AD Replication is occurring slowly.

Discussion:

(too old to reply)

Alan

2008-01-31 14:58:01 UTC

I contiously receive this alert as well as the related "AD Replication is
slower than the configured threshold" event; however, every time I verify
replication using "repadmin /replsum /bysrc /bydest /sort:delta" I confirm
that all domain controllers are replicating (InterSite) within 15 minutes.

I even tried bumping up the threshold to 30 and 45 minutes on the alert, but
I still get reported latencies of up to 50 minutes which is not accurate of
actual replication latencies which are all well within both the 15 minute
IntraSite and 5 minute IntraSite values.

Any ideas? Am I missing something within the Overrides for these alerts?

Thanks in advance,
Alan

Thomas CR

2008-02-02 19:24:02 UTC

Permalink

have you checked that the Action account have the right permission to update
the object, that is used to check replication.
have you also tried to check the obejcts to see if they gets updated.
--
BR
Thomas CR

Post by Alan
I contiously receive this alert as well as the related "AD Replication is
slower than the configured threshold" event; however, every time I verify
replication using "repadmin /replsum /bysrc /bydest /sort:delta" I confirm
that all domain controllers are replicating (InterSite) within 15 minutes.
I even tried bumping up the threshold to 30 and 45 minutes on the alert, but
I still get reported latencies of up to 50 minutes which is not accurate of
actual replication latencies which are all well within both the 15 minute
IntraSite and 5 minute IntraSite values.
Any ideas? Am I missing something within the Overrides for these alerts?
Thanks in advance,
Alan

Alan

2008-02-02 23:04:00 UTC

Permalink

Yes, I verified permissions and the Action account has the right to modify
the objects. I also verified that the MOMLatencyMonitors objects are getting
updated; each one has a current timestamp for both the whenChanged and
modifyTimeStamp attributes. In addition, these attribute values change within
acceptable timeframes for the settings on the monitor.

I'm stumped on this one. I even increased the latency thresholds to 2 hours
and it still triggers alerts even though repadmin shows latency below 15
minutes for IntraSite. My DC's are all connected via MPLS with the branch
offices having a T3 and the hub sites having a 45Mb pipe. It's not a true
latency issue; however, the alert continuously triggers as if there were
issues with latency.

-Alan

Post by Thomas CR
have you checked that the Action account have the right permission to update
the object, that is used to check replication.
have you also tried to check the obejcts to see if they gets updated.
--
BR
Thomas CR

Thomas CR

2008-02-03 19:06:00 UTC

Permalink

and all the objects are getting updated ? old servers are not automatic
deleted.

As I read you response, the objects are gettings updated, but you still get
the errors. Then I would call PSS. :-)
--
BR
Thomas CR

Post by Alan
Yes, I verified permissions and the Action account has the right to modify
the objects. I also verified that the MOMLatencyMonitors objects are getting
updated; each one has a current timestamp for both the whenChanged and
modifyTimeStamp attributes. In addition, these attribute values change within
acceptable timeframes for the settings on the monitor.
I'm stumped on this one. I even increased the latency thresholds to 2 hours
and it still triggers alerts even though repadmin shows latency below 15
minutes for IntraSite. My DC's are all connected via MPLS with the branch
offices having a T3 and the hub sites having a 45Mb pipe. It's not a true
latency issue; however, the alert continuously triggers as if there were
issues with latency.
-Alan

Alan

2008-02-03 23:16:00 UTC

Permalink

Hehe, I was hoping you wouldn't say that. One more question. I'm runing SCOM
side by side with MOM 2005 which is also running the AD MP. Could this be
posing an issue? MOM 2005 is currently reporting correctly on AD replication,
only SCOM return incorrect alerts for replication.

Thanks,
Alan

Post by Thomas CR
and all the objects are getting updated ? old servers are not automatic
deleted.
As I read you response, the objects are gettings updated, but you still get
the errors. Then I would call PSS. :-)
--
BR
Thomas CR

Robin Drake

2008-02-07 11:46:30 UTC

Permalink

Have you understood that thresholds for AD Replication latencies have to be
overriden in 5 separate places identically? If you fail to do so the 5
wrkflows will work independently with their own values. (1 Monitor + 4
rules allfor AD Replication)

Post by Alan
Hehe, I was hoping you wouldn't say that. One more question. I'm runing SCOM
side by side with MOM 2005 which is also running the AD MP. Could this be
posing an issue? MOM 2005 is currently reporting correctly on AD replication,
only SCOM return incorrect alerts for replication.
Thanks,
Alan

Alan

2008-02-07 12:46:01 UTC

Permalink

No, I was not aware that it was in five places. I dug a little deeper and
have overidden the "IntersiteExpectedMaxLatency" parameter for the following
rules:
1. AD Replication Performance Collection - Metric Replication Latency:Average
2. AD Replication Performance Collection - Metric Replication Latency:Maximum
3. AD Replication Performance Collection - Metric Replication Latency
4. AD Replication Performance Collection - Metric Replication Latency:Minimum

I've also overidden the following monitor:
1. AD Replication Monitoring

All five now have the same value for the parameter. I will continue to
monitor once the changes are pushed to the agents. I've been stumped by this
one for several days, thanks for the tip!

