Re: Database Performance Problem between 3:00PM and 4:00PM

From: Robert Klemme <shortcutter_at_googlemail.com>
Date: Wed, 12 Dec 2007 09:00:40 +0100
Message-ID: <5s9ip8F1869hdU3@mid.individual.net>

On 12.12.2007 04:18, zigzagdna_at_yahoo.com wrote:

> On Nov 7, 1:50 am, Jan Krueger <j..._at_stud.uni-hannover.de> wrote:
>> NetComrade wrote:

>>> On Sun, 04 Nov 2007 11:38:47 -0800, DA Morgan <damor..._at_psoug.org>
>>> wrote:

>>>> Jan Krueger wrote:
>>>>>> I talked to my SAN team and they do not see anything wrong with SAN.
>>>>>> It definteley looks like some scheduled task but no one knows what it
>>>>>> is?
>>>>> I have never heard SAN people saying something else. Same is true for
>>>>> network people.
>>>>> Jan
>>>> I agree. SOP is to point the finger elsewhere. Ask them what metrics
>>>> they gathered that support their conclusion.

>>> Ask them to enable SNMP on the array and how IO patterns for the
>>> night. Find out what else is on the SAN.. if they insist there are no
>>> problems, tell them you want a dedicated SAN :)
>>> This is likely some backup job.
>>> -a
>>> .......
>>> We run Oracle 9iR2,10gR1/2 on RH4/RH3 and Solaris 10 (Sparc)
>>> remove NSPAM to email

>> No no, just get two SQL trace files. Don't look somewhere else than in
>> the db. Once you got the information from the trace that storage access
>> is causing the difference, bother the other department.
>>
>> Jan- Hide quoted text -
>>
>> - Show quoted text -
> We have finally nailed down this problem. It is something in SAN. When
> we moved database from SAN to local disk on Windows server, problem
> went away. I had never heard eralier that problem in SAN can happen
> certian time of day, really strange.

Thanks for the update! Can you share more details (SAN vendor, what exactly went on in there)? I am suspecting some kind of SAN internal snapshotting or backup

Kind regards

        robert Received on Wed Dec 12 2007 - 02:00:40 CST