Crash recovery

1. Crash recovery

2. Review: The ACID properties • Atomicity: All actions in the Xaction happen, or none happen. • Consistency: If each Xaction is consistent, and the DB starts consistent, it ends up consistent. • Isolation: Execution of one Xaction is isolated from that of other Xacts. • D urability: If a Xaction commits, its effects persist. • CC guarantees Isolation and Atomicity. • The Recovery Manager guarantees Atomicity & Durability.

3. Transaction failure : Logical errors: application errors (e.g. div by 0, segmentation fault) System errors: deadlocks Aborts System crash: hardware/software failure causes the system to crash. Disk failure: head crash or similar disk failure destroys all or part of disk storage Why is recovery system necessary? • The lost data can be in main memory or in disk

4. Volatile storage: does not survive system crashes examples: main memory, cache memory Nonvolatile storage: survives system crashes examples: disk, tape, flash memory, non-volatile (battery backed up) RAM Stable storage: a “mythical” form of storage that survives all failures approximated by maintaining multiple copies on distinct nonvolatile media Storage Media

5. Recovery and Durability • To achieve Durability: Put data on stable storage • To approximate stable storage make two copies of data

6. Solution: Write to the first disk Write to the second disk when the first disk completes The process is complete only after the second write completes successfully Recovery (from disk failures, etc): Detect bad blocks with the checksum (e.g. parity) Two good copies, equal blocks: done One good, one bad : copy good to bad Two bad copies: ignore write Two good, unequal blocks? Stable-Storage Implementation • Ans: Copy the second to the first

7. Example: transfer $50 from account A to account B goal is either to perform all database modifications made by Tior none at all. Requires several inputs (reads) and outputs (writes) Failure after output to account A and before output to B…. DB is corrupted! Recovery and Atomicity

8. Recovery Algorithms • Recovery algorithms are techniques to ensure database consistency and transaction atomicity and durability despite failures • Recovery algorithms have two parts • Actions taken during normal transaction processing to ensure enough information exists to recover from failures • Actions taken after a failure to recover the database contents to a state that ensures atomicity, consistency and durability

9. Simplifying assumptions: Transactions run serially logs are written directly on the stable storage Log: a sequence of log records; maintains a record of update activities on the database. (Write Ahead Log, W.A.L.) W.A.L. Log: record the operation on the log, before you write it on the db. Record everything else on the log before commit Log-Based Recovery

10. Log based approach • Log records for transaction Ti: • <Tistart > • <Ti, X, V1, V2> • <Ticommit > • Two approaches using logs • Deferred database modification • Immediate database modification

11. Log example Log <T1, start> <T1, A, 1000, 950> <T1, B, 2000, 2050> <T1, commit> Transaction T1 Read(A) A =A-50 Write(A) Read(B) B = B+50 Write(B)

12. Ti starts: write a <Tistart> record to log. Tiwrite(X) write <Ti, X, V> to log: V is the new value for X The write is deferred Note: old value is not needed for this scheme Tipartially commits: Write <Ticommit> to the log DB updates by reading and executing the log: <Tistart> …… <Ticommit> Deferred Database Modification

13. How to use the log for recovery after a crash? Redo: if both <Tistart> and <Ti commit> are there in the log. Crashes can occur while the transaction is executing the original updates, or while recovery action is being taken example transactions T0and T1(T0executes before T1): T0: read (A) T1: read (C) A: - A - 50C:- C- 100 Write (A) write (C) read (B) B:- B + 50 write (B) Deferred Database Modification

14. Below we show the log as it appears at three instances of time. Deferred Database Modification (Cont.) <T0, start> <T0, A, 950> <T0, B, 2050> <T0, commit> <T1, start> <T1, C, 600> (b) <T0, start> <T0, A, 950> <T0, B, 2050> (a) <T0, start> <T0, A, 950> <T0, B, 2050> <T0, commit> <T1, start> <T1, C, 600> <T1, commit> (c)

15. Database updates of an uncommitted transaction is allowed Tighter logging rules are needed to ensure transactions are undoable Write records must be of the form: <Ti, X, Vold, Vnew > log record must be written before database item is written Output of DB blocks can occur: Before or after commit In any order Immediate Database Modification

16. Log Write Output <T0start> <T0, A, 1000, 950> <To, B, 2000, 2050> A = 950 B = 2050 <T0commit> <T1start> <T1, C, 700, 600> C = 600 BB, BC <T1commit> BA Note: BXdenotes block containing X. Immediate Database Modification Example

17. Recovery procedure : Undo : <Ti, start > is in the log but <Ticommit> is not. Undo: restore the value of all data items updated by Ti to their old values, going backwards from the last log record for Ti Redo: <Tistart> and <Ticommit> are both in the log. Redo:sets the value of all data items updated by Tito the new values, going forward from the first log record for Ti Both operations must be idempotent: even if the operation is executed multiple times the effect is the same as if it is executed once Immediate Database Modification (Cont.)

