Current Position:Home > Looking to see if this NFS issue is universal with everyone else

Looking to see if this NFS issue is universal with everyone else

Update:10-11Source: network consolidation
Advertisement
First I followed these directions exactly (and our filer is a Netapp):
http://www.oracle-base.com/articles/11g/OracleDB11gR1RACInstallationOnRHEL5UsingVMwareESXAndNFS.php
So the database has been created as per the end of these instructions. Nothing customized. No data populated or anything of any substance has been done yet.
But if I now follow the instructions in this link (you do need a Netapp NOW login, but basically on the filer you do a nfstat -z to clear the stats, then nfsstat -l to see the top talkers, and then nfsstat -h <IP> to see the kind of traffic per host):
https://now.netapp.com/Knowledgebase/solutionarea.asp?id=kb36582
My top two servers by a long shot are the brand new RAC Nodes. Even more than the VMware NFS IP's used to carry all the guest vm traffic. After 10 minutes the top host had 450,000 NFSOPS. The third highest was less than 50,000.
The bottom of that netapp link states "If it looks like it is the culprit that is generating too many getattr, access, or other requests at a rate of 50% or higher at the expense of reads and writes, you should work with your UNIX vendor or support team to analyze that data and determine what processes on the host are running against the filer to cause so many unoptimized requests."
My values for "getattr" and "access" are 62% and 35% of the total traffic. My reads and writes are 2% and 1% respectively. According to Netapp this is a problem and hugely unoptimized. What the heck is Oracle or CRS doing? Anybody with a similar setup see the same thing? Is this a cause for concern? Should we bring it up with Technical Support or is this expected behavior for Oracle/RAC? Let me know what you all are seeing.
Thanks!

The Best Answer

Advertisement
Hi,
I have same issue on one of my two RAC nodes.
We have NetApp 3020c and using a physical Dell PE 1435 server with RHEL 5.3 and Oracle 10.2.0.3 RAC.
Did you found a solution for problem?
Best Regards
  • Looking to see if this NFS issue is universal with everyone else Update:10-11

    First I followed these directions exactly (and our filer is a Netapp): http://www.oracle-base.com/articles/11g/OracleDB11gR1RACInstallationOnRHEL5UsingVMwareESXAndNFS.php So the database has been created as per the end of these instructions. Nothing

  • Issue In Universe Creation Update:10-11

    Hi, I have One query in BI which contains two structures one at row level and one in column. column stuture has various selections which are defined using characteristics whereas raw sturcture level selections are dfined using characteristics as well

  • Parsing Issue at Universe Level Update:10-11

    Dear All, I am a BO Consultant, for the past few months we (our report developing team) were trying to create new objects at the Universe level (based on requirement) so that we can use the same in the report level, but unfortunately we failed to do

  • DateTime issue at Universe Level Update:10-11

    Hello Experts, We have stored date in database as datetime format. In Universe, we have created object on this date. But we have to keep type of this objet as u2018Dateu2019 as there is no datetime  object type available  in universe. When we dragged

  • Issues with universes for AS400 DB2 databases Update:10-11

    I create Business Objects universes and Webi reports accessing data from AS400 DB2 databases (via ODBC). I have a couple of queries that I have been having trouble resolving. I hope someone in the forum is working on a similar set-up. My queries are

  • IChat AV issues with university Update:11-30

    I'm able to use iChat to chat but not for video and i know that one can access skype on windows through my university proxies, but I cant seem to use in through the proxies on my mac (os X 10.3) does it use the network location settings?your message