
8. The  conversation   •  Operations:  “Something  is  broke  in  production”   •  Developers:  “It  worked  in  the  Dev  and  Test  environment”  (that  we  set  up)   •  Operations:  “It  looks  like  there  is  a  permission  issue  with  the  database”   •  Developers:  “What  makes  you  think  that?”   •  Operations:  “When  I  log  into  the  box  I  see  errors  in  the  app  log”   •  Developers:  “Can  we  get  permission  to  the  production  box”  
9. Operations  Response  
10. Mean  time  to  Innocence   Engineer:  “I  have  great  news!”   Regular  Human  Being:  “Did  you  fix  the  problem?”   Engineer:  “No,  but  I  can  prove  it’s  not  our  fault”  
11. Incentives  Matter  
12. Operation’s  Incentives   • Uptime   • Stability  
13. Developer’s  Incentives     • Speed  to  market   • New  functionality   • Rapid  change   • Slamming  in  cool  new  architectures   • Agility  
14. The  Conflict      
15. Agile  is  making  this  problem  worse      
16. DevOps   "is  a  software  development   method  that  stresses   communication,  collaboration   and  integration  between   software  developers  and   information  technology  (IT)   professionals"    
17. Adopting  DevOps   •  Process   •  Culture   •  Tools  
18. Process   •  Version  Control   •  Automate   •  Build   •  Testing   •  Deployment   •  Configuration   •  Monitor   •  Reflect   •  Continuous  Improvement   •  Treat  infrastructure  as  Code  
19. Tools   •  Version  Control  (TFS,  Git,  SVN)   •  Resource  Pools  (Windows  Azure,  Hyper  V,  VMWare,  Amazon  EC2)   •  Configuration  Management  (Windows  Azure,  TFS,  Chef,  Puppet)   •  Continuous  Integration  (TFS,  Team  City,  Hudson)   •  Monitoring  (SCOM,  Avicode,  Log4Net)  
20. Culture   •  Overcome  Stereotypes   •  Developers  are  lazy,  reckless,  and  have  large  egos   •  Sys  admins  just  want  to  say  “No”   •   Engage  early  and  often   •  Transparency  –  What  will  this  change  do  to  the  other  side.   •  Define  done  (last  mile  syndrome)   •  Invite  everyone  to  standups   •  …and  retrospectives  
21. Practice   •  Build  Often   •  Test  Often   •  Deploy  Often   •   Engage  Often  
22. Infrastructure  as  Code   Imagine  your  data  center  suddenly  disapeared.  Which  of  the  below  would  take   the  longest  to  bring  back  online?     •  Source  Code   •  Application  Data   •  Server  Infrastructure  
23. Implementing  Infrastructure  as  Code   1.  Create  a  resource  pool   2. Compose  infrastructure  into  network  addressable  resources   3.  Connect  addressable  resources  
24. Demo  
25. Further  Reading    
26. Q  &  A  
27. About  Me   •  Name:  Josh  Harrison   •  Company:  Improving  Enterprises   •  Twitter:  @josh_r_harrison   •  Email:  josh.r.harrison@gmail.com