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: Hot Backup referenced external data

Re: Hot Backup referenced external data

From: <dickweisinger_at_my-deja.com>
Date: Mon, 20 Nov 2000 08:25:07 GMT
Message-ID: <8van51$ro0$1@nnrp1.deja.com>

Hello Howard and Jake,

You are right that I don't know that much about backups, but I think you didn't understand my question.
We are backing up data from a system that uses Oracle for archiving electronic documents and images (MS Office files, large TIFF images, ...).
When a document is checked in, Oracle is updated with metadata for the document. The file(s) that make up the document are also moved to a secure volume and there is a column in an Oracle table that has a reference to this location. The document file(s) are not stored as blobs in the DB -- they reside outside the DB. When we do an incremental backup, we are backing up the Oracle Metadata, but we'd also like to be able to backup the document files on the volume referenced in the Oracle table. Right now, a backup of these files is being done based on file timestamps. If we can parse out the path information from the table in the incremental backup data, then we can backup these "extra files". I will try to find out more information about Log Miner. Thanks
Dick

In article <3a132bb2_at_news.iprimus.com.au>,   "Howard J. Rogers" <howardjr_at_www.com> wrote:
> Hi jake: good questions. No, I think RMAN is an excellent utility.
 But I'd
> be using it because (a) the scripting language means that what works
 can be
> preserved, and protected behind the usual Oracle security mechanisms,
 (b)
> the fact that hot backups don't generate block-level backups, and
 thus the
> redo sub-system is not overly stressed, (c) recoveries become
 relatively
> simple affairs 'restore database - recover database' -I've seen too
 many
> restores stuffed up by failure to remember precisely what gets
 restored etc
> etc and (d)the fact that it's an Oracle utilitiy means it understands
 the
> concept of an Oracle block -and can report on any corruptions
 straight away
> (no need to remember to dbverify the results). You've only got to
 recall
> some of the posts here a few weeks back about 'three weeks of backup
 turns
> out to have included corrupt data files' to realise how important that
> little feature is.
>
> I think the incremental possibilities come in around point (e). It
 can make
> all the difference between achieving a backup in the available
 maintenance
> window, and not.
>
> I confess to having been utterly intimidated by the syntax from Hell
 with
> which RMAN is blessed (cursed?) in my early days with it. But that's
 just
> another language to learn, and given enough time, it's not *that*
 bad, with
> hindsight. I therefore recommend RMAN with conviction these days.
 Indeed,
> just last month, for the first time, one of my students did the entire
> Backup and Recovery workshop using nothing but RMAN. I was somewhat
> gobsmacked, but also deeply impressed, and persuaded that this (or
 something
> like it) ought to be the way of the future as regards backup and
 recovery.
>
> Regards
> HJR
> --
> ----------------------------------------------------------------------



> Opinions expressed are my own, and not those of Oracle Corporation
> Oracle DBA Resources:

 http://www.geocities.com/howardjr2000
> ----------------------------------------------------------------------


>
> "Jake" <jjstrauss_at_yahoo.com> wrote in message
> news:T%wQ5.13583$3u4.67355_at_e420r-sjo2.usenetserver.com...
> > > I mentioned earlier. Unless you are talking huge amounts of
 data, the
> > > incremental backup feature is not the first one that springs to
 mind as
 a
> > > reason for moving to RMAN, however.
> >
> > Howard, what is your first reason for moving to RMAN, or are you
 saying
 you
> > prefer not to use RMAN?
> >
> > --
> > Jake
> > "Howard J. Rogers" <howardjr_at_www.com> wrote in message
> > news:3a111d39$1_at_news.iprimus.com.au...
> > > I'm slightly confused about what you are after, but here goes for
> > > nothing....
> > >
> > > 1. How can we best backup data associated with a hot backup?
> > >
> > > I *really* don't understand this question. A hot backup is a
 backup of
 the
> > > data files associated with a database. You want to backup a
 backup? If
 you
> > > mean, 'how do you do hot backups', the short answer is: do it a
 tablespace
> > > at a time, issue the command 'alter tablespace blah begin backup;
 take a
> > > host copy of the data files associated with that tablespace; when
 the
 copy
> > > has finished, issue the command 'alter tablespace blah end
 backup'. You
> > > then move on to the next tablespace and so on, until all
 tablespaces
 have
> > > been processed. Full instructions are contained (amongst other
 places)
 in
> > > my backup and recovery document, available on my website.
> > >
> > > 2. I think you then ask how (or whether) you can just backup the
 extra
> > > column that has been added to a table. Short answer is, you
 can't.
 Even
> > > using RMAN's incremental backup feature, since that is row-level
 sensitive
> > > to data changes, not column-sensitive.
> > >
> > > 3. Yes, you can 'parse the data' from your archives, provided
 you have
> > > 8.1.5 or above -the feature is called Log Miner (also 'parses
 data' from
> > > on-line archive logs). Anything earlier than 8.1.5, and you'll
 have to
 find
> > > a third-party product that does the deed (and I don't know of any
 off
 hand).
> > >
> > > 4. 'Can we read incremental RMAN data'? No, but (rather
 fortunately)
 RMAN
> > > can. If you mean, 'can RMAN do incremental backups', the answer
 is yes,
 as
> > > I mentioned earlier. Unless you are talking huge amounts of
 data, the
> > > incremental backup feature is not the first one that springs to
 mind as
 a
> > > reason for moving to RMAN, however.
> > >
> > > Not sure whether you are entirely familiar with backup and
 recovery
 concepts
> > > and techniques. L=ike Isay, go to my site, find the Backup and
 Recovery
> > > course page, and download my small document there -and then get
 back to
 me
> > > with any questions you may have.
> > >
> > > Regards
> > > HJR
> > > --
> >
> > --------------------------------------------------------------------


> > -
> > > Opinions expressed are my own, and not those of Oracle Corporation
> > > Oracle DBA Resources:

 http://www.geocities.com/howardjr2000
> >
> > --------------------------------------------------------------------


> > -
> > >
> > >
> > > <dickweisinger_at_my-deja.com> wrote in message
> > > news:8uqq5f$e1p$1_at_nnrp1.deja.com...
> > > > Hi,
> > > > Can you please help me?
> > > > How can we best backup data associated with a hot
> > > > backup?
> > > > We have a table that references volume and file
> > > > data. As part of the hot backup, we want to do
> > > > an incremental backup of all files that are newly
> > > > referenced in the incremental hot database backup.
> > > > (Someone added a column to the table with path to
> > > > a file.)
> > > > How can we best do this? If we keep an
> > > > ARCHIVELOG, can we parse the data from that
> > > > file? We may move to RMan in the near future --
> > > > can we read incremental RMan data?
> > > > Thanks
> > > >
> > > >
> > > >
> > > >
> > > > Sent via Deja.com http://www.deja.com/
> > > > Before you buy.
> > >
> > >
> > >
> >
> >
> >
> >
>
>

Sent via Deja.com http://www.deja.com/
Before you buy. Received on Mon Nov 20 2000 - 02:25:07 CST

Original text of this message

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