By using asynchronous
communication, the principal and the mirror only need to do their best to keep
up. This may seem like a significant deterrent to using High Performance, but database
mirroring is designed to minimize the likelihood of the two nodes being out of sync, even
in this scenario.
The largest risk of data loss occurs when the two nodes sync up initially and transfer
new data from the transaction log of the primary database to the mirror database after it
has been restored without recovery. At that point, failure of the principal would guarantee
data loss if the mirror were brought online. Once the two nodes are synchronized,
database mirroring no longer uses the active transaction log at all when sending new
information over the wire; it comes directly from the in-memory log cache directly to the
service broker automation. Barring any huge data changes within a short period of time,
it??™s unlikely that the asynchronous connection will ever be out of sync for long. In the
event of a failure, there is no witness defined to facilitate automated failover.
CHAPTER 8 n DATABASE MIRRORING 206
Failover
During a failure situation, mirroring simply stops, and the mirror database is left in an
unrestored state.
Pages:
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395