Looking for suggestions for DRP (Disaster Recovery Plan) solutions

[erich13]

Looking for suggestions for DRP (Disaster Recovery Plan) solutions, like but not limited to using Ignite/UX and DRD (Dyanmic Root Disk). Like, tape or dvd backups of user data 

In general, I am looking for references to other people's DRP Best Practices, on HP-UX. 

And I don't just mean what tools/products are usefu. Although those would be useful too :-)    I am mostly looking for Best Practices of how to successfully combine and use tools/products.

[Ken Grabowski]

A Disaster recovery plan is a complex document that starts first with the business owners (not IT) defining what business services are important to be recovered in hours, days, weeks, etc.  There is a cost to everything that must be recovered and the business has to decide what they are willing to pay for.

From HP-UX you have the Ignite tools to create bootable images of the operating system. Those have to be combined with tape backups of the application/data sections of the server storage. Off site storage of the media for both of those along with documents defining how to recover and where to recover, and the source of recovery systems (ie. a fixed DR site, a service from a DR vendor, etc).

Depending on what your business has decided to invest, you may have multiple data centers with replicated data, storage, and ready to run servers on one extreme or  buying new equipment and recovering to that on the other.

There are numerous vendors, consultants, and reference books available on this subject.   You can't really answer this with a few suggestions from internet forums. Has anybody defined what needs to be recovered and at what priority? Is there an owner of the process?

[Bill Hassell]

And to add to Ken's comments...

Do you have just one HP-UX computer and nothing else (routers, switches, external storage, windows and/or Linux servers, etc)?  Or does your HP-UX computer use or depend on data from these external systems? A DR plan should always start with "smoke-and-rubble", that is, complete and total loss of the entire computer room or building. That's where the business costs for this loss must be evaluated, not only the replacement costs (insured or not) but the complete loss of the business data and compute power. Now you add the daily cost of this loss until the replacement machines are running, connected to a network and the OS and data restored to each machine. Now your DR plan would then outline each step to restore the lost systems. Finally, you run a DR rehearsal at least twice a year to verify that the steps are complete and actually work as designed.

Once a complete DR plan has been created (which should include partial disasters such as rm -rf /), then you should be protected. There are several realtime and near realtime recovery solutions which use both redundant local and remote systems sharing the data. With such designs, recovery may be only a few minutes.