| Nologging [message #159246] |
Fri, 17 February 2006 00:31  |
gajini
Messages: 262
Registered: January 2006
|
Senior Member |
|
|
I read the following sentences in setting Nologging mode, "Redo logs do not contain the values that were inserted when the table was in Nologging mode,therefore the datafile pertaining to the table or partition should be backed up immediately upon completion of the direct-load operation"
My doubt in this is, Why the datafile should be immediately backed up?
|
|
|
|
| Re: Nologging [message #159256 is a reply to message #159246] |
Fri, 17 February 2006 00:54  |
deepa_balu
Messages: 74
Registered: March 2005
|
Member |
|
|
As per my knowledge...
Suppose ur database goes down and u need to perform a recovery operation by applying the redologs/archive logs...whatever operation using nologging mode will not be recovered when applying the redologs if u are using a backupprior to the nologging operation.
So in case if u backing up the datafile after a nologging operation, u can use that datafile during recovery so that the nologging data is available.
|
|
|
|
| Re: Nologging [message #159285 is a reply to message #159256] |
Fri, 17 February 2006 03:09 |
gajini
Messages: 262
Registered: January 2006
|
Senior Member |
|
|
Thanks Deepa_balu..
I'm new to DBA & I'm very confused with the concept of Checkpoint occurrence and SCN.Can you explain me these concepts briefly in layman terms?
Please reply asap,so that I'll get clear idea about this.
|
|
|
|
| Re: Nologging [message #159685 is a reply to message #159285] |
Mon, 20 February 2006 23:16 |
deepa_balu
Messages: 74
Registered: March 2005
|
Member |
|
|
This is from metalink....
Here is a brief overview of what occurs that may help your understanding. Oracle records every change in a redo record, which is an entry in the redo buffer describing what has changed. Each atomic change in a redo record is called a change vector. A redo record is composed of all the change vectors involved in a change. If the update is committed, Oracle generates another redo record and assigns the change an SCN.
Each instance is associated with a 'thread' of redo. A checkpoint structure contains information describing the the position in the redo thread from which crash or instance recovery would need to begin for the database (as all redo entries recorded before this point in the online redo log are already written to the datafiles). The information from this structure is stored in the datafile headers and in the controlfile.
Beginning in Oracle8, the dirty buffers in the buffer cache are linked on a queue called the checkpoint queue. This queue contains dirty, logged buffers ordered according to the position in the redo log corresponding to the first change to the buffer. A Redo Byte Address (RBA) points to a specific location in a particular redo thread. In response to a checkpoint request, DBWR writes out all
the buffers on this queue in ascending low RBA order. Every checkpoint request specifies an RBA. Once DBWR writes out buffers whose low RBAs are equal to or greater than the checkpoint's RBA, the checkpoint is declared complete, and the relevent information is recorded in the control file and datafile headers. The information stored includes the checkpoint SCN, thread number, and RBA.
Checkpoint is the only operation which updates this structure in the controlfile and datafile headers. Recovery uses the information in this structure to determine what change recovery needs to begin with.
|
|
|
|
] 
Blog
Search
Help
Members
Register
Login
Home
