Re: sqlplus shutdown "time-out"

From: Vitalis <vitalisman_at_gmail.com>
Date: Fri, 22 Apr 2005 13:15:33 +0200
Message-ID: <68b1285505042204152cb6343a@mail.gmail.com>

Hi Jared,

I agree with your description, but the topic is about shutting down an instance before a cold backup i.e. issuing "shutdown immediate" as soon as we get to the "logon" step you described. And by using "shutdown abort" as the first step for this backup purpose, the whole procedure would be slower.

Regards,
Jerome

On 4/22/05, Jared Still <jkstill_at_gmail.com> wrote:
> When you shutdown abort, Instance Recovery requires
> rolling forward through the redo before opening the database.
> =20
> Rolling back the undo is done in the background.
> =20
> You can shutdown abort, startup and logon to the database
> faster than you can shutdown immediate, startup and logon
> to the database.
> =20
> Jared

>=20

> =20
> =20
> On 4/21/05, Vitalis <vitalisman_at_gmail.com> wrote:
> > On 4/21/05, Jared Still <jkstill_at_gmail.com> wrote:
> > > On 4/20/05, Vitalis <vitalisman_at_gmail.com> wrote:
> > > >
> > > > (why use "abort" and force a recovery everytime while we could use=
=20
> > > "immediate"
> > > > in most cases?)
> > > Because it is faster.
> >=20
> > I don't think so, if we're talking about delays caused by active
> transactions.
> >=20
> > The recovery would have to roll forward the updates that the initial=20
> > transaction(s)
> > had made; then Oracle would have to roll back the work again...
> > AFAIK it would take longer because of the recovery.
> >=20
> > Jerome
> >=20

>=20
>=20
>=20

> --=20
> Jared Still
> Certifiable Oracle DBA and Part Time Perl Evangelist=20
>=20
>

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

Received on Fri Apr 22 2005 - 07:19:43 CDT