+

Search Tips   |   Advanced Search

Disaster Recovery Site

IBM MobileFirst Platform Foundation supports the creation of a separate disaster recovery site that becomes operational if the original site goes down.

A disaster recovery site is a second, physically separate IT center on which a copy of the IT systems exists, and springs into operation if the original site is down. IBM MobileFirst Platform Foundation has such a site for some of its customers.

Within the site, MPF provides redundancy at every level: compensating load balancers, multiple IBM MobileFirst Platform Foundation servers that scale linearly, and database redundancy through Oracle RAC. Some customers prefer to provide another level of redundancy using a disaster recovery site.

The key administrative factors for such a site are:

Architecture

The architecture of the backup site is a copy of the original site. Special care must be taken to:

  • Provide access to all corporate back-end systems.

  • Create a switch that transfers incoming requests from master to backup site.

Data mirroring

For the backup site to work, data on the master site must be mirrored to the backup regularly:

Component Description Mirror frequency
IBM MobileFirst Platform Foundation Database All tables must be mirrored. The exceptions to this rule are cache tables (SSO_LOGIN_CONTEXTS) and report tables (which are large in size). Highly dependent on implementation and can range from a few minutes to 24 hours. For more information, contact software support.
IBM MobileFirst Platform Foundation Software, customization, and content Any change in IBM MobileFirst Platform Foundation software, customization, or content must also be installed on the mirror servers. As it occurs.

Switching to back up site

When we switch to the backup site, some information might be lost:

  • All clients lose context and disconnect. In the case of an authenticated app, the user is prompted to log in again.

  • Report information is lost (unless previously mirrored).

  • Cache is lost. If Cache was implemented for various queries, an additional server fetch is required to fill cache.

Switching back to Master Site

Before you switch back to the master site, we must mirror the database back to the master site.

Important: The success of a recovery site is in the details. To ensure the successful functioning of such a site, we must develop and follow a strict written procedure, which you test regularly.


Parent topic: High availability