Whichever node has ownership of the quorum is the active node. If the quorum resource
fails, all nodes will go offline, because the nodes will not be able to know which is the active node.
Although this design provides high availability, it has some shortcomings. The challenges behind a
Single Copy Cluster are complexity, cost, and there are still single points of failure. To address these
issues, Microsoft added a continuous replication model. Other Microsoft technologies, such as Microsoft
SQL Server, already offer a similar form of clustering. In the continuous replication model, the cluster
nodes do not share disk resources. This solution addresses one of the single points of failure in SCC,
which is that a single copy of the data exists across all nodes. Now in continuous replication, each node
maintains its own copy of the data. How is this accomplished? To keep the nodes in sync, Exchange uses
asynchronous replication, also called log shipping. The first step in forming the cluster is to copy the
Exchange database to the passive node. This process is called seeding. Once the passive node has a full
copy of the database, it pulls closed log files from the active node. After the copy process completes and
the log file is checked for integrity, the passive node replays the log into its copy of the database.
Pages:
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467