Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Mailing Lists -> Oracle-L -> FW: EMC monitoring - slightly OT

FW: EMC monitoring - slightly OT

From: Bobak, Mark <Mark.Bobak_at_il.proquest.com>
Date: Wed, 15 Mar 2006 10:45:12 -0500
Message-ID: <AA29A27627F842409E1D18FB19CDCF27072A937D@AABO-EXCHANGE02.bos.il.pqe>


Henry,  

I forwarded your question to one of our Sys Admins. See below for what he had to say.  

-Mark
 

--

Mark J. Bobak
Senior Oracle Architect
ProQuest Information & Learning

"Exception: Some dividends may be reported as qualified dividends but are not qualified dividends. These include:

  --IRS, Form 1040-A Instruction Booklet, Line 9b: Qualified Dividends  


From: Suiter, Thomas
Sent: Wednesday, March 15, 2006 10:38 AM To: Bobak, Mark
Subject: RE: EMC monitoring - slightly OT

As to more "nice" monitoring tools it all depends upon what you want to pay (or not):  

If you are into the whole integrated world view with larger $$, you can use ECC with PerformanceManager, that will tie in end to end: Database tables (which are hot, I/O's to each, etc), host information (cpu, memory, etc), san switch (ports running at max throughput or counting errors degrading perf), than into the storage  

Many people use navisphere analyzer which is a fairly inexpensive add on to navisphere that you can monitor your clariion, this will give storage specific information: physical disk getting hot, cache information, connectivity port information, etc (highly recommend this, gives nice trend information using graphs)  

Lastly you can turn on a snmp agent for the clariion and allow read requests, we have the above two so I don't know what information you would be able retrieve from this; I assume some performance information should be available    

For his specific problem he should also be able to continue to use the navicli command with the "getcache" attribute to retrieve information about each individual SP (after you have enable statistics logging inside each sp, else you won't get meaningful information): meaningful switches would be:  

-pdp - percent of dirty pages in cache (pages modified in cache but not
written to disk, high number means most write changes are occurring in cache rather than in physical disk)
-high - write cache high watermark (when percent dirty pages reaches
this force flushing of cache)
-low - write cache low watermark (force flush cache down to this
watermark than stop flushing)  

These two commands should retrieve what I believe he's interested in navicli -h <spA ip address> getcache -pdp -high -low navicli -h <spB ip address> getcache -pdp -high -low  

here is the output from one of our clariions: navicli -h 192.168.20.169 getcache -pdp -high -low

Prct Dirty Cache Pages =            49
High Watermark:                     80
Low Watermark:                      60

On the SP at 192.168.20.169 if 80% of cache is dirty it will flush cache down to 60%, it currently is at 49%


From: Bobak, Mark
Sent: Wednesday, March 15, 2006 9:53 AM
To: Suiter, Thomas
Subject: FW: EMC monitoring - slightly OT

Thomas,  

This question comes from an Oracle mailing list I'm on. What tools do you use to monitor our EMC boxes?  

-Mark
 

--

Mark J. Bobak
Senior Oracle Architect
ProQuest Information & Learning

"Exception: Some dividends may be reported as qualified dividends but are not qualified dividends. These include:

  --IRS, Form 1040-A Instruction Booklet, Line 9b: Qualified Dividends  


From: oracle-l-bounce_at_freelists.org
[mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Henry Poras Sent: Tuesday, March 14, 2006 12:47 PM
To: oracle-l_at_freelists.org
Subject: EMC monitoring - slightly OT

I'm trying to find some simple way of monitoring our EMC SAN (Clariion). I assume emc provides something, but I haven't turned it up as of yet. All I really need is a way to check out our cache (is it saturated yet?) and disk i/o utilization. I have found navicli which seems to be good for providing static information (how things are configured), but not so good for dynamic. I've been looking at the EMC web site and will head back there, but they don't seem to be good at posting docs online.

Thanks.

Henry

--

http://www.freelists.org/webpage/oracle-l Received on Wed Mar 15 2006 - 09:45:12 CST

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US