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: RMAN and two LOG_ARCHIVE_DEST_<n>

Re: RMAN and two LOG_ARCHIVE_DEST_<n>

From: koert54 <nospam_at_nospam.com>
Date: Sun, 01 Dec 2002 19:43:04 GMT
Message-ID: <cDtG9.30185$Ti2.5497@afrodite.telenet-ops.be>


> upgrade to 9i would be recommended, because in 9i, RMAN works perfectly
with
> RAC.
Are you running RAC production servers with RMAN ? :-) RMAN still needs a little push on RAC - especially when an instance crash occurs and
the surviving nodes performs instance recovery on the crashed node - this means that archives
coming from crashed node's thread will reside on the surviving node... this freaks out RMAN completely.
It works - but perfectly is not a word I would use :-)

"Howard J. Rogers" <howardjr2000_at_yahoo.com.au> wrote in message news:IJsG9.87856$g9.247249_at_newsfeeds.bigpond.com...
> Marc,
>
> To put it bluntly: RMAN doesn't work in a Parallel Server environment.
>
> Full stop. Period.
>
> You're doing nothing wrong, but merely encountering the reason why an
> upgrade to 9i would be recommended, because in 9i, RMAN works perfectly
with
> RAC.

>
> I know, I know... that's out of the question. In which case: get another
> backup strategy, 'cos RMAN isn't reliable in OPS.
>
> Regards
> HJR
>
> "Marc Blum" <marc_at_marcblum.de> wrote in message
> news:pspjuukohpbcc3qv6pd2dffouuvcvsq33n_at_4ax.com...
> > Hi everybody,
> >
> > situation:
> > Oracle Parallel Server 8.1.6 (test, produktion will be 8.1.7, I have no
> > possibility to upgrade the test DBs)
> >
> > each instance archives into LOG_ARCHIVE_DEST_1 and LOG_ARCHIVE_DEST_2 on
> shared
> > directories, so I have two complete sets of archived redologs
> >
> > backups are performed with RMAN and catalog db, media manager is Legato
> > Networker
> >
> >
> > requierement:
> > two separate RMANs, each with it's own Catalog DB, backup one distinct
set
> of
> > archived redologs on tape
> >
> >
> > problem:
> > the option DELETE INPUT within BACKUP raises this behaviour:
> >
> > - instance 1 generates arcs a, b, c
> >
> > - instance 2 generates arcs 1, 2, 3
> >
> > - LOG_ARCHIVE_DEST_1 contains: 1,2,3,a,b,c
> >
> > - LOG_ARCHIVE_DEST_2 cantains: 1,2,3,a,b,c
> >
> > - RMAN1 backupos ars with DELETE INPUT, one copy of the arcs is deleted,
> > everythings fine!
> >
> > - LOG_ARCHIVE_DEST_1 now contains: 1,2,3
> >
> > - LOG_ARCHIVE_DEST_2 now contains: a,b,c
> >
> > - RMAN1 backups again with DELETE INPUT, now the remaining set of arcs
is
> > written to tape and deleted
> >
> > - LOG_ARCHIVE_DEST_1 now contains:
> >
> > - LOG_ARCHIVE_DEST_2 now contains:
> >
> > - RMAN2 is unable to backup his arcs!
> >
> >
> > question:
> > why doesn't RMAN1 recognise the already backuped arcs? Is it a bug or
> well-known
> > behaviour?
> >
> > many thanks in advance
> >
> >
> > (I found a workaround by opening two channels and assigning each channel
> to a
> > distinct LOG_ARCHIVE_DEST_<n>, but I would prefer a solution, where I
> don't have
> > to tell RMAN the LOG_ARCHIVE_DEST explicitely)
> >
> >
> > Marc Blum
> > mailto:marc_at_marcblum.de
> > http://www.marcblum.de
>
>
Received on Sun Dec 01 2002 - 13:43:04 CST

Original text of this message

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