ch12 - Chapter 12 Distributed DBMS Reliability Chapter 12...

Info iconThis preview shows pages 1–7. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Chapter 12 Distributed DBMS Reliability Chapter 12 Distributed DBMS Reliability Reliability of DDB refers to (1) atomicity, and (2) durability. Specific reliability protocols related include (1) commit protocols, and (2) recovery protocols. 121 RELIABILITY CONCEPTS AND MEASURES 122 FLAILURES AND FAULT TOLERANCE IN DISTRIBUTED SYSTEMS 123 FAILURES IN DISTRIBUTED DBMS 12.31 Transaction Failures Incorrect input data, The detection of deadlock …. can cause this type of failures. 12.32 Site (System) Failures 12.31 Media Failures 12.32 Communication Failures 12.4 LOCAL RELIABILITY PROTOCOLS 1 DB Buffers (Volatile DB) Chapter 12 Distributed DBMS Reliability LRM (Local Recovery Manager) maintain the atomicity and durability properties of local transactions by performing some functions. LRM accepts “Begin_transaction”, “read”, “write”, “commit”, and “abort” commands. 12.41 Architectural Considerations Commands from scheduler Main memory Fetch Flush Stable Read Write DB Write Read Interfaces between LRM, Buffer and DB LRM executes operations only on the volatile DB. Buffers are organized in pages. 2 Chapter 12 Distributed DBMS Reliability Fetch command Get a page. If the page is in DB buffers then the Buffer Manager returns it; otherwise the buffer Manager reads it from the Stable DB and puts it in buffers. An algorithm is needed if the buffer is full. Flush command Force pages to be written from buffers to the stable DB. 12.42 Recovery Information In-place update – physically change the value of data items in stable DB. The previous values are lost. Out-of-place update – do not change the value of data items in stable DB but maintain the new value separately. Most DBMS use in-place update for better performance. In-place update recovery information Recovery information are kept in DB log. Each 3 New stable DB state Old stable DB state Old stable DB state New stable DB state Chapter 12 Distributed DBMS Reliability update not only changes DB, but also saves records in DB log. Update An example of system failure Begin T 1 Commit Begin T 2 t Assume buffer pages are written back to the stable DB only when Buffer Manager needs new buffer space. T1: from user’s point of view it is committed. But updated buffer pages may get lost. REDO T 1 4 New stable DB state Old stable DB state Old stable DB state New stable DB state Chapter 12 Distributed DBMS Reliability REDO T 2 : not terminated, but some updated pages may have been written to the stable DB. Undo is needed. UNDO T 2 UNDO Both REDO and UNDO are idempotent. Contents of DB log – transaction ID, begin/end of transaction, before image, after image, etc. Log file maintenance Logging Interface 5 Log Buffers DB Buffers (Volatile DB) Chapter 12 Distributed DBMS Reliability Commands from scheduler Main memory Log Read Write Fetch Read Flush Write Stable Read Write DB Write Read Interfaces between LRM, Buffer and DB and Log file Two ways to write log pages 1)Synchronous (forcing a log) – the adding of each...
View Full Document

This note was uploaded on 12/23/2009 for the course DBST 663 taught by Professor Tba during the Spring '09 term at MD University College.

Page1 / 37

ch12 - Chapter 12 Distributed DBMS Reliability Chapter 12...

This preview shows document pages 1 - 7. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online