Previously, a disk failover had to be performed at the RAID array level. For example,
say you had a number of items set up as clustered resources, including network shares.
If a SQL Server instance failed, it would have failed over the disk array that it used. If
you had any other resources on that disk, they would have failed over too. An instance
could fail for any number of reasons that didn??™t involve the disk subsystem. Since the
array sizes would be so large, management would usually want to leverage that space
for additional functions other than SQL, such as file shares. I worked with one client
who used the same data array to house both data for a SQL Server instance and the
home directories for each user??™s Windows login. The SQL Server instance failed, and
so did every home directory, causing a momentary but highly visible disruption in enduser
operation. Needless to say, the CEO wasn??™t happy that none of the drives mapped
when the users logged in. To make matters worse, the failure occurred at 7:55 a.m., just
as everyone was starting their day.
By using mount points, only the individual drive letters fail over, not the entire array.
Pages:
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350