An application had an outage during an Avi upgrade. The Service Engine Group assigned to the Virtual Service is using the default configuration. What…

An outage during a rolling upgrade in an N+M Service Engine (SE) Group occurs if there is no available SE to migrate the Virtual Service (VS) to when the active SE is taken down. The default HA mode is N+M, but if the group is configured with M=0 (zero buffer SEs), there is no redundancy. When the single active SE hosting the VS is rebooted for the upgrade, the VS has nowhere to fail over, causing an outage. While the default value for M is 1, a "default configuration" could imply a group with only one provisioned SE, making M effectively 0.

A. Active/Standby HA (if an option): This configuration is designed for failover, which, while causing a brief blip, is not typically considered a full outage and is not the default SE Group HA mode.

C. N+M HA with M>0 (if an option): This is the default and is designed to prevent outages during upgrades by providing a buffer SE for the VS to migrate to.

D. Disruptive Upgrade (if an option): This would cause an outage, but the default upgrade method for an SE Group is "Rolling Upgrade," which is non-disruptive if HA is correctly configured.

1. VMware NSX Advanced Load Balancer Documentation, "HA Modes for Service Engine Group," Section: Elastic HA N+M Mode. This section explains that M represents the number of buffer SEs for failover. An M value of 0 provides no buffer.

2. VMware NSX Advanced Load Balancer Documentation, "Upgrade Service Engines in an SE Group," Section: Rolling Upgrade for an SE Group. This document states, "Virtual services are migrated away from an SE before it is upgraded." This migration requires a destination SE.