Not possible to change Failover configuration after migration to BrokerGateway

Affected versions

DataMiner Systems using BrokerGateway-managed NATS from DataMiner 10.5.0 [CU4]/10.5.7 up to 10.6.0 [CU4]/10.6.7.

Cause

Starting from DataMiner 10.5.0 [CU4]/10.5.7, in systems that have been migrated to BrokerGateway, a change in NATS behavior can cause inconsistent cluster routing and authority resolution between Failover Agents. As a result, redirects can loop, and no clear authoritative Agent is selected, so Failover configuration updates can be intercepted or aborted.

Fix

Install DataMiner 10.6.0 [CU5]/10.6.8.

Note

To upgrade an automatic Failover setup that is affected by this issue, in the upgrade options, make sure the Failover policy is set to Simultaneously or Upgrade main and backup Agent simultaneously.

Description

In DataMiner Systems that have been migrated to BrokerGateway, switching from a manual to an automatic Failover configuration or vice versa can fail.