Failure and Availibility

1
Failure and Availibility

• DS Lecture 12

2
Optimistic Replication

• Let everyone make changes
• Only 3 transactions ever abort
• Make changes, send updates
• If someone elses changes come through with T_him
  lt T_you, your changes are overridden
• Wait for a bit before committing
• deadlocks

3
Two Phase Commit

• Blocking Trades availability for correctness
• How?

Can commit?
TM
Cohort
yes
Do Commit
Committed
4
2PC Blocking
Can commit?
TM
Cohort
yes
Do Commit
Committed
5
2PC Blocking

• After Can_Commit, TM fails
• Everyone blocks
• After Do_Commit, chort fails
• TM blocks
• Single host failure can compromise the
  availability of the system

6
2PC Blocking

• Blocks because of the fear of unknown
• The system can be in an unknown state, so
  everyone blocks hoping for stability

7
3PC

• Remove the fear of unknown
• Structure the state transition to remove
  ambiguity between commit/abort

8
3PC Three Phase Commit

• Non-blocking consistency
• Combines agreement with transactions
• 1. There's no single state from which it's
  possible to make a transition directly to either
  a commit or abort state.
• 2. There is no state in which it is not possible
  to make a final decision and from which a
  transition to a commit state can be made.

9
Three-phase Commit
10
3PC Protocol

• Timeout is more meaningful
• Phase 1 yes/no
• Failure, Timeout abort
• Phase 2 Prepare to commit
• Cohort Failure
• Before vote
• abort
• After vote
• Commit
• Phase 3 Commit
• All acks commit
• After failure commit

11
Recovery

• Checkpoints
• Independent
• Coordinated
• Logging
• Independent checkpoints
• Store activity

12
Quorum-based Protocols

• Read and Write quorum
• K G gt N
• If Write quorum gt n/2 1
• Always an overlap with read quorum
• Improve availbility

13
Starbucks 2PC?

• Interesting story from the web ..
• Synchronous Vs Asynchronous
• Failure Model