18. I M Recovery Example <T0, start> <T0, A, 1000, 950> <T0, B, 2000, 2050> <T0, commit> <T1, start> <T1, C, 700, 600> <T1, commit> (c) <T0, start> <T0, A, 1000, 950> <T0, B, 2000, 2050> <T0, commit> <T1, start> <T1, C, 700, 600> (b) <T0, start> <T0, A, 1000, 950> <T0, B, 2000, 2050> (a) Recovery actions in each case above are: (a) undo (T0): B is restored to 2000 and A to 1000. (b) undo (T1) and redo (T0): C is restored to 700, and then A and B are set to 950 and 2050 respectively. (c) redo (T0) and redo (T1): A and B are set to 950 and 2050 respectively. Then C is set to 600

19. Problems in recovery procedure as discussed earlier : searching the entire log is time-consuming we might unnecessarily redo transactions which have already output their updates to the database. How to avoid redundant redoes? Put marks in the log indicating that at that point DB and log are consistent. Checkpoint! Checkpoints

20. Checkpoints • At a checkpoint: • Output all log records currently residing in main memory onto stable storage. • Output all modified buffer blocks to the disk. • Write a log record < checkpoint> onto stable storage.

21. Recovering from log with checkpoints: Scan backwards from end of log to find the most recent <checkpoint> record Continue scanning backwards till a record <Ti start> is found. Need only consider the part of log following above start record. Why? After that, recover from log with the rules that we had before. Checkpoints (Cont.)

22. Example of Checkpoints Tc Tf T1 T2 T3 T4 checkpoint system failure checkpoint • T1 can be ignored (updates already output to disk due to checkpoint) • T2 and T3 redone. • T4 undone

23. To permit concurrency: All transactions share a single disk buffer and a single log Concurrency control: Strict 2PL :i.e. Release eXclusive locks only after commit. Logging is done as described earlier. The checkpointing technique and actions taken on recovery have to be changed since several transactions may be active when a checkpoint is performed. Recovery With Concurrent Transactions

24. Checkpoints for concurrent transactions: < checkpointL>L: the list of transactions active at the time of the checkpoint We assume no updates are in progress while the checkpoint is carried out Recovery for concurrent transactions, 3 phases: Initialize undo-list and redo-list to empty Scan the log backwards from the end, stopping when the first <checkpointL> record is found. For each record found during the backward scan: if the record is <Ticommit>, add Tito redo-list if the record is <Ti start>, then if Ti is not in redo-list, add Ti to undo-list For every Ti in L, if Ti is not in redo-list, add Tito undo-list Recovery With Concurrent Transactions (Cont.) ANALYSIS

25. Scan log backwards Perform undo(T) for every transaction in undo-list Stop when reach <T, start> for every T in undo-list. Locate the most recent <checkpoint L> record. Scan log forwards from the <checkpoint L> record till the end of the log. perform redo for each log record that belongs to a transaction on redo-list Recovery With Concurrent Transactions UNDO REDO

26. Go over the steps of the recovery algorithm on the following log: <T0start> <T0, A, 0, 10> <T0commit> <T1start> <T1, B, 0, 10> <T2start> <T2, C, 0, 10> <T2, C, 10, 20> <checkpoint {T1, T2}> <T3start> <T3, A, 10, 20> <T3, D, 0, 10> <T3commit> Crash!!!! Example of Recovery Redo-list{T3} Undo-list{T1, T2} Undo: Set C to 10 Set C to 0 Set B to 0 Redo: Set A to 20 Set D to 10 DB A B C D Initial 0 0 0 0 At crash 20 10 20 10 After rec. 20 0 0 10

27. Recovery with Buffered Log • Log records buffered in main memory, instead of being output directly to stable storage • We output the log records when then block if full or a force operation requests that • Log force is performed to commit a transaction. Why?

28. Recovery with Buffered Log • The rules must be followed when log records are buffered: • Log records are output at the order of creation • Ti commits when the log record <Ti, commit> has been output to the log • We must use WAL • Postgres Recovery: • WAL, Pg_log