Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: relation bet REDO LOG FILE and BEGIN BACKUP

Re: relation bet REDO LOG FILE and BEGIN BACKUP

From: joel garry <joel-garry_at_home.com>
Date: 11 Jul 2006 15:12:07 -0700
Message-ID: <1152655927.813931.317140@35g2000cwc.googlegroups.com>

Jonathan Lewis wrote:
> "sybrandb" <sybrandb_at_yahoo.com> wrote in message
> news:1152605925.870552.276390_at_m79g2000cwm.googlegroups.com...
> >
> > peter wrote:
> >> Hello DBAs,
> >>
> >> when we issue
> >> alter tablespace <tb_name> begin backup;
> >> what extra information go in redo log, why archive is fast and if there
> >> is shutdown abort why backup at that time is not useful........
> >>
> >> Because in books it was mentioned that archive is fast because all
> >> information of tablespace would go in redo log but even in simple way,
> >> everything goes in redo log so what extra..........
> >
> > According to the documentation, begin backup forces Oracle *for every
> > change* to write the complete block to redolog, instead of changed
> > bytes only.
> >
> >
> > --
> > Sybrand Bakker
> > Senior Oracle DBA
> >
>
> If that's what the manual says, it's wrong.
> 'begin backup' forces a tablespace checkpoint.
>
> After the checkpoint is complete, the first change
> to a block after it has been loaded into the cache
> (or the first change for any block already in the
> cache) causes a complete block write of the current
> version of the block, followed by the normal redo.
> Subsequent changes __while the block stays in the
> buffer__ do NOT cause a further block write.

asktom points out the error here:
http://asktom.oracle.com/pls/ask/f?p=4950:8:::::F4950_P8_DISPLAYID:271815712711 (search for the word "proof")

jg

-- 
@home.com is bogus.
http://jilltxt.net/txt/FeralHypertext.pdf
Received on Tue Jul 11 2006 - 17:12:07 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US