Re[2]: About standby redo logs

  • From: Kamus <kamusis@xxxxxxxxx>
  • To: greg.loughmiller@xxxxxxxxxxxx
  • Date: Thu, 06 Jan 2005 10:59:17 +0800

Oracle document said very clearly.
http://download-west.oracle.com/docs/cd/B10501_01/server.920/a96653/log_tra=
nsport.htm#1038566

To minimize data loss in the event of a primary database failure, you want =
to copy data from the primary database to the standby database as it is bei=
ng generated. You can choose to have either the log writer process or the a=
rchiver process transmit redo logs to a destination.
=2E...
Choosing the LGWR attribute indicates that the log writer process (LGWR) wi=
ll transmit redo data to the associated destination as it is generated. As =
redo is generated for the primary database, it is also propagated to the st=
andby system where the RFS process writes the redo to either a standby redo=
 log or to a standby archived redo log.



On Wed, 5 Jan 2005 12:52:14 -0600=20
"Loughmiller, Greg" <greg.loughmiller@xxxxxxxxxxxx> wrote:

> If I may ask a similar question, yet not exactly related to this thread..
> What are the advantages of using LGWR as compared to the ARCH process for
> the transport mechanism of the redo data to the standby database?
>=20
> Thanks
> Greg
> -----Original Message-----
> From: Carel-Jan Engel [mailto:cjpengel.dbalert@xxxxxxxxx]=20
> Sent: Wednesday, January 05, 2005 11:55 AM
> To: oracle-l@xxxxxxxxxxxxx
> Subject: Re: About standby redo logs
>=20
> Kamus,
> Oracle documents that one needs at least the # of redo log groups on the
> primary + 1 on the standby, when standby redologs are used. I never saw
> a logswitch happen to another standby redologfile than the one
> previously used, so your observation is quite normal, though somewhat
> unexpected regarding the documented requirements. DG always seems to
> reuse the same logfile.  I guess it has to do with the speed of the ARCH
> on the standby. I need to investigate this further if I get a chance on
> a testsystem, before I can get more conclusive on this.
>=20
> Because normally DG-configurations have real HA requirements (if not,
> why was DG implemented), I tend to stay on the safe side and follow the
> Oracle docs. It's a little bit CYA, I admit, so be it.
>=20
> Best regards,
>=20
> Carel-Jan Engel
>=20
> =3D=3D=3D
> If you think education is expensive, try ignorance. (Derek Bok)
> =3D=3D=3D
>=20
> Upcoming appearances:=20
>=20
>       * Jan 27, 2005: London, UKOUG Unix SIG: Data Guard Best Practices=
=20
>       * Feb 9-10, 2005: Denver, RMOUG Training Days: Data Guard
>         Performance Issues=20
>       * Mar 6-10, 2005: Dallas, Hotsos Symposium: Data Guard Performance
>         Issues
>=20
>=20
> On Wed, 2005-01-05 at 15:33, Kamus wrote:=20
>=20
> > I create 4 standby redo log groups, and use LGWR in primary site to
> > transfer redo data, all are good.=20
> > But when I query the V$STANDBY_LOG view,
> > I found that only the status column of GROUP# 4(the first group of my
> > standby redo logs) value is "ACTIVE" while all the others(5-7) are
> > "UNASSIGNED".
> > also, only GROUP# 4's sequence# is a value greate than 0
> > and THREAD# is 1, all others are 0 and 0.
> >=20
> > Any explains.
>=20
>=20
>=20
>            =20
>       *=20
>=20
>=20
> --
> //www.freelists.org/webpage/oracle-l
>=20
>=20
> --
> //www.freelists.org/webpage/oracle-l

--=20
Kamus <kamusis@xxxxxxxxx>

=C4=C7=C3=B4=B6=E01G=B5=C4=D3=CA=CF=E4=A3=AC=CE=D2=C4=DC=D3=C3=C0=B4=B8=C9=
=CA=B2=C3=B4:-)
A Oracle8i & 9i Certified DBA from China

--
//www.freelists.org/webpage/oracle-l

Other related posts: