RE: index maintenance tracing and debugging event

From: Mark W. Farnham <mwf_at_rsiz.com>
Date: Thu, 20 Jun 2013 10:33:38 -0400
Message-ID: <053201ce6dc3$28508920$78f19b60$_at_rsiz.com>



Good one!

By the way, I personally think that anyone who does a direct load into a table concurrently subject to an online index rebuild has worse problems than an Oracle "bug."

Perhaps the "bug" is that online index rebuild does not make a direct load attempt wait.

mwf

-----Original Message-----
From: oracle-l-bounce_at_freelists.org [mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Tanel Poder
Sent: Wednesday, June 19, 2013 5:41 PM
To: Josh Collier
Cc: oracle-l_at_freelists.org
Subject: Re: index maintenance tracing and debugging event

One way would be to dump redo :)
There's also a following event, which traces delayed index maintenance operations (so, possibly not every single index maintenance op, but only the bulk-delayed ones):

12837, 00000, "Delayed index maintenance debugging event"

// *Document: NO
// *Cause:
// *Action: Control dumping of debugging information for dealyed maintenanc

It's mentioned here:

*Bug 11795952 - Direct load insert may result in ORA-1 errors during concurrent online index rebuild operation [ID 11795952.8] *

*"If event 12837 is set to level 31, the we see that there are multiple rows being inserted into the index bufffer (kxib) for the same index."*

I've never used it myself, so better ask from support before running this in production ...

--
*Tanel Poder*
Enkitec (The Exadata Experts)
Training <http://blog.tanelpoder.com/seminar/> |
Troubleshooting<http://blog.tanelpoder.com/>
 | Exadata<http://www.amazon.com/Expert-Oracle-Exadata-Apress/dp/1430233923>
 | Voicee App <http://voic.ee/>



On Wed, Jun 19, 2013 at 9:04 PM, Josh Collier
<Josh.Collier_at_banfield.net>wrote:


> Hi,
> I am looking for an event to set that will debug/trace oracle index
> maintenance operations. Such as the type of maintenance that occurs
> automatically when you work on a table with global indexes. I am hoping
> there is an event to set that will give me more information than what I
get
> from 10046,
>
> Thanks!
>
> Josh C.
>
> --
> http://www.freelists.org/webpage/oracle-l
> > > -- http://www.freelists.org/webpage/oracle-l -- http://www.freelists.org/webpage/oracle-l
Received on Thu Jun 20 2013 - 16:33:38 CEST

Original text of this message