Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Mailing Lists -> Oracle-L -> RE: Rebuilding indexes

RE: Rebuilding indexes

From: Jeremiah Wilton <jwilton_at_speakeasy.net>
Date: Fri, 01 Jun 2001 10:41:18 -0700
Message-ID: <F001.00318188.20010601094714@fatcity.com>

On Fri, 1 Jun 2001, Hatzistavrou Giannis wrote:

> I don;t quite understand you quoting
> "Once the indexes have all
> been rebuilt nologging into the original tablespace, all that's left
> is to
> refresh the datafiles for that tablespace onto the primary. "
> Is primary a typo error (i.e. standby)?

Argh! It was a typo. Of course, after doing many unrecoverable operations, you should refresh the affected datafiles from the primary to the standby.

Vadim Gorbounov mentioned the column 'unrecoverable_change#' in v$datafile. This looks like an SCN, not a count of unrecoverable changes. Either way, it seems like a useful way to decide if a datafile shoule be refreshed onto the standby. I can even imagine improving a standby log applier that could be made to refresh such files from the primary automatically.

On another topic, I believe I qualify for some sort of award as uber geek. This week I obtained the Washington State license plate "ORA DBA" for my '74 Volvo.

It got me thinking that it would be fun to have some bumper stickers made up for members of this list. Some ideas I had:

"ARE YOU AN IDIOT?" or

"SHUTDOWN ABORT"

--
Jeremiah Wilton
http://www.speakeasy.net/~jwilton



> > -----Original Message-----
> > From: Jeremiah Wilton [SMTP:jwilton_at_speakeasy.net]
> > Sent: Thursday, May 31, 2001 22:15
> > To: Multiple recipients of list ORACLE-L
> > Subject: RE: Rebuilding indexes
> >
> > No need to copy the whole file to the standby when adding a datafile!
> > Just wait
> > for recovery to fail on the new file (ORA-01670), then type on the standby
> > (in
> > 'mount standby database' mode):
> >
> > SQL> alter database create datafile '<filename>' as '<filename>';
> >
> > As long as you are using identical paths on both primary and standby, both
> > filenames are the same. Our standby log applier does this automatically,
> > so no
> > manual intervention is required.
> >
> > If you use logical volumes or raw, just make sure the volume is where the
> > system
> > expects it to be on the standby.
> >
> > If the original poster wants to create a scratch tablespace where the
> > indexes
> > will be initially rebuilt, then move the indexes back into the "real"
> > tablespace, he can probably just 'offline drop' the datafiles for the
> > scratch
> > space on the standby when recovery encounters them. Once the indexes have
> > all
> > been rebuilt nologging into the original tablespace, all that's left is to
> > refresh the datafiles for that tablespace onto the primary. When the
> > scratch
> > tablespace is dropped on the primary, the offline dropped datafile records
> > in
> > the standby's controlfile will also disappear.
> >
> > With regard to logging/nologging, with index rebuilds, I would rebuild the
> > indexes NOLOGGING, and when you are all done, refresh the datafiles for
> > the
> > newly reorganized tablespace onto the standby. That saves on space,
> > operations,
> > time and reduces the logs you have to back up.
> >
> > Nologging operations are not harmful to the standby - you just have to
> > keep
> > track of nologging operations and refresh datafiles on the standby
> > accordingly.
> >
> > --
> > Jeremiah Wilton
> > http://www.speakeasy.net/~jwilton
> >
> > On Thu, 31 May 2001, Rachel Carmichael wrote:
> >
> > > It adds the datafile NAME only, not the file itself. And if the physical
> > > file has not been moved before the archived log that has that redo in it
> > is
> > > applied, the recovery dies.
> > >
> > > >From: Hatzistavrou Giannis <j.hatzistavrou_at_telesoft.gr>
> > > >
> > > >You might be mistaken, since Oracle Manual (Oracle 8.1.5 Backup and
> > > >Recovery) , chapter 16 -- Managing a standby database --- Adding
> > Datafiles
> > > >states the following:
> > > >
> > > >"Adding a datafile to your primary database generates redo data that,
> > when
> > > >applied at your standby, automatically adds the datafile name to the
> > > >standby
> > > >control file....."
> > > >There is also a detailed example on how to do it
> > > >
> > > > > -----Original Message-----
> > > > > From: Richard Huntley [SMTP:rhuntley_at_mindleaders.com]
> > > > >
> > > > > For those of you that have implemented a standby database, what
> > method
> > > >do
> > > > > you use to rebuild your indexes tablespace.
> > > > > My plan was to create a new tablespace and rebuild the indexes into
> > the
> > > > > new
> > > > > tablespace and then reverse the process to move
> > > > > back to the original tablespace after I drop and recreate it in
> > order to
> > > > > get
> > > > > rid of fragmentation in the originally indexes tablespace.
> > > > > However, creating the new datafile associated with the new
> > tablespace,
> > > > > cancels media recovery associated with the standby database...just
> > > > > wondered
> > > > > if anyone has a better method of defragging an index tablespace when
> > > >there
> > > > > is a standby database catching the redo it generates.
-- Please see the official ORACLE-L FAQ: http://www.orafaq.com -- Author: Jeremiah Wilton INET: jwilton_at_speakeasy.net Fat City Network Services -- (858) 538-5051 FAX: (858) 538-5051 San Diego, California -- Public Internet access / Mailing Lists -------------------------------------------------------------------- To REMOVE yourself from this mailing list, send an E-Mail message to: ListGuru_at_fatcity.com (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
Received on Fri Jun 01 2001 - 12:41:18 CDT

Original text of this message

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