Unexpected Replication

[Sean Rooney]

This is strange...

I used repadmin to disable outbound replication on my schema master. I then
did an ADprep on my schema master. It quarantined the changes on my server
for a period, but whilst monitoring in replmon I noticed that the changes
had managed to get to my other DC before I'd opened up outbound replication.

I know this sounds like BS, but it really happened like that. Anyone know of
any scenario in AD whereby replication would override my settings? I've seen
articles on forced replication, but not sure if that would cause this. Any
help appreciated.

 

[Guest]

Sean

urgent replication (lockouts, etc) ignores normal schedules, but I don't
think it has the ability to enable replication on the server (urgent gets
pushed anyway?).
Not much of an answer I know.

Gordon

 

[ptwilliams]

(urgent gets pushed anyway?).

Gordon,

Urgent replication triggers are pushed (like normal triggers) but outside of
the normal replication Window. They use the normal replication topology
(the existing connection objects and replication partners). Urgent
replication triggers are intra-site only; Intersite replication still
adheres to the original schedules.

Password changes (and one or two other changes) are pushed directly to the
PDCe via a discreet RPC call. This doesn't use the replication topology,
and doesn't care about replication windows, etc.


To Sean, the original poster...

Are you sure that outbound replication was disabled?
Have you made any configuration changes to the site links? Particularly the
options value?

 

[Sean Rooney]

Thanks for the info guys.

Definitely intra site was disabled on the schema master. It was the first
thing we checked, but it was still in place. Strange thing was we were
watching the replication in Replmon and it confirmed that the intra-site and
inter-site replication was definitely stopped, but within a 10 minute period
there was at least one successful replication cycle after that the
replication quarantine we'd setup kicked in again. So intra-site and
inter-site replication was still disabled. Seemed to kick in for some
unforseen reason.

It could have been urgent replication, the schema master was also the pdc
emulator, so that might explain it. Can't find any details on how urgent
replication works - does it work through all the queued replication and then
apply the urgent changes, or does it ignore everything else and just deal
with the urgent changes?

More questions than answers at this stage, but we've logged a call with MS.
If I hear anything I'll post it.

 

[ptwilliams]

Can't find any details on how urgent replication works

Stick this into Google: "urgent replication" trigger site:microsoft.com

Anyway, here's one or two articles to get you going:
--
http://www.microsoft.com/technet/pr...Ref/1465d773-b763-45ec-b971-c23cdc27400e.mspx
-- http://support.microsoft.com/default.aspx?id=232690

does it work through all the queued replication and then apply the urgent
changes, or does it ignore everything else and just deal with the urgent
changes?

Urgent replication is simply a change notification sent to direct,
intra-site replication partners outside of normal replication schedule
(straight-away, basically).

See the above link for more info.

More questions than answers at this stage, but we've logged a call with
MS. If I hear anything I'll post it.

I'd be interested in knowing what PSS have to say. Hope to hear from you
soon.