Re: Do dirty buffers and redo log buffers contain same information ?

"Akinwande Seigmund Walter-Johnson III" <asjohns_at_midway.uchicago.edu> wrote in message news:Lcts6.179$E4.9551_at_uchinews...
> What is the difference in content ? , I know redo log buffers contain the
> actual statement changes

Sort of. They contain the sort of information you'd get out of v$logmnr_contents if you were using Log Miner -the before and after image of all DML statements (hence, the sql_redo and sql_undo columns), and a transaction identifier (serial_number in v$logmnr_contents), and a rowid where the transaction is being applied.

A dirty buffer simply means a data block in memory that doesn't agree with the corresponding data block on disk (because some form of transaction is taking place (or has already taken palce)). So 'what's in a dirty buffer?' is really just a way of saying 'what's in a data block?'.

Data blocks contain multiple entire rows (when redo is generated, it's only the actual *fields* being changed that are logged). There are no rowids. There's the block header and its transaction slots. There are lock bytes.

It's really like trying to compare apples and oranges. At the end of the day, if you looked at a dirty buffer, it would look like a piece or selection of a table. If you looked at the log buffer, it would look like a list of terse instructions to change bits of data in all sorts of tables.

HJR
> "Howard J. Rogers" <howardjr_at_www.com> wrote in message
> news:3aa022f9_at_news.iprimus.com.au...
> > No
> >
> > "Akinwande Seigmund Walter-Johnson III" <asjohns_at_midway.uchicago.edu>
 wrote
> > in message news:4Wtn6.101$w4.1549_at_newsfeed.uchicago.edu...
> > >
> > >
> > >
> >
> >
>
>
Received on Sun Mar 18 2001 - 06:26:31 CST