Nutanix NX-3000 review: Virtualization cloud-style
Nutanix Virtual Computing Platform brings resilient, cloud-like server and storage infrastructure to traditional virtualization deployments
Nutanix defines a Storage Pool as a group of physical storage devices that may include PCIe SSD, SSD, and rotating disk. Naturally, a Storage Pool may span multiple nodes and will expand when the cluster scales out to include new nodes. A Container is defined as a group of VMs or files and is a logical subset of a Storage Pool. Each Container typically corresponds to a single data store in a VMware environment, for example.
The beauty of the Nutanix architecture is that day-to-day operation of the appliance requires little to no management intervention. Once the system has been configured, it should run without any operator input until you need to expand capacity. This typically happens when you add another appliance to the network and need to expand the number of nodes in the cluster. For this scenario, a menu option on the management home page labeled Expand Cluster will lead you through the process of bringing the new system online. Other operator actions might include managing available storage by creating containers and storage pools.
Primary management of a Nutanix appliance takes place from a Web browser. You can also use SSH to open a terminal session on any node and run scripts or manually start and stop services. Many of the settings that control how often different processes run or what triggers specific events reside in configuration parameters called GFlags, which you can set using the browser interface.
Nutanix provides advanced management and monitoring features in addition to the standard HTML5 management pages. These include direct insight into the individual functions, such as the Curator and Stargate. To access these pages, simply type in the URL of the Nutanix Controller VM and add the path to the specific service. Each of the primary functions mentioned earlier has its own page number or exclusive URL. Figure 2 below shows a typical dashboard screen with a default layout of informational widgets. This page is user configurable to include virtually any system level detail you could think of.
Figure 2: The Nutanix dashboard
You'll find plenty of detailed information about the inner workings of NDFS if you dig deep enough. NDFS makes heavy use of logging, and those logs provide insight into key performance parameters. If you don't see the information you're interested in on the dashboard, you only have to add a new chart from a long list of options.
Nutanix uses a RESTful API for its management interface along with plug-ins for VMware vCenter Server and other virtualization management tools. If you're really ambitious, you can write your own code using a language like Python. Using this approach takes about 10 lines of code to get a ton of information about currently running VMs. Inquiring minds can browse the REST API through the main portal page to manually explore the interface, although you probably wouldn't want to do that on a production machine.
All Nutanix products were designed from the ground up with performance and scale as the two driving principles. Data typically passes through the local OpLog with a copy sent over the network to another node in the cluster for redundancy. Sequential writes skip the OpLog and go directly to disk, and they may optionally skip the SSD tier entirely for specific use cases. This reduces the amount of storage needed on the SSD tier while taking advantage of the suitability for sequential writes of HDDs.