Both methods have major
benefits and major shortcomings.
In the first scenario, you could create an automated job to generate snapshots
on an hourly basis. The snapshots should be named so that the time of day is easily
recognizable??”something like AdventureWorks_SS_0700. If you have a serious user error,
simply apply the appropriate snapshot. The upside of this technique is that there is
almost no downtime for your database; the downside is that there will be data loss.
If availability is your primary concern, this technique works nicely.
The second scenario works essentially the same way in that you create periodic
snapshots, but probably less frequently (maybe once or twice a day). When a user error
occurs, you can query the snapshot to see what the database looked like before the mistake
was made and generate a script to either replace the lost data or remove improper
information. The key benefit to this approach is that data loss is unlikely; the drawback
is that it could take quite some time to diagnose the situation, especially if you don??™t
have access to an automated database-compare tool.
Pages:
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439