Implementation Plan October 20, 2009

Personnel

  • Programmer is Michael Wenk
  • MW Developer is Alex Alfieri
  • QC is Sandra

Preparation for production release

During maintenance window

S = maintenance window start

  • (S  + 0m Programmer) Put all application servers into maintenance mode and shutdown Sakai on each one
  • (S + 5m Programmer) Deploy sakai release tag via setup-sakai-node.sh to sakaiprod1.
  • (S + 5m MW Developer) Change MyUCD's messages to use CAS.
  • (S + 20m Programmer) Restart sakaiprod1 sakai instance
  • (S + 22m QC) Verification ???
    • Log in to see if CAS is working
    • Log out and make sure it is complete
  • (S + 30m) Go/No Go
  • (S + 31m Programmer) Deploy release tag via setup-sakai-node.sh to sakaiprod2 . . . sakaiprod6, and sakaiprodmail1.
  • (S + 40m Programmer) Start Sakaiprod2 . . . sakaiprod6, and sakaiprodmail1.
  • (S + 45m Programmer) Take smartsite-prod out of maintenance mode 
  • (S + 47m Sandra) Send Release Management email

Fall-back Procedure (If needed) Time D is the decision time made to fallback

  • (D + 0m Programmer) Shutdown Sakai on all nodes
  • (D + 2m MW Developer) Change MyUCD's messages to use DistAuth.
  • (D + 2m Programmer) Remove sakai.properties, toolOrder.xml, and bash_profile from config-staging and rename sakai.properties-prod-022, toolOrder.xml-prod-022, and bash_profile-pord-022 to sakai.properties, toolOrder.xml, and bash_profile
  • (D + 4m Programmer) Run mvn sakai:deploy from previous tag directory
  • (D + 15m Programmer) Deploy previous tag on head node and all other app servers via setup-sakai-node.sh
  • (D + 20m Programmer) Restart sakai on head node
  • (D + 22m Programmer) Make sure sakai running is old tag (CAS is removed, etc)
  • (D + 25m Programmer) Restart sakai on remaining nodes
  • (D + 27m Programmer) Take application servers out of maintenance mode