Image For Disaster Recovery

JSS

Active Member
Hello - I have been requested by management to get JDE up and running at an alternate site from our current production site for disaster recovery. The alternate will have exact same servers as the current production servers hardware wise. My question is...has anyone tried to image their JDE servers and copy the images to another set of servers? Has this been successful? Or am I better of installing JDE from scratch on the alternate servers and then copy over the directories? Please share your experiences on this subject.

We are currently running 8.11 SP1 on SQL 2005(64 bit), Websphere 6.1 on all Windows 2003 servers
 
Jay,

Our strategy was to utilize microsoft clustering and replicated sans. My servers are clustered, with one node in each of our two data centers. The servers write their data to a SAN. The San writes it's data in real time to the san in the other data center. In case of a disaster, the infrastructure guys have to "break" the san replication, and then we bring up the alternate nodes in the DR room. If you are a subscriber to JDEtips Journal, you can read a case study about my setup. It was published in the January 2007 edition.

- Gregg
 
- HyperV server images for Enterprise, Dev machines, CreateForms, RFGen and Web servers
- Database log shipping every 15 minutes

The only physical hardware at the DR site is a Deployment server which is kept in sync for code by log shipping and weekly file copies to update the C files.

We have tested it twice with the busienss community and it runs great.

- 16 hour startup if pulling "newer" images from tape
- 8 hour startup if using "older" images and doing a full build
- 2-3 hour startup if using "older" images and running old code base for startup (with a full build starting up to refresh ASAP)

Easy to do if done properly.
 
Back
Top