IBM FileNet P8, Version 5.2.1            

Disaster recovery

The section on high availability discusses putting into place systems and devices that assure your FileNet® P8 system is available to your customers and employees with as little downtime as possible. High availability addresses localized failures such as a hard disk crash or server failure. But what if your entire production system is lost due to a catastrophic event such as earthquake or flood? Disaster recovery goes a step further than high availability by planning replication of your entire data center.

The first step in disaster recovery planning is to decide which approach you want to take:

The low-cost option is to back up all application data at regular intervals (preferably daily) as well as making less frequent full backups, and to keep copies of the data off site. All data is then available, but setting up a replacement system after a disaster occurs can take days or weeks. The danger of this solution is that the application is unavailable for a time after a disaster occurs and some amount of data (data that was changed or created after the last backup) is permanently lost. In addition, this solution is difficult to test on a regular basis because the replacement system is assembled only after a disaster occurs.

A hot site recovery service from a vendor is a more expensive solution that offers quicker recovery, based on dedicated hardware for the recovery system. This service consists of a shared resource environment provided for disaster recovery on a subscription basis. As with the low-cost option, periodic backups are stored off site. In this case, though, hardware and other resources are standing by at a vendor disaster recovery site that is shared with other clients of the vendor. IBM® Professional Services offers a hot site recovery service in partnership with SunGard. SunGard provides the server recovery facility, the hardware infrastructure, telecommunications, and workstation recovery capabilities, and IBM delivers dedicated technical support, software, and all necessary documentation. Regular testing is done to ensure the validity of the recovery. This solution provides end-to-end service for both clients and servers.

The most expensive solution, but the one that offers the quickest and most complete recovery, is a redundant data center at a remote disaster recovery site, outside the potential disaster impact zone. The recovery site has a dedicated copy of the production system hardware, possibly without the extra servers that were needed to make the production site highly available. Rather than relying on data backups and restores, this approach typically employs a storage solution from vendors like EMC, Network Appliance, or Hitachi that enables data replication to the recovery site. At their best, these replication products can provide real-time protection of the data, so that the data, right up to the moment of disaster, is available at the recovery site. They also offer less costly options for replication that allow some lag in the replication, and hence a small loss of data in a disaster. For transactional systems like FileNet P8, the optimal option is a hot standby recovery site as shown in the following diagram. A hot standby recovery site reduces the complexity of the disaster recovery solution and avoids conflict resolution problems between active sites. This solution can be tested on a regular basis.

Diagram showing the relationship of a production site to a standby disaster recovery site



Last updated: October 2015
p8sov094.htm

© Copyright IBM Corporation 2015.