Solve this disaster scenario

(1) The setup relies on remote office people to change tapes. It is very likely that one day somebody will just forget to insert a tape. Luck will have it that this is the very night disaster strikes.

(2) Who monitors the backup successes and failures? There is potential here for people making assumptions about each others' responsibilities.

(3) Is the tape at each remote site sent off-site or is an off-site duplicate produced? Without it, this is not a DR solution.

(4) What if the tape is faulty? If it is a single tape device, the backup system cannot move on to the next tape.

(5) If funds are available, a more automated solution would be better. There are two options:

- Change the backup strategy to backup the data to the central site. This will require network funding. However, depending on the type of data held at the remote sites, there are a number of tools which specialise in low-bandwith backups.

- Change the storage strategy to a central model and deliver the data to the remote sites using thin client technology, if possible, or deploy local caching servers.

Is there a check in place to be certain that the tapes are actually being changed?

Is there a check in place to ensure that the backups are actually producing readable tapes?

Are the tapes being sent off site?

Is there a documented procedure for restoring these tapes?

Is there a server available to which the tapes can be restored? If not, how long will it take to obtain one in case of need?

If a tape is restored to another server at another site, will the users be able to accessthe data?

Is this "solution" only supposed to enable recovery of the data or must it recover the function of the remote office?

The primary flaw is the apparant lack of true disaster recovery planning. Sure, this works great if the file server died, or the file system became corrupt, they could restore it from tape.

The problem is that if there's a disaster, generally that means that more than just the server is affected. The entire computer room, office, city, state, etc. could be affected.

Using the infrastructure in place, I could create a disaster recovery system that would insure maximium data recovery per site. However, this is assuming that each remote office is at least 300 miles away from one another. Have the remote offices send their backups to the central location. If there's a disaster at one of those sites, the technicians could get new hardware, restore the tape they received, and ship the server out to the affected remote site.

If the remote sites are closer, say various places within a city, I would set up a SAN backup, where each remote site might back up another remote site and the 5th site is backed up by the central site. Better yet, put all the backup drives in the central location and run the backups from there.