Re: ACID
- From: saurabh manroy <saurabhmanroy@xxxxxxxxx>
- To: Laimutis.Nedzinskas@xxxxxx
- Date: Wed, 24 Aug 2011 13:20:02 +0200
Hi Laimutis,
Even Oracle is using WAL mechanism. Changes are first written to logfile
before flushing the change to datafiles on disk.
For DB2, I see that its incorporated that select query from another session
has to wait until User commit is 'complete'.
http://webdocs.caspur.it/ibm/db2/8.1/doc/htmlcd/en_US/index.htm?openup=admin/c0005425.htm
" The data in the log buffer is written to disk by the logger process. In
the following circumstances, query processing must wait for log data to be
written to disk:
- On COMMIT
- Before the corresponding data pages are written to disk, because
DB2(R) uses
write-ahead logging. The benefit of write-ahead logging is that when a
transaction completes by executing the COMMIT statement, not all of the
changed data and index pages need to be written to disk.
- Before some changes are made to metadata, most of which result from
executing DDL statements
- On writing log records into the log buffer, if the log buffer is full"
On Wed, Aug 24, 2011 at 12:03 PM, <Laimutis.Nedzinskas@xxxxxx> wrote:
>
> Implementation of durability by Oracle deserves some discussion.
>
> The term REDO is may be fine but now days there exists a better term:
> Write-ahead logging. Three words tell the essence: write to log ahead of
> modifications.
>
> Obviously for some reasons (a few of them can be named actually) oracle
> implemented the commit command using Write-after logging. One wonders what
> about rollback ? For me, Rollback seems to be fine even with write-after
> logic.
>
>
>
>
>
>
>
>
>
>
> ---------------------------------------------------------------------------------
>
> Please consider the environment before printing this e-mail
> ----- Forwarded by Laimutis Nedzinskas/VBankas on 2011.08.24 12:45 -----
>
> From: Laimutis Nedzinskas/VBankas
>
> To: jonathan@xxxxxxxxxxxxxxxxxx
>
> Cc: "Oracle L" <oracle-l@xxxxxxxxxxxxx>,
> oracle-l-bounce@xxxxxxxxxxxxx
>
> Date: 2011.08.24 12:37
>
> Subject: Re: ACID
>
>
>
>
>
> http://en.wikipedia.org/wiki/ACID
>
> There is one property which has very special comments about it:
>
> " Isolation ...
> This property of ACID is often relaxed (i.e. partly respected) because of
> the huge speed decrease this type of concurrency management implies.
> ...
> In reality, many alternatives are used due to speed concerns, but none of
> them guarantees the same reliability. "
>
> Could not be said better.
>
>
>
>
> ---------------------------------------------------------------------------------
>
> Please consider the environment before printing this e-mail
>
>
>
> From: "Jonathan Lewis" <jonathan@xxxxxxxxxxxxxxxxxx>
>
> To: "Oracle L" <oracle-l@xxxxxxxxxxxxx>
>
> Date: 2011.08.23 11:54
>
> Subject: ACID
>
>
>
>
>
>
>
> So who knows the meaning of ACID (not the Timothy Leary kind) ?
>
> http://tonyhasler.wordpress.com/2011/08/22/why-is-acid-important/
> http://jonathanlewis.wordpress.com/2011/08/19/redo-2/
>
>
> Regards
>
> Jonathan Lewis
> http://jonathanlewis.wordpress.com
>
>
> --
> //www.freelists.org/webpage/oracle-l
>
>
>
>
>
>
> --
> //www.freelists.org/webpage/oracle-l
>
>
>
Other related posts:
