In my previous blog, I presented an overview of the new NetApp® OnCommand® Unified Manager 7.2, which simplifies day-to-day management of NetApp storage running on ONTAP® data management software. We made things simpler by reducing the need for two servers to one server. Some of you may be thinking, “There’s no such thing as a free lunch—how can one server do the work of two? Did you impact monitoring at scale by unifying the servers?” In today’s blog I’ll answer this question in depth.
In measuring scale with Unified Manager, our customers usually measure scale based on the number of clusters or nodes that a single server can manage. For the past few years, NetApp has been working to optimize Unified Manager (and Performance Manager) analytics, and in the 7.2 release we are pleased to announce that there is not a specific upper limit. That’s right, the number of clusters monitored by Unified Manger 7.2 is limited only by your server resources.
Most Unified Manager installations are on VMWare ESX, where we distribute a complete vApp. In the past, this required resource reservations (dedicated processor and memory allocations) for you to be able to meet the maximum limit guaranteed scale. This is no longer required due to several innovations.
- We have increased the efficiency of data collection and analytics.
- We have introduced a “scale monitor” feature that identifies when the resources provided to the Unified Manager server are insufficient.
- We no longer require VM resource reservations because Unified Manager tells you when the server is not getting sufficient resources.
This is especially good news for customers with labs or smaller environments, because the Unified Manager VM consumes only the minimal amount of resources required, which is lower than ever before.
With the Scale Monitor in place, Unified Manager monitors its own performance so that it can warn you when more memory or disk space is required to continue monitoring. These warnings typically show up after a new cluster is added or after a major configuration change to the environment; for example, when you configure a high number of quotas. (High is millions of quotas.)
Results Equals Real Savings
So what have we seen in terms of real use results? Most of the customers that I have met with are able to fold the performance management functionality into their Unified Manager server and decommission their old OPM servers, saving at least 50% in management server resources. Large enterprise customers have successfully used one server to manage more than 100 clusters when using a dedicated midrange server. Customers who are running labs and certification environments can now run them continuously, because the Unified Manager server consumes very little resources when the lab is idle.
Over the past few years, NetApp has made a consistent investment in optimization of the server code and database layout. This effort has brought the Unified Manager to a point where it can scale and meet the needs of most customers—small, midsize, and enterprise.
After identifying how many servers you need to manage your ONTAP based storage, the next logical step is to review your strategy for IT management tools and high availability. That’s the topic of my next blog, so join me again next week to explore storage management high availability.