That way, you can kill the individual backup if it has a significantly negative
impact on the usability of the application.
nNote Be careful who you inform about a backup test. Informing all the end users might result in false
positives: often when told that they might see a performance impact, users will look for the slightest change
in performance. You may receive reports that it takes 20 seconds to do what normally takes 2 seconds, but
the real increase may be only a second or two. This isn??™t a case of deliberate exaggeration; it??™s just a trick
the mind can play on us. Instead, inform managers and supervisors to be on the lookout for reactions from
staff. You still may get a false-positive report of significant degradation, but at least you won??™t be planting the
seed that there could be performance issues.
CHAPTER 5 n CREATING A BACKUP/RECOVERY PLAN 103
Minimizing the time to back up usually involves frequent log backups and/or frequent
file and filegroup backups. Also, consider examining the overall system load
throughout the day, looking for regularly occurring low spots (perhaps right before and
after a shift change, or during commute times for online commercial applications).
Pages:
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219