-Alan

Post by Robin Drake
Have you understood that thresholds for AD Replication latencies have to be
overriden in 5 separate places identically? If you fail to do so the 5
wrkflows will work independently with their own values. (1 Monitor + 4
rules allfor AD Replication)

Post by Alan
I contiously receive this alert as well as the related "AD Replication is
slower than the configured threshold" event; however, every time I verify
replication using "repadmin /replsum /bysrc /bydest /sort:delta" I
confirm
that all domain controllers are replicating (InterSite) within 15
minutes.
I even tried bumping up the threshold to 30 and 45 minutes on the
alert, but
I still get reported latencies of up to 50 minutes which is not
accurate of
actual replication latencies which are all well within both the 15 minute
IntraSite and 5 minute IntraSite values.
Any ideas? Am I missing something within the Overrides for these alerts?
Thanks in advance,
Alan

LearningCurve

2008-02-14 20:24:57 UTC

Permalink

Good stuff Alan.

I too continue to struggle with this under-developed MP, a port from MOM 05
with
"no" new features I can find. Regardless, are you guys seeing high-CPU
utilization on
your less-powered DC's. After my DC guys continue to complain about the
increase
in processor utilization and the "major" differences between the MOM 05
agent and the
07 agent in the context of processor utilization, I finally had an epiphany,
it appears to
me that basically 90% of the components within the 07 version of this MP are
"on" by
default, whereas, if you look at the 05 rules/monitors, a much lower
percentage of rules
are enabled by default.

Things that make you go.. hmmmmm.

-richard

Post by Alan
No, I was not aware that it was in five places. I dug a little deeper and
have overidden the "IntersiteExpectedMaxLatency" parameter for the following
1. AD Replication Performance Collection - Metric Replication
Latency:Average
2. AD Replication Performance Collection - Metric Replication
Latency:Maximum
3. AD Replication Performance Collection - Metric Replication Latency
4. AD Replication Performance Collection - Metric Replication
Latency:Minimum
1. AD Replication Monitoring
All five now have the same value for the parameter. I will continue to
monitor once the changes are pushed to the agents. I've been stumped by this
one for several days, thanks for the tip!
-Alan

Post by Thomas CR
and all the objects are getting updated ? old servers are not automatic
deleted.
As I read you response, the objects are gettings updated, but you
still
get
the errors. Then I would call PSS. :-)
--
BR
Thomas CR

Post by Alan
Yes, I verified permissions and the Action account has the right to modify
the objects. I also verified that the MOMLatencyMonitors objects are getting
updated; each one has a current timestamp for both the whenChanged and
modifyTimeStamp attributes. In addition, these attribute values
change
within
acceptable timeframes for the settings on the monitor.
I'm stumped on this one. I even increased the latency thresholds to
2
hours
and it still triggers alerts even though repadmin shows latency
below
15
minutes for IntraSite. My DC's are all connected via MPLS with the branch
offices having a T3 and the hub sites having a 45Mb pipe. It's not a true
latency issue; however, the alert continuously triggers as if there were
issues with latency.
-Alan

Post by Thomas CR
have you checked that the Action account have the right permission
to
update
the object, that is used to check replication.
have you also tried to check the obejcts to see if they gets updated.
--
BR
Thomas CR

Post by Alan
I contiously receive this alert as well as the related "AD
Replication is
slower than the configured threshold" event; however, every time
I
verify
replication using "repadmin /replsum /bysrc /bydest /sort:delta" I
confirm
that all domain controllers are replicating (InterSite) within 15
minutes.
I even tried bumping up the threshold to 30 and 45 minutes on the
alert, but
I still get reported latencies of up to 50 minutes which is not
accurate of
actual replication latencies which are all well within both the
15
minute
IntraSite and 5 minute IntraSite values.
Any ideas? Am I missing something within the Overrides for these
alerts?
Thanks in advance,
Alan

Jennifer Chu [MSFT]

2008-05-14 15:43:40 UTC

Permalink

Hi Richard,

Did you have to disable some of the monitor/rules in ADMP 2007 in order to
free up the CPU?

Thanks!
Jennifer

Post by LearningCurve
Good stuff Alan.
I too continue to struggle with this under-developed MP, a port from MOM
05 with
"no" new features I can find. Regardless, are you guys seeing high-CPU
utilization on
your less-powered DC's. After my DC guys continue to complain about the
increase
in processor utilization and the "major" differences between the MOM 05
agent and the
07 agent in the context of processor utilization, I finally had an
epiphany, it appears to
me that basically 90% of the components within the 07 version of this MP
are "on" by
default, whereas, if you look at the 05 rules/monitors, a much lower
percentage of rules
are enabled by default.
Things that make you go.. hmmmmm.
-richard

