Home
|
Previous Page
|
Next Page
Logging and Log Administration
>
Checkpoints and Fast Recovery
In This Chapter
How the Database Server Achieves Data Consistency
Critical Sections
Checkpoints
Full Checkpoint
Fuzzy Checkpoint
Fuzzy Operations
Write-Ahead Logging and Fast Recovery
Fuzzy Checkpoints Improve Performance
Events That Initiate a Fuzzy Checkpoint
Events That Initiate a Full Checkpoint
Forcing a Full Checkpoint
Forcing a Fuzzy Checkpoint
Sequence of Events in a Checkpoint
User Threads Cannot Enter a Critical Section
Logical-Log Buffer Is Flushed to the Logical-Log File on Disk
Page-Cleaner Thread Flushes the Physical-Log Buffer
Page-Cleaner Threads Flush Modified Pages in the Buffer Pool
Checkpoint Thread Writes Checkpoint Record
Physical Log Is Logically Emptied
Backup and Restore Considerations
Time Stamps
Time Stamps on Disk Pages
Time Stamps on Logical-Log Pages
Fast Recovery
Need for Fast Recovery
Situations When Fast Recovery Is Initiated
Fast Recovery and Buffered Logging
Fast Recovery and No Logging
Details of Fast Recovery After A Full Checkpoint
Returning to the Last-Checkpoint State
Finding the Checkpoint Record in the Logical Log
Rolling Forward Logical-Log Records
Rolling Back Incomplete Transactions
Details of Fast Recovery After A Fuzzy Checkpoint
Returning to the Last-Checkpoint State for Nonfuzzy Operations
Locating the Oldest Update in the Logical Log
Applying the Log Records for Fuzzy Operations
Rolling Forward Logical-Log Records
Rolling Back Incomplete Transactions
Fast Recovery of Tables
Monitoring Checkpoint Information
Using onstat Options
Using SMI Tables
Home
| [
Top of Page
|
Previous Page
|
Next Page
|
Contents
|
Index
]