RE: Linux buffer cache monitoring

From: Henry Poras <>
Date: Fri, 25 Jul 2008 16:00:52 -0400
Message-ID: <000001c8ee91$24a59880$>


It's been a while since I have used Orion. I forget if it can be used with Direct I/O. If so, that should bypass the OS cache.

With an EMC backend I was able to do what you are looking for by using navcli to measure the physical storage I/O. It was a good match with the Orion numbers.


> -----Original Message-----
> From:
> [] On Behalf Of Allen, Brandon
> Sent: Friday, July 25, 2008 2:46 PM
> To: Hemant K Chitale; oracle-l
> Subject: RE: Linux buffer cache monitoring
> Thanks Hemant, but that's not exactly what I'm trying to monitor. I'm
> not trying to see how much of the RAM is used/free - rather, I'm
> basically trying to see the Linux buffer cache hit ratio, and the
> components that make up that ratio (logical/buffered reads & physical
> reads).
> I'm running Orion (Oracle's I/O capacity testing tool) and it's
> reporting about 5000 IOPS, but that is too good to be true considering
> I'm only striped across 5 disks (in RAID5), I should only be able to
> sustain about 625 IOPS (125 x 5, as long as they're 100%
> reads, which is
> what I'm testing with Orion). According to sar and iostat, I'm only
> doing about 250 IOPS and my understanding is that those are physical
> reads from the OS perspective (I know they might still be
> cached by the
> disk, controller, array cache, etc. in the backend). So, my guess is
> that the other 4500 IOPS are being done in the Linux buffer cache, but
> I'd like to see something to confirm that. I've got 8GB of
> RAM on this
> box and I'm testing Orion with a 64GB file, so I don't see
> how it could
> be satisfying so much of its I/O from cache.
> Regards,
> Brandon
> -----Original Message-----
> From: Hemant K Chitale []
> I gave up trying to monitor this !
> See

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.


Received on Fri Jul 25 2008 - 15:00:52 CDT

Original text of this message