If everything runs smoothly, the only time that second node will ever be used is
when you install a service pack or need to reboot the active node due to some sort of
maintenance.
You also risk that the passive node will sit unused. Often hardware issues only crop
up when the server is under constant use. A good example is the RAID/SAN controller.
You??™ll only see problems with an I/O controller when it??™s under heavy use. If you never
run SQL Server for any period of time on the passive node, you won??™t be able to tell if the
hardware is defective. To deal with this risk, you can reverse the roles of the active and
passive nodes for a few weeks at a time during testing and sometimes for a brief time
while in production. In doing that, you??™re likely to discover a defect in hardware.
nNote If the hardware does have some minor manufacturing defect, it usually fails within the first 30
days. It??™s best to have the hardware running continually for at least two weeks before moving to production.
Never purchase hardware and put it immediately into production unless it??™s to replace an existing piece of
equipment that failed.
Pages:
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358