SpeechMiner Disaster Recovery?

Architect & Flows
1

Hi all,

Has anybody deployed SpeechMiner+GIR for a Disaster Recovery scenario? Any thoughts and considerations that should be accounted for?

Genesys doesn’t officially support this deployment, and we are trying to work around this gap. So far, no success.

Here’s what we have done:

  • Shared storage between both sites on WebDav
  • Shared storega for the keystore for RCS
  • Shared database for SpeechMiner UPlatform
  • Two redundant pairs of RCS and RPS
  • One IVR Profile for each DataCenter (Primary / DR) pointing to corresponding RPS for call metadata
  • Redundant ICONs
  • One cluster of Cassandra and ElasticSearch covering both DataCenters

On Disaster Recovery scenario, we attempt to change the RCS mapping on SpeechMiner so it points to either the primary datacenter or the DR datacenter.

Any suggestions or is it really not feasible at all?

2

What do you want to exactly achieve - High-availability mode or business continuity mode?

3

A “Disaster Recovery” mode - does not need to be fully automated (some manual actions may be performed), but needs to be able to access, playback and download the recordings made by the Primary site, and be able to record interactions that would be accessible by the Primary site once the primary site is recovered.

4

We actually found out what was wront… There was incorrect configuration for towards IRWS url in the RPS configuration file. After changing that and some manual configuration on the SMConfig interface for the recording RCS path, the Disaster Recovery mode was successfully achieved.