Academy/VCAP — VCF Networking (3V0-25.25)/Lab NN1: BGP Convergence & Graceful Restart on NSX Edges
This lab targets VCF 9.0

Lab NN1: BGP Convergence & Graceful Restart on NSX Edges

VCF 9.0Intermediatevcap-advanced⏱ 75 min

Objectives

  • Design and validate fast, BFD-based BGP convergence between NSX Edges and ToR switches.

Prerequisites

VCF lab environment deployed and operational

Lab Environment

Standard VCF lab environment for Advanced VCF 9.0 Networking (NSX & Advanced Routing)

Tasks

Task 1 Lab NN1: BGP Convergence & Graceful Restart on NSX Edges

Design and validate fast, BFD-based BGP convergence between NSX Edges and ToR switches.

Step 1

Enable BFD on the Tier-0 BGP neighbors towards the ToR with a 500 ms interval and multiplier 3 (Timers & Password on the neighbor; 500 ms is the minimum and the default for an Edge node running as a VM — only bare-metal Edges go lower). Record the resulting detection time (~1.5 s).

Step 2

Leave the Tier-0 Graceful Restart mode at its default, Helper Only (the VCF design recommendation), and do not select 'Graceful Restart and Helper' while BFD is in use: NSX documents that Graceful Restart on a tier-0 increases the time a remote neighbor takes to select an alternate tier-0 and delays BFD-based convergence (it is also unsupported with a single BGP peer). Record the GR mode configured on the ToR as well.

Step 3

Kill one Edge via forced reboot and measure steady-state traffic drop.

Step 4

Comparison run: measure convergence with Helper Only (BFD detects the dead peer and traffic moves to the surviving Edge in about 1.5 s), then temporarily set 'Graceful Restart and Helper' and repeat — the stale-route behaviour keeps the neighbor forwarding to the failed Edge longer. Revert to Helper Only.

Step 5

Document tuning knobs with their trade-offs (CPU vs convergence speed).

Validation Gate

Check: Verify lab completion

Expected: Lab exercise completed successfully

Common Errors

BGP graceful restart timer mismatch between NSX Edge and physical router
Fix: Graceful restart requires matching restart and stale-route timers on both peers. If NSX Edge restarts and the physical router's stale timer is shorter than the Edge's restart time, routes are withdrawn during Edge reboot. Set the stale-route timer on the physical router longer than the Edge restart time — and note that with BFD the recommended tier-0 mode is Helper Only, so this timer matters only when the Edge acts as the restarting router.
BFD false positives during Edge VM vMotion
Fix: BFD with aggressive timers (e.g., 500 ms x 3, the Edge VM minimum) can detect Edge VM vMotion latency as a failure, causing route withdrawal. Either: disable BFD during planned Edge maintenance, or tune BFD timers to tolerate vMotion switchover latency (set detect-multiplier × interval > vMotion switchover time).
Not verifying BGP convergence time meets SLA requirements
Fix: BGP convergence after Edge failover includes: BFD detection (ms) + BGP session timeout (if no BFD) + route propagation (ms-seconds) + FIB update (ms). Measure actual convergence time with traffic during planned Edge failover test. Compare against SLA requirement.

Final Validation

Lab completed successfully

✓ All steps completed → No errors observed

Cleanup / Restore

• Revert to snapshot if needed

Design Reflection (VCDX)

BGP convergence tuning shows deep networking expertise. VCDX panelists test whether you measured actual convergence times and tuned timers for your SLA.

⚠ Known Pitfalls

Using aggressive BFD timers without accounting for vMotion latency — causes false failovers.
Assuming default BGP timers meet SLA — always measure actual convergence during Edge failover test.
Was this page useful?
Type to search. ↑ ↓ to move, Enter to open, Esc to close.