Re: Possible Oracle Bug, Oracle Claims No

Sybrand Bakker wrote:
> On Tue, 11 Mar 2003 09:52:35 -0800, Chuck <ccarson_at_echeeba.com> wrote:
>
>

>>So, we have absolutely no logs at the OS, Veritas, or Raid conrtoller 
>>level, we have an error detected in an oracle background process, only 
>>one datafile was affected even though there are many on that disk group, 
>>AND we just happen to be running a heavy job on the datafile at the time 
>>of the error. Oracle claims there was no oracle problem and that there 
>>was a hardware I/O problem, like a bad block or something. In my 10+ 
>>years of experience I have never seen a raid group produce an I/O due to 
>>a bad disk block AND not log anything what-so-ever. Solaris is usually 
>>very very good at logging file system inconsistencies.
>>
>>Just wanted to get feedback from other impartial DBA's.

Incorrect. Solaris always has and will probably always will log I/O errors, and is very, very effective in detecting them. (ever since the early days of SunOS)

> You would need to run the Sun explorer utility on the server to
> exclude you didn't encounter any errors.

Sun explorer would be _compeltely_ useless in this instance.

> Also I'm not sure why you are running Veritas Database Edition (which
> is playing trics to both the O/S *and* Oracle) and aren't running
> quick i/o. AFAIK the very *purpose* of this software product is to be
> capable to use quick i/o.

Veritas is a journaled file system, unlike ufs. Which is the sole point of using it. Quick I/O only supposedly gives you performance similiar to raw filesystems but more and more people are claiming they fail to see any improvements. Also, when you have a very fast SAN, as we do, you do not see the added benefit, so you are merely adding an unnecessary level of complexity. If you claim otherwise, I would love to see some bonafide performance metrics, because I have been in very large enterprise environments with hard empiral evidence supporting otherwise.

> That all said I have the distinct feeling you started finger pointing
> too early, and I agree with Oracle's analysis.

> Solaris, when not being set up properly, can be a can of worms.

Hmm. never in my 10 years experience with Sun would I refer to it as a can of worms. Any capable admin should have no problems installing, configuring, tuning, and maintaining a Solaris box. I usually delegate these tasks to my more junior admins and interns.

> My department is performing remote site administration, and if I would
> have to state which Unix and which vendor is providing the most
> trouble, it is Solaris and Sun support. They are *very* quick to
> withdraw support when you didn't install *ALL* patches.

I have asked many an expert, including a software engineer formerly at oracle for 8 years, and given the evidence stated before, it is clearly pointing to an oracle error, whether it is a bonafide bug or not I cannot say.

I can always tell when someone is a true SA or a DBA wanna-be SA.

-CC

-----------== Posted via Newsfeed.Com - Uncensored Usenet News ==----------

   http://www.newsfeed.com The #1 Newsgroup Service in the World! -----= Over 100,000 Newsgroups - Unlimited Fast Downloads - 19 Servers =----- Received on Tue Mar 11 2003 - 17:20:17 CST