HA Config for Hub to Hub Failover including probe and queues operation

Products

• CA Unified Infrastructure Management On-Premise (Nimsoft / UIM)
• CA Unified Infrastructure Management SaaS (Nimsoft / UIM)
• DX Unified Infrastructure Management (Nimsoft / UIM)

Issue/Introduction

Customers often ask for guidance on how to configure hubs for failover so remote/secondary hubs can act as failover hubs where remote (agentless) probes have been deployed. This currently only applies to an environment that doesn't use/rely upon MCS for probe/profile deployment.

The HA probe can be used for infrastructure components' high availability, but it does not support monitoring probes with MCS-based configurations to failover.

This feature/functionality may be considered in a future release.

Environment

• DX UIM 20.4 or higher

Cause

• Guidance

Resolution

Sample HA Config for Hub<->Hub Failover including probe and queues operation

Probes installed and running on Secondary in normal mode of operation.

HA Probe installed and configured on <ABC>_Secondary watching <ABC>_Primary.

Queues and probes listed below from the HA.cfg on a Secondary hub:

<queue_up>

   queue_0 = action_manager

   queue_1 = alarm_manager

   queue_2 = data_engine

   queue_3 = ems

   queue_4 = event_manager

   queue_5 = HAProxy01_AlarmDiscovery

   queue_6 = HAProxy01_QOS

   queue_7 = HAProxy02_AlarmDiscovery

   queue_8 = HAProxy02_QOS

   queue_9 = legacy_alarm_manager

   queue_10 = probeDiscovery

   queue_11 = udm_inventory

</queue_up>

<queue_down>

   queue_0 = Secondary_QOSDiscovery

</queue_down>

<probes_up>

   probe_0 = data_engine

   probe_1 = ems

   probe_2 = ace

   probe_3 = wasp

   probe_4 = qos_processor

   probe_5 = discovery_server

   probe_6 = mpse

   probe_7 = maintenance_mode

   probe_8 = udm_manager

   probe_9 = sla_engine

   probe_10 = trellis

   probe_11 = service_host

</probes_up

 

• Configure data_engine with the same credentials as the Primary data_engine and verify it can login to the database. It is ok to start the data_engine and have them both running for a limited amount of time but make sure that the distsrv is deployed and kept up and running otherwise the data_engine will not start.

• Deploy and configure the AdminConsole on <ABC>_Secondary Hub:
  How to install a second instance of the web-based Admin Console application on a secondary hub

• Secondary hub is configured for adminconsole on port 80
  http://<hostname or IP_ADDRESS>

 

HA Queues need to be configured, but disabled on the Secondary. They will be enabled by the HA probe on failover, with some connecting to other hubs/Proxy Hubs.

 

• The Secondary_QOSDiscovery Queue connects to the Primary Hub during normal operation and is disabled on failover.

• Nas replication is running between Primary and Secondary during normal operation so no Alarm Queue is needed.

 

Proxy Fail-Over

The Proxy hub fail-over is straightforward forward with the 2 <City/Location> Proxy hub servers watching each other, ready to pick up the tertiary Hub Queues if one should fail.

 

<ABC>_Proxy01 ç=è <ABC>_Proxy02

 

<ABC>_Proxy01 watches <ABC>_Proxy02 which manages Queues for <ABC>_SiteHub03 and 04.

Example in HA.cfg on <ABC>_Proxy01 which will have Queues enabled if <ABC>_Proxy02 stops.

<queue_up>

   queue_0 = HASiteHub03_AlarmDiscovery

   queue_1 = HASiteHub03_QOS

   queue_2 = HASiteHub04_Alarmdiscovery

   queue_3 = HASiteHub04_QOS

</queue_up>

 

<ABC>_Proxy02 in turn watches <ABC>_Proxy01 and will enable its Queues to SiteHub01 and 02 if needed.

Using the Infrastructure Manager (IM), under hub->Advanced Settings->General, make sure the ‘paired’ hubs are configured with the same origin. This is important for all high-availability scenarios.