Datasheet

14 CHAPTER 1 INTRODUCING EXCHANGE SERVER 2010
Continuous Replication Basics
If we had to pick a single technology that is the most compelling in Exchange Server 2010,
it would be the continuous replication technology. This new technology supports the ability
to replicate a database to one or more additional Exchange Mailbox servers within your
organization.
Unlike many tools from third-party vendors, which replicate data either at the disk block
level or by taking snapshots of the disk and replicating changes, Exchange continuous replica-
tion is more similar to the SQL Server log shipping technology. This is considered similar to a
pull model, but it is the active copy of the database that does the work. The replication service
managing the passive copy of the database communicates with the active copy and indicates
which logs the passive copy needs to keep the database in sync. The active source Exchange
database, logs, and database engine do not even realize they are being copied. The Microsoft
Exchange Replication Service (Microsoft.Exchange.Cluster.ReplayService.exe)handles
copying the logs and managing the passive databases.
Initially (as when continuous replication is set up or reconfigured) the current copy of the
database is copied to the passive location; this is called seeding. As an Exchange t ransaction
log is filled up and renamed, (that is, when the E00.LOG file is filled and then renamed to
E000000001.LOG), the renamed and closed log file is then copied t o the passive location. The
information store service then verifies the log file and commits it to the passive copy of
the database. So the actual database file is not replicated at all, but it is kept in sync by
copying the log files and replaying them.
You will probably understand this concept better with an illustration. Figure 1.5 shows an
example o f how this process works. The Exchange database engine is run by the Microsoft
Exchange Information Store; transactions fill up the current transaction log (E00.LOG). The
transaction log file (E00.LOG) is renamed to the next a vailable transaction log filename (in this
case E0000000001.LOG). All of this is handled by the Information Store service.
Figure 1.5
How continuous
replication works
E00.LOG
Microsoft Exchange Information Store
Microsoft Exchange Replication Service
Active database
Active database transaction logs
Inspector directory Target log directory
Passive
database
E0000000001.LOG
If continuous replication is enabled, the Microsoft Exchange Replication Service copies
the E0000000001.LOG file to the Inspector directory. This folder exists on any server within the
database availability group (DAG) that has a copy of the database.
The service performs an intensive verification of the log files in the Inspector directory to
ensure they are not corrupted. Once the log files are verified as not being corrupted, they are