From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mx1.redhat.com (mx1.redhat.com [172.16.48.31]) by int-mx1.corp.redhat.com (8.11.6/8.11.6) with ESMTP id k1LALI126146 for ; Tue, 21 Feb 2006 05:21:18 -0500 Received: from moutng.kundenserver.de (moutng.kundenserver.de [212.227.126.183]) by mx1.redhat.com (8.12.11/8.12.11) with ESMTP id k1LALCFI006610 for ; Tue, 21 Feb 2006 05:21:13 -0500 Message-ID: <43FAE99A.4060609@gmx.net> Date: Tue, 21 Feb 2006 11:21:14 +0100 From: Klaus Strebel MIME-Version: 1.0 Subject: Re: [linux-lvm] Using LVM to snapshot Oracle References: <20060220220531.GA16484@percy.comedia.it> <20060220221114.GB16484@percy.comedia.it> In-Reply-To: <20060220221114.GB16484@percy.comedia.it> Content-Transfer-Encoding: quoted-printable Reply-To: LVM general discussion and development List-Id: LVM general discussion and development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , List-Id: Content-Type: text/plain; charset="iso-8859-1"; format="flowed" To: LVM general discussion and development Luca Berra schrieb: > On Mon, Feb 20, 2006 at 11:05:32PM +0100, Luca Berra wrote: > to clarify: >> you can do hot backup, but you will need to tell oracle you are doing >> so: >> >> you will need to have oracle in archive log mode >> >> you will need to manually set all tablespaces in backup mode: >> alter tablespace XXX begin backup; (for all tablespaces) >> alter system switch logfile; > the snapshot creation part goes here >=20 > NO >do your backup including the archive logs >=20 >> alter tablespace XXX end backup; (for all tablespaces) >=20 > do your backup here including the archive logs > destroy the snapshot here. >=20 >> it is a simple sql script to code >> >=20 Hi Mitchell, hi Luca, well, despite i prefer using RMAN ( since 9iR2 it's once configured and=20 then you can always do your backup with a simple 'backup database' ) but=20 you need your database running in archived mode, which brings additional=20 space and backup needs. The way Luca describes for a online-backup is the usual way that's=20 working since Oracle 7 ( or even 6, but just work with Oracle since 7.0=20 ;-) ), but i think, if all ( really all !) files are on the same=20 snapshout and the snapshot is big enough( ;-) ), you should make a=20 logswitch, create a readonly snapshot, copy all the files ( including=20 control-files, and redo-logs ) and remove the snapshot. You can then=20 carefully try to use these files to restore the database. If you have=20 enough diskspace, you can shutdown your instance, rename all the files (=20 again incl. control-files and redo-logs ), restore the copies from the=20 snapshot and start the database carefully ( startup nomount; alter=20 database mount; select file#, name, status from v$datafile; -> all file=20 should be in status system or online ; alter database open; -> up again ). I have nither the time nor the resources to test this on my own. But you=20 can share your results with us :-) Ciao Klaus --=20 Mit freundlichen Gr=EF=BF=BDssen / best regards Klaus Strebel, Dipl.-Inform. (FH), mailto:klaus.strebel@gmx.net /"\ \ / ASCII RIBBON CAMPAIGN X AGAINST HTML MAIL / \