Post by Alan
Hehe, I was hoping you wouldn't say that. One more question. I'm
runing
SCOM
side by side with MOM 2005 which is also running the AD MP. Could this be
posing an issue? MOM 2005 is currently reporting correctly on AD replication,
only SCOM return incorrect alerts for replication.
Thanks,
Alan

Post by Thomas CR
and all the objects are getting updated ? old servers are not automatic
deleted.
As I read you response, the objects are gettings updated, but you
still
get
the errors. Then I would call PSS. :-)
--
BR
Thomas CR

Post by Alan
Yes, I verified permissions and the Action account has the right to modify
the objects. I also verified that the MOMLatencyMonitors objects
are
getting
updated; each one has a current timestamp for both the whenChanged and
modifyTimeStamp attributes. In addition, these attribute values
change
within
acceptable timeframes for the settings on the monitor.
I'm stumped on this one. I even increased the latency thresholds to
2
hours
and it still triggers alerts even though repadmin shows latency
below
15
minutes for IntraSite. My DC's are all connected via MPLS with the branch
offices having a T3 and the hub sites having a 45Mb pipe. It's not
a
true
latency issue; however, the alert continuously triggers as if there were
issues with latency.
-Alan

Post by Thomas CR
have you checked that the Action account have the right
permission to
update
the object, that is used to check replication.
have you also tried to check the obejcts to see if they gets updated.
--
BR
Thomas CR

Post by Alan
I contiously receive this alert as well as the related "AD
Replication is
slower than the configured threshold" event; however, every
time I
verify
replication using "repadmin /replsum /bysrc /bydest
/sort:delta" I
confirm
that all domain controllers are replicating (InterSite) within 15
minutes.
I even tried bumping up the threshold to 30 and 45 minutes on the
alert, but
I still get reported latencies of up to 50 minutes which is not
accurate of
actual replication latencies which are all well within both the
15
minute
IntraSite and 5 minute IntraSite values.
Any ideas? Am I missing something within the Overrides for these
alerts?
Thanks in advance,
Alan

Jennifer Chu [MSFT]

2008-06-07 01:59:45 UTC

Permalink

This issue should be improved in ADMP 2007 SP1. The beta will be released
in Q3CY2008. If you're interested in participating the beta program, please
let me know and I will provide the instructions.

Post by Jennifer Chu [MSFT]
Hi Richard,
Did you have to disable some of the monitor/rules in ADMP 2007 in order to
free up the CPU?
Thanks!
Jennifer

Post by Alan
No, I was not aware that it was in five places. I dug a little deeper and
have overidden the "IntersiteExpectedMaxLatency" parameter for the following
1. AD Replication Performance Collection - Metric Replication Latency:Average
2. AD Replication Performance Collection - Metric Replication Latency:Maximum
3. AD Replication Performance Collection - Metric Replication Latency
4. AD Replication Performance Collection - Metric Replication Latency:Minimum
1. AD Replication Monitoring
All five now have the same value for the parameter. I will continue to
monitor once the changes are pushed to the agents. I've been stumped by this
one for several days, thanks for the tip!
-Alan

Post by Alan
Hehe, I was hoping you wouldn't say that. One more question. I'm
runing
SCOM
side by side with MOM 2005 which is also running the AD MP. Could this be
posing an issue? MOM 2005 is currently reporting correctly on AD replication,
only SCOM return incorrect alerts for replication.
Thanks,
Alan

Post by Thomas CR
and all the objects are getting updated ? old servers are not automatic
deleted.
As I read you response, the objects are gettings updated, but you
still
get
the errors. Then I would call PSS. :-)
--
BR
Thomas CR

Post by Alan
Yes, I verified permissions and the Action account has the right
to
modify
the objects. I also verified that the MOMLatencyMonitors objects
are
getting
updated; each one has a current timestamp for both the whenChanged and
modifyTimeStamp attributes. In addition, these attribute values
change
within
acceptable timeframes for the settings on the monitor.
I'm stumped on this one. I even increased the latency thresholds
to 2
hours
and it still triggers alerts even though repadmin shows latency
below
15
minutes for IntraSite. My DC's are all connected via MPLS with the branch
offices having a T3 and the hub sites having a 45Mb pipe. It's not
a
true
latency issue; however, the alert continuously triggers as if
there
were
issues with latency.
-Alan

Post by Thomas CR
have you checked that the Action account have the right
permission to
update
the object, that is used to check replication.
have you also tried to check the obejcts to see if they gets updated.
--
BR
Thomas CR

Post by Alan
I contiously receive this alert as well as the related "AD
Replication is
slower than the configured threshold" event; however, every
time I
verify
replication using "repadmin /replsum /bysrc /bydest /sort:delta" I
confirm
that all domain controllers are replicating (InterSite) within 15
minutes.
I even tried bumping up the threshold to 30 and 45 minutes on the
alert, but
I still get reported latencies of up to 50 minutes which is not
accurate of
actual replication latencies which are all well within both
the 15
minute
IntraSite and 5 minute IntraSite values.
Any ideas? Am I missing something within the Overrides for these
alerts?
Thanks in advance,
Alan