HP OpenVMS Systems

ask the wizard
Content starts here

Disaster recovery, hot standby site, cluster?

» close window

The Question is:

We are setting up the Hot Standby option with Oracle Rdb, but some of our
 important production data is in RMS files. The production and standby sites
 are networked and have a 45Mbps (DS3) connection.
What are my options for keeping a close to real-time copy of the flat files at
 our offsite location?
We are going to be upgrading VMS soon so if new features will be available for
 use soon also.

The Answer is :

  Why have a standby site when you can have a distributed cluster and
  can use the systems and the data from all hosts concurrently?
  By its very nature, a standby site without this capability is wasteful
  -- the resources at the standby site are unused in normal production
  save for the journaling- and BACKUP-related operations.
  Most customers would clearly prefer to get some benefit and some use
  from all available servers and these resources, prefering to see some
  degraded operations during failures.  (Further, regular use of these
  remote "spare" systems ensures that the remote site is actually
  operational, something you clearly don't want to have to determine
  during an actual fail-over.)
  With a distributed multi-site cluster, you can have local shadowset
  volumes containing the actual files on the volumes, given the clustering
  capabilities that are possible with a DS3-class (or faster) link.
  If you wish to maintain journaled and recoverable copies of your RMS
  files, the RMS journaling capabilities are available as well.
  Disaster-Tolerant Cluster Services consulting and support is available
  from HP.  In general, establishing a reliable and functional
  disaster-tolerant configuration (or even a fail-over configuration)
  is far more difficult than it might initially look, with many and
  particularly many subtle considerations involved.

answer written or last revised on ( 16-NOV-2002 )

» close window