>Unless I'm heavily mistaken, RMAN knows about "skip offline"? it knows but there is allways a but: documentation states "The RECOVER DATABASE command does not recover any files that are offline normal or read-only at the point in time to which the files are being recovered." I.e. my understanding is RMAN double checks. Which is a good thing. It makes rman recovery more foul proof. Indeed "alter database datafile 4 offline drop;" for example does not prevent rman from erroring with "datafile 4 needs to be restored" --------------------------------------------------------------------------------- Please consider the environment before printing this e-mail From: Stefan P Knecht <knecht.stefan@xxxxxxxxx> To: "Laimutis.Nedzinskas@xxxxxx" <Laimutis.Nedzinskas@xxxxxx> Cc: "Mark W. Farnham" <mwf@xxxxxxxx>, "oracle-l@xxxxxxxxxxxxx" <oracle-l@xxxxxxxxxxxxx> Date: 2011.11.09 21:18 Subject: Re: Remove datafile from controlfile: recover partial database until time Unless I'm heavily mistaken, RMAN knows about "skip offline"? On 09.11.2011, at 19:12, Laimutis.Nedzinskas@xxxxxx wrote: > yes, exactly. dumping(backing in oracle terms) controlfile to trace gives > me the "create controlfile" command which I can edit to skip datafiles I do > not need to restore. > > My question is if this (removing datafile record from control file) can be > done by using individual SQL commands. > My understanding is that this is not possible: the whole tablespace can be > dropped together with datafiles but for that the database has to be opened. > > The problem is RMAN: I do not know how to feed into rman what datafiles to > skip. RMAN skips at tablespace level only which is fine but the tablespace > I deal with is a few TBs in size. > > If one looks into alert log while RMAN recovers then it is possible to see > that rman builds a list of datafiles to recover but ok, this interface is > hidden as far as I understand and the least I want is to use hidden > features unless really necessary. But it looks like RMAN operates at > tablespace level only (?) > > > > > > > > --------------------------------------------------------------------------------- > > Please consider the environment before printing this e-mail > > > > From: "Mark W. Farnham" <mwf@xxxxxxxx> > > To: <Laimutis.Nedzinskas@xxxxxx>, <oracle-l@xxxxxxxxxxxxx> > > Date: 2011.11.09 18:33 > > Subject: RE: Remove datafile from controlfile: recover partial database until time > > > > > > > Backing up the controlfile to trace gives you an ascii version you can edit > and use to create a controlfile for the partial clone recovery with the > syntax all conveniently in place. Is that what you're talking about? > > -----Original Message----- > From: oracle-l-bounce@xxxxxxxxxxxxx [mailto:oracle-l-bounce@xxxxxxxxxxxxx ] > On Behalf Of Laimutis.Nedzinskas@xxxxxx > Sent: Wednesday, November 09, 2011 8:33 AM > To: oracle-l@xxxxxxxxxxxxx > Subject: Remove datafile from controlfile: recover partial database until > time > > > Hi > > Is it possible to remove datafile entry from control file while database is > mounted ? > > I managed to achieve that by using recreating control file with "CREATE > CONTROLFILE..." > Is there another option ? > "alter database datafile ... offline drop" does not do that. It offlines > the > datafile. > > If anyone is interested why I need that I can tell: > > I develop a procedure to partially recover only some datafiles of database. > It all works fine with "alter database datafile ... offline drop" except > that sqlplus RECOVER UNTIL TIME has to be used instead of RMAN. RMAN does > not allow such recovery. It works with SKIP TABLESPACES option but my task > is more than that: I know which datafiles I have to restore exactly. > However if controlfile is recreated from scratch and the unneeded files are > omitted then RMAN recover works fine. RMAN recover has a few nice options > like auto restoring of required archive logs and deleting them on a fly. Of > course one can guess why I need all that: out of multy TB database only a > few tabless needs to be restored :-] Sure, the storage space is an > issue :-] > > > > Thank you in advance. Laimis N > > ---------------------------------------------------------------------------- > > ----- > > Please consider the environment before printing this e-mail > > -- > //www.freelists.org/webpage/oracle-l > > > > > > -- > //www.freelists.org/webpage/oracle-l > > -- //www.freelists.org/webpage/oracle-l