Failback in a Cluster
Just as important as preparing for failover, you should invest time in planning the
process to use in restoring the failed node. Restoring a failed node to its former position
in the cluster is referred to as failback. Will things fail back automatically, or will
you have to move them back over to the other node manually? As Figure 7-6 shows, if
you choose to fail back automatically, you can configure how long the repaired node
will need to be up before the cluster services decides it is restored. You can also set the
failback process for a specific resource, or you can group resources together and have
them behave as one.
Figure 7-6. Here SQL Agent is set to restart automatically in a failback situation.
CHAPTER 7 n CLUSTERING 191
Personally, I prefer to control failback manually. As long as I have the resources to
run both applications on a single node with minimal performance impact, I like to control
exactly what happens when. I??™ve encountered scenarios in which the replacement
server is installed and ready to receive cluster resources, but it hasn??™t been completely
rebuilt.
Pages:
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370