For SAM with Notification Cache Publishing Enabled which interacts with VMware M&R UI, several failover scenarios exist and not all of them require hook scripts for VMware M&R collectors. In the VMware M&R platform, the smarts-collector typically resides on a separate host and performs event and performance collecting from one or more Domain Managers.

The Failover Manager monitors the Domain Managers and does not monitor the collectors. You can configure the failover_post-AMPM-sample.pl hook script so that the script can stop the smarts-collector on a separate host and start another one near the newly-promoted Domain Manager.

Table 1. Scenarios for hook scripts for collectors

Scenario or situation

Is configuring a hook script needed?

If SAM with Notification Cache Publishing Enabled or any one of its services (Tomcat, Rabbit MQ, and ElasticSearch) fail, the Failover Manager performs the actions described in “Active SAM failure scenario” on page 27

Not applicable. Configuring a hook script to start VMware M&R collectors is not applicable, because the collectors gather data from Domain Managers and not from Global Managers like Service Assurance Managers or Adapter Platform.

For a Site failover which includes the failure of VMware M&R UI

VMware M&R failover is not covered in this guide and it is a custom solution.

For a Site failover where VMware M&R UI is still running, the Failover Manager performs the actions described in Chapter 12, “Site Failover,” for the failed VMware Smart Assurance components.

It depends on your VMware Smart Assurance deployment:

  • A hook script is not needed if the VMware M&R collector is located near the Domain Manager at both Active (Site A) and Standby (Site B) locations in the same geographical area. In this case, delays due to high latency are not a concern and the same collector can get information from Domain Managers regardless of the site from which they are running.

  • A hook script is needed if the Domain Managers at both Active (Site A) and Standby (Site B) locations are geographically dispersed and the VMwareM&R collector is not located near the Domain Managers. Delays due to high latency are a concern when an Active Domain Manager associated with a collector fails. Also, for latency reasons, you should consider installing a second collector near the Standby location (Site B). Typically, you need at least one collector per site.

“Example of smarts-collector messages that indicate high latency” on page 69 provides additional information about high latency.