Re: Oracle recommends rebuilding IOTs in AQ to reduce redo
- From: "Niall Litchfield" <niall.litchfield@xxxxxxxxx>
- To: greg@xxxxxxxxxxxxxxxxxx, jonathan@xxxxxxxxxxxxxxxxxx, ryan_gaffuri@xxxxxxxxxxx, oracle-l@xxxxxxxxxxxxx
- Date: Mon, 7 Jan 2008 07:41:26 +0000
It's interesting to me that a coalesce is what is happening for
multiple subscriber queues, but not single and that there is a
compatible check as well. It rather suggests to me that some thought
has gone into the procedure - perhaps as a result of relevant bugs.
On 07/01/2008, Greg Rahn <greg@xxxxxxxxxxxxxxxxxx> wrote:
> The coalesce would be a better option IMO (I've personally used this
> as a recommendation from the AQ dev team) and it can be done while the
> workload is in progress, while a rebuild requires the queues to be
> stopped.
>
> Another recommendation I received is *not* to use ASSM tablespaces for
> the queues.
>
> What db version is this?
> What is the queue navigation type (first or next message)?
> Are you doing single row or array enqueues/dequeues?
>
>
> On 1/6/08, Jonathan Lewis <jonathan@xxxxxxxxxxxxxxxxxx> wrote:
> > In simple indexes, the optimum solution to this type of problem
> > is usually a coalesce, rather than a rebuild. Given that Queues
> > are IOTs that you're not really supposed to know about, it's
> > not too surprising that Oracle has suggested a rebuild instead.
>
> --
> Regards,
>
> Greg Rahn
> http://structureddata.org
> --
> //www.freelists.org/webpage/oracle-l
>
>
>
--
Niall Litchfield
Oracle DBA
http://www.orawin.info
--
//www.freelists.org/webpage/oracle-l
Other related posts:
