Data Center

• All Analytic Nodes for a single vRealize Operations cluster must be in the same Data Center.

Cluster Management

• If HA or CO-Stop issues, all nodes should be on separate hosts.
• Maintain proper Log Levels.
• Don’t use Admin accounts for local integrations such as vRealize Operations for Horizon, vRealize Log Insight, or Endpoint Operations.

Hardware Resources

• Same amount of CPU, Memory, and Disk on Analytics Nodes.
• VMware Tools: Updated through Upgrade process. 
• Validate Sizing Guidelines.

Storage

• Add VMDK Instead of Extending.
• Consistent 10ms with spikes to 15ms.
• Storage Tiers:  Same Type.
• DRS Anti-Affinity Rules:  Keep nodes on separate data stores, and hosts.
• Storage DRS: SDRS to Manual (unintentional migrations), vMotions are unadvised.

Networking

• Static IP: Master / Replica must use Static tied to FQDN.
• DHCP: Data and RC can use DHCP (DHCP client syncing).
• Latency:
  • Analytics to Analytics:  Sub 5ms
  • RC to Analytics:  Sub 200ms
  • < 5% Packet Loss
• Forward and Reverse DNS lookup all nodes.
• Same Subnet, Stretched Layer 2 and Layer 3 network not supported.
• Network Bandwidth >= 1gbps.
• WAN: Do not distribute analytics nodes across.
• Analytics Nodes Must Not be Separated by:
  • NAT
  • Load Balancers (In front is fine, but not between)
  • Firewalls
  • Proxy (That inhibits bidirectional communication by IP or FQDN)
• Scanning tools can DOS