#OOW2012: Assuring Infrastructure Performance Supporting Oracle Applications

For most enterprises, Oracle applications are mission-critical.  If you’re the DBA, you need assurance from the infrastructure team that your service level agreements will be met.  If you’ve decided to virtualize your Oracle applications, you want assurance that performance will be the same or better than when it was physically hosted.


It’s highly recommended that you have monitoring tools in place to troubleshoot problems if they occur, optimize resources to maximize performance, and prevent issues before they happen. 


One solution from NetApp is OnCommand Balance.  It monitors infrastructure performance (VMs, physical servers, and storage) and provides visibility and predictability into performance and capacity growth. 


OnCommand Balance and Oracle Applications

In an Oracle environment, it provides the added benefit of mapping Oracle database schemas so the DBA, virtualization/system administrator, and the storage administrator have a shared tool that provides insight into the logical connections in the environment.  Here’s an example of the topology mapping in Balance showing Oracle 11g schemas on a physical server environment.


Balance also enables the infrastructure team to answer the following key performance questions:


  • Do I have end to end visibility from the VM to the spindle?
  • If performance is slow, will I know where the problem is?  Database? VM? Host? Storage?
  • Can I predict problems before users notice?
  • If I decide to move Oracle to a private cloud, will I know how to assess performance and capacity requirements?
  • Will I know when I am going to run out of storage headroom/performance capacity?


How to Prevent a Storage Headroom Problem 

So, let’s say I’m the storage administrator.  Our virtualization admin requested storage for a new Oracle database deployed to handle ticketing operations.  We have NetApp storage and we recently upgraded to DATA ONTAP Cluster-mode, so I’m going to open Balance and navigate to our topology view and take a look to make sure the layout I set up was implemented correctly.  I logged in and navigated to the Cluster node 1’s Topology View.  Below we can see all of those Oracle volumes at the ready for the virtualization admin.


Wait, what’s this – Oh, I have an email from the Balance Predictor.  This is a really cool feature in OnCommand Balance.  Balance can monitor my cluster aggregates and proactively warn me when an aggregate is running out of headroom, or performance capacity.  This gives me the chance to correct developing issues before any users are impacted.  Let’s see…Uh oh, trouble.  It looks like one of my Security servers is going crazy.  While it normally does 32-76 IOPS, Balance just reported a surge to nearly 1200.  Let’s click ‘full analysis’ to take a closer look.


Here’s the utilization surge.  So, that security server is driving this disk group, agg4, beyond its performance capacity.



Let’s have a look at the Security Server Topology View.  That server uses cluster-mode Volume 6, which is sharing agg4 with these other servers.  I better move this volume now, before everyone suffers slow performance.  This can be accomplished with NetApp Workflow Automator.  With it I can dynamically move Vol6 off of this aggregate and onto an aggregate on my other cluster node.



See OnCommand Demonstrations at OOW

If you’ll be at Oracle Open World next week, please stop by NetApp’s booth #1301 for a demonstration of:


  • OnCommand Balance
  • OnCommand Workflow Automator for customized storage workflow automation including provisioning a new Oracle database from a gold image
  • OnCommand Insight, which offers storage performance, configuration, and capacity management 



Here’s a full lineup of all of the NetApp activities at the show:


For all of you that want to follow the show blogging and tweeting, you can use the hashtag #NetAppOOW.