DevOps Classroom Series – 19/Sept/2020

Story of a organization Learning Thoughts

• LT has a product LT-HRMS

• LT has the

  • Management Team
  • Sales Team
  • Research & Development
    • Architecture
    • Development
    • Testing
    • IT Infra Team

• LT HRMS product is developed, It can be customized. LT HRMS has got a deal from two different customers with some different customizations. There are some other customers in pipeline

• Now we have three project created

  1. Customer Apple
  2. Customer Amazon
  3. Core Features

• How should we handle code repository for these situations. To handle parallel tracks in projects we generally create branch

• Branches in version controls allows to merge changes from one branch to other.

• How do i create a branch in git

  • Git already has default branch master
  • Git Branch is a pointer which looks at latest commit.

• HEAD is a pointer which points to commit id in git. making HEAD to move to different location is considered as checkout

• Lets create branches for apple, core and amazon

• Moving around branches

• If you want create a branch and move the HEAD to that branch (Our sales struck a new deal of tesla, so we need to create)

git checkout master
git checkout -b testla => create a new branch & checkout to that branch

How Git Works?

• Git is a content tracker

• Create a new directory and execute git init, a .git folder will be created.

• Lets add two simple files (Readme.md, .gitignore)

• Now execute git log

• Lets understand hashing: Hash algorithms are functions that generate a fixed length result (the hash) from a given input.

• Git uses SHA-1 Hashing to generate commit ids

• Git commit id is SHA-1 hash of

  • Snapshot of changes in staging area
  • Author name
  • Author email
  • Date time
  • Commit message

• How git maintains versions

  • We have create changes in git which appear as
  • Lets use plumbing commands

  git cat-file -t # This command will tell the type 
  git cat-file -p # This will print the contents of SHA1 hash
  

  • When we print commit id we get
    • tree => snapshot of changes done in staging area
    • parent => parent commit id
    • author =>
    • committer =>
    • commit message
  • Now lets explore the snapshot
  • blob stands for file and the hash is hash of contents in the file and tree is folder.
  • Now for the current commit the previous commit will be parent and git travels to the initial commit (commit with no parent)
  • Now lets create a new commit by copying the same file in 3 different folders
  • Now lets look into .git folder, The hashes are encoded in to .git/objects
  • Our current working tree depends on pointer HEAD
  • Summarize:
    • HEAD points to Branch which in turns points to latest commit id
  • Git always recommends that HEAD should be pointing to branch not commit, when we force git to do this then it gets into detached head
  • Internally Branch is nothing but a moving pointer which points to latest commit

• Apple and Amazon have asked for features, There are some common features which

  • core team will work
  • Rest of features will be worked by customer teams

• How can i make change in one branch and ensure that change is available to other branches

• Next Steps:

  • Merge
  • Cherry pick
  • Rebase
  • Remotes
  • Tags
  • Un staging the changes
  • modifying commits
  • git sub modules & hooks ( After we finish jenkins concepts)
  • git branching strategy ( After we finish jenkins concepts)