Comments on: Audit Ouch!

By: Correlation oddity « Oracle Scratchpad

Correlation oddity « Oracle Scratchpad — Mon, 19 Dec 2011 18:26:17 +0000

[...] anomaly may look familiar – but this time it’s nothing to do with the audit_trail, it just seems to be the way [...]

By: Audit Excess « Oracle Scratchpad

Audit Excess « Oracle Scratchpad — Tue, 07 Jun 2011 18:20:09 +0000

[...] by session was a light-weight operation, think again. Interestingly, the auditing anomaly that I reported recently re-appears in the “audit by session” option. You will recall that when auditing is [...]

By: Jonathan Lewis

Jonathan Lewis — Sat, 04 Jun 2011 10:29:37 +0000

Thanks for the information. It’s nice to know that it’s so localised.

I noticed the document had a summary of the redo OPs – do you know if there’s a quick way to generate this summary. At present I just dump the log file and then grep my way through it – it would be convenient if there were a dump option that dumped just the one necessary line. (I really don’t want to have to write a C program to unravel the file, given the infrequency of the occasions when I need to do that type of analysis).

By: Anonymous

Anonymous — Fri, 03 Jun 2011 11:12:52 +0000

Bug 5166745 Excessive redo generation with auditing enabled
Versions confirmed as being affected
* 10.2.0.1
* 10.2.0.2
* 10.2.0.3
This issue is fixed in
* 10.2.0.4 (Server Patch Set)
* 11.1.0.6 (Base Release)
This problem is introduced in 10g R2 (10.2.0.1).
Excessive redo generation occurs in 10.2 with standard auditing enabled
(around 15%-20% increase compared to 10.1/9.2)

By: Jonathan Lewis

Jonathan Lewis — Thu, 02 Jun 2011 20:41:08 +0000

Franck,

Thanks for the comment.
I hope the rest of the audience viewed the presentation the same way. I have to say, when things “go wrong” I prefer it when I’ve planned for them to go wrong so that I don’t have to backtrack and contradict myself. Still, with hindsight I can say the audience did see the most important bit of the “science of Oracle”, the bit where someone says: “wait a minute – that can’t be right.”

By: Heitor Kirsten

Heitor Kirsten — Thu, 02 Jun 2011 20:40:04 +0000

I didn’t pay attention when you wrote “appeared only on 10.2.0.3″, sorry.

By: Jonathan Lewis

Jonathan Lewis — Thu, 02 Jun 2011 20:26:50 +0000

Heitor,

Given that it doesn’t happen in the previous version or the newer versions, it’s reasonable to assume (at least for the moment) that it’s not intended behaviour (i.e. it’s a bug).

By: Franck Pachot

Franck Pachot — Sun, 29 May 2011 09:08:46 +0000

Hi Jonathan,
>> Footnote [...] I gave what must have been my worst ever presentation
No, don’t worry, it was a good presentation.
You proved that it is not so easy to match expected results with experimental observations, because there are so many parameters that can change the behaviour of such a complex system as Oracle is. But you have shown as well that an expert will not give up and look at the system as a black box. That he still can try to understand what happens, even if it takes a few hours…
Regards,
Franck.

By: Heitor Kirsten

Heitor Kirsten — Fri, 27 May 2011 20:47:13 +0000

Any idea why Oracle is doing such thing?

Redo size higher by 50% when audit is turned on is something to think about…

Heitor