Node

• A physical server contained in a Nutanix block
• Model Comparison:
  • 2 Node (6000/7000)
  • 4 Node (1000/2000/3000/3050)
• Each node runs a hypervisor (ESXi/HV/KVM) + Nutanix CVM (Controller VM)

Block

• A set of Nutanix nodes contained in a single enclosure
• Uniform populated blocks recommended
  • Prevents storage skew
  • Ensures if block fails/maintenance needed, system can run without interruption

Cluster

A group of nodes contained in one or more Nutanix blocks

Node/Block/Rack Awareness

• Determines component/data placement
• DSF is node/block aware
• Node awareness is default
  • Replicas are replicated to another node to protect against node failure
• Min. 3 blocks needed for block awareness
  • Replicas are written to other blocks in cluster
  • Provides data availability in case of block outage
  • As of AOS 4.5, block awareness is best effort
• Increases to rack aware as cluster scales

• Within block, PSU + fans are the only shared components

• In event of block failure, block awareness maintained and re-replicated blocks are replicated to other blocks in cluster: