Re: Long IO Latency
- From: Paul Drake <bdbafh@xxxxxxxxx>
- To: Brandon.Allen@xxxxxxxxxxx
- Date: Tue, 16 Aug 2005 13:04:28 -0400
On 8/16/05, Allen, Brandon <Brandon.Allen@xxxxxxxxxxx> wrote:
> According to Oracle, the average I/O time should be < 20ms for acceptable
> performance. I usually see about 5-10ms on my systems - some of which are
> RAID10, and some are RAID5 with lots o' cache and a high (.9+) r/w ratio.
>
> 50ms is ridiculous speed and 700ms is ludicrous (slow) speed (anyone remember
> Spaceballs?)
Brandon,
Imagine if a particular datafile is only opened once in a statspack
snapshot interval.
The overhead of opening the file would entirely occur during the
single IO as far as oracle is concerned - which may be on the order of
1000 milliseconds. This is where "averages" tends to hide what is
going on.
The OP should examine the individual wait events and determine if in
fact there is a problem.
I agree with the numbers that you are seeing ... but have grown used
to seeing under 2 ms averages when using an EMC Clarion unit (where
the cache hits are likely reducing the average).
Paul
>
>
>
> -----Original Message-----
> From: oracle-l-bounce@xxxxxxxxxxxxx
> [mailto:oracle-l-bounce@xxxxxxxxxxxxx]On Behalf Of Shivanischal A
>
>
> . . .
>
> Just one more question, while reading the I/O latency in the Statspack
> report, how slow is slow? I mean lets say the latency is 50ms, how do I
> determine that it is slow?
>
> Privileged/Confidential Information may be contained in this message or
> attachments hereto. Please advise immediately if you or your employer do not
> consent to Internet email for messages of this kind. Opinions, conclusions
> and other information in this message that do not relate to the official
> business of this company shall be understood as neither given nor endorsed by
> it.
>
> --
> //www.freelists.org/webpage/oracle-l
>
--
#/etc/init.d/init.cssd stop
# f=ma, divide by 1, convert to moles.
--
//www.freelists.org/webpage/oracle-l
Other related posts:
