Keep in mind that you must configure each active node so that in the event of a failure,
all nodes can run on a single server without clobbering each other. In an eight-node
active environment, that means a lot of configuration and a lot of testing. I??™minterested
to see one of these in production (as of this writing, I haven??™t).
If you have more than three physical nodes, I strongly advocate using the N+1 principle.
One server should always be sitting fallow, waiting to help out in the event of a
hardware failure. That node should always be physically as ???beefy??? as the strongest node
in your cluster. Trust me, when the hardware failure occurs, you??™ll be happy you configured
it that way. An active/active configuration on a two-node cluster requires more care
in configuration, but it??™s an acceptable situation. If you have the funds for a cluster of
three or more nodes, there really is no excuse for not having a passive node for failure
scenarios.
CHAPTER 7 n CLUSTERING 186
Figure 7-4. A four-node cluster
Multiple Instances
While a standalone installation supports 50 instances, a clustered environment can
only support 24 instances.
Pages:
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362