See also "Improve Mobility scalability with load-balancing and/or clustering" - https://ideas.microfocus.com/MFI/mf-gw/Idea/Detail/15526

In the event of a total loss of a GMS server or a site containing a GMS server there is no automatic failover to a second server or site. A cold standby (ready-built GMS server) can be brought up and users provisioned but this necessarily takes time and the process incurs an interruption to service which could be significant when there are large numbers of users and/or devices.

As an example, a model such as that used by ZENworks, with multiple primary servers, would be preferable although the database represents a single point of failure.


MRFE / MG / 101289577275

Comments

  • This problem can probably be addressed running GMS database in cluster mode, replicating attachments and other needed files with drbd and running a load balancer like nginx in front of GMS's, but I'm afraid this is not a supported deployment. Anyway, nowadays GMS is a critical part of mail infrastructure to keep it so unprotected