In a complex model of several dc’s, replication would be
complex since the dc’s will replicate the same data in an endless manner. Here
comes UDTV for the help.
Up-to-dateness vector (UTD or UDTV) is a replication
attribute which avoids endless replication. Each dc maintains a UDTV table for
every other dc’s. This is as per the following naming context scenario:
For Domain NC, each dc in a domain maintains every other dc’s
UDTV values.
For
schema and Configuration NC’s, this is done for every dc in the forest.
UDTV mainly keeps track of two things:
- Highest USN that each dc has received from its replication partners
- Highest USN that it has received from every dc that replicates a given NC
Each replicated change contains:
-
GUID of dc that is replicating the change.2. USN from the dc that is replicating the change.3. GUID of dc that originated the change.4. USN from the dc that originated the change.
Let us consider an example of 3 dc’s – DC1, DC2 and DC3.
Currently all dc’s are updated.
Now DC3 performs an update in its dc like a password reset.
As usual, DC3 notifies DC1 and DC2 about the change and the change is
replicated to both the dc’s as shown below:
Now there is a chance to contact DC1 from DC2, for the same
replication given by DC3 to DC2. DC1 contacts DC2 with the following information’s:
DC2’s
GUID : 4501
DC3’s
GUID : 7002
Based
on DC2’s HWMV received (4501) it assumes that it has not replicated the change.
But then it again checks for DC3’s GUID and sees that this initiated
replication doesn’t actually needed (since the DC3’s GUID sent by DC1 is same
as that of DC2).
At
this point DC1 updates its HWMV entry for DC2, and this process is carried out
until all the DC’s are updated with other DC’s current values.
*Thanks to Microsoft official technet site for the images
*Thanks to Microsoft official technet site for the images
No comments:
Post a Comment