Local Continuous Replication
(LCR) is another way in which Exchange Server 2007 provides improved
resiliency against failure. LCR works similarly to CCR in that a
secondary copy of the databases and log exist on separate storage. In
the case of LCR, the secondary copy exists on the same server. This
helps protect against media failures because the administrator can
quickly switch over to the secondary copy of the logs and databases.
Configuring LCR
Setting up LCR is a fairly simple process and
must be performed on each storage group that is to utilize LCR. To
enable Local Continuous Replication on a storage group, complete the
following steps:
1. | Click Start, Programs, Exchange, Exchange Management Console.
|
2. | In the left pane, expand Server Configuration.
|
3. | Click Mailbox.
|
4. | In the center pane, click the storage group for which you intend to enable LCR.
|
5. | From the Action bar, click Enable Local Continuous Replication.
|
6. | When the Setup Wizard launches, click Next.
|
7. | Browse
to the paths that will host the replica logs and system files. This
should be a different disk than the original storage group location.
Click Next.
|
8. | Browse
to the path that will host the replica database. This should be a disk
other than the one that hosts the primary copy of the database. Click
Next.
|
9. | Review the configuration. If it is correct, click Enable.
|
10. | When the configuration completes successfully, click Finish.
|
Recovering a Storage Group with LCR
In the event of a failure of either the primary
database disks or the primary log disks, you can switch over to the
secondary copy via the following process:
1. | Verify
that the corruption is not the result of an offline log drive, offline
database drive, or a disk volume configuration error. If the log volume
of the production storage group is not available (and could be
available) at the time of the failover, more data might be lost than
necessary.
|
2. | Assess
if the data in the passive copy of the database is acceptable.
Typically, the system should be able to recover with all data from the
active copy of the database. Thus, the assessment should show that all
necessary log files are available. If this is not the case, you should
investigate why some or all log files are unavailable.
|
3. | Disable LCR for the storage group containing the corrupt data.
|
4. | Dismount the corrupt database. You can use the Dismount-Database
cmdlet in the Exchange Management Shell or the Dismount shortcut menu
option for the database in the Exchange Management Console.
|
5. | If you determine the passive copy of the database is acceptable, activate it using either of the following methods:
If you are using the NTFS file system volume mount points for your production and LCR storage group and database files, use the Restore-StorageGroupCopy
cmdlet to copy all remaining logs and make the database mountable.
Remove the existing drive letter assignments for the mount points that
contain the storage group and database files. Create new drive letter
assignments that point to the mount points that contain the storage
group and database files. Continue to step 6. If
you are not using NTFS file system volume mount points for your
production and LCR storage group and database files, move the corrupted
files to a safe location. Use the Restore-StorageGroupCopy cmdlet with the ReplaceLocations
parameter to push the copy’s locations into the product locations. This
automatically attempts to copy the final logs. Continue to step 6.
|
6. | Use the Restore-StorageGroupCopy cmdlet as follows:
Restore-StorageGroupCopy -Identity:<Server>\<StorageGroupName> -ReplaceLocations:$true
|
7. | At the confirmation prompt, type Y, and then press Enter.
|
8. | The database can now be mounted.
|
9. | The Restore-StorageGroupCopy cmdlet automatically disables LCR for the storage group. LCR must be reenabled after the recovery is complete.
|
Limitations of LCR
When using LCR on a storage group, that storage
group is limited to a single database. This means that if you planned to
use more than 20 databases, you need to either put multiple databases
on a single hard drive or you need to mount the hard drives as logical
mount points. By using logical mount points, you can get around the
limitations on available drive letters.
To take full advantage of the redundancy offered
by LCR, you must place the local replica of the logs and databases on
drives other than those that host the primary copy. This way, if a drive
fails, the replica data is still safe and usable.
If an Exchange Server 2007 Mailbox server is configured as a CCR cluster, it will be unable to utilize LCR for local redundancy.
