Discussion:
How often does DFS replicate when set to 'Full' on 24x7 schedule?
(too old to reply)
Bo
2006-05-21 18:25:15 UTC
Permalink
I have 2 servers on the same site, and with DFS, both servers will
answer different users in the facility. This is good, it lets us load
balance. The problem is, user A creates a document and tells user B to
view it, and user B can't see it because it hasn't replicated yet. Is
there anything we can do to improve the speed? It's on a GB backbone
and set as noted in the subject - Full bandwidth 24x7.

Thanks in advance!
Ned Pyle (MSFT)
2006-05-21 23:32:51 UTC
Permalink
Hi Bo,

Ordinarily a pair of servers in the configuration you have described will
replicate with DFSR pretty much instantaneously. A couple questions:

1. Is RDC still enabled?
2. How large are these files, on average?
3. Has user A actually saved the file and closed it at that point?
4. How long is actually taking to copy (using the dfsrdebug logs will allow
you to say precisely).
5. Does DFSRDIAG.EXE show any backlog indicating your staging area quota is
too low?

Ned
Post by Bo
I have 2 servers on the same site, and with DFS, both servers will
answer different users in the facility. This is good, it lets us load
balance. The problem is, user A creates a document and tells user B to
view it, and user B can't see it because it hasn't replicated yet. Is
there anything we can do to improve the speed? It's on a GB backbone
and set as noted in the subject - Full bandwidth 24x7.
Thanks in advance!
Bo
2006-05-22 17:02:19 UTC
Permalink
Update:
I moved from using psexec to run a remote command prompt, to terminal
services and running dfsrdiag locally. The result is that it told me I
have 26030 backlogged files, however, the staging area still does not
show as full, and it told me this number (or very close to it)
yesterday as well, when the staging area was set to 20GB. Today it is
at 30GB

Thanks again for any help you can provide.

On Sun, 21 May 2006 19:32:51 -0400, "Ned Pyle \(MSFT\)"
Post by Ned Pyle (MSFT)
Hi Bo,
Ordinarily a pair of servers in the configuration you have described will
1. Is RDC still enabled?
2. How large are these files, on average?
3. Has user A actually saved the file and closed it at that point?
4. How long is actually taking to copy (using the dfsrdebug logs will allow
you to say precisely).
5. Does DFSRDIAG.EXE show any backlog indicating your staging area quota is
too low?
Ned
Post by Bo
I have 2 servers on the same site, and with DFS, both servers will
answer different users in the facility. This is good, it lets us load
balance. The problem is, user A creates a document and tells user B to
view it, and user B can't see it because it hasn't replicated yet. Is
there anything we can do to improve the speed? It's on a GB backbone
and set as noted in the subject - Full bandwidth 24x7.
Thanks in advance!
Bo
2006-05-22 17:13:45 UTC
Permalink
Additional info in case it helps you help me ;)


H:\>dfsrdiag propagationtest /rgname:"ocfs01 to ocfs02" /rfname:home
/member:tc0
33 /testfilename:dfsrdiag.txt /cleanup:dfsrdiag.txt
[WARNING] Failed to delete file:
\??\UNC\tc033.ourdomain.com\D$\Shares\Home\__DFSR
_DIAGNOSTICS_TEST_FOLDER__\dfsrdiag.txt Err: 2 (0x2)


Operation Succeeded


H:\>dfsrdiag propagationtest /rgname:"ocfs01 to ocfs02" /rfname:home
/member:tc0
33 /testfilename:dfsrdiag.txt /cleanup:dfsrdiag.txt /v
[INFO] Computer Name: tc033
[INFO] Computer DNS: tc033.ourdomain.com
[INFO] Domain Name: ourdomain
[INFO] Domain DNS: ourdomain.com
[INFO] Site Name: OC
[INFO] Connected to WMI services on computer: tc033.ourdomain.com
[INFO] Issuing query: SELECT * FROM DfsrReplicationGroupConfig WHERE
Replication
GroupName="ocfs01 to ocfs02"
[INFO] Found DfsrReplicationGroupConfig object, guid:
FA423A3D-4722-46CF-B76E-F9
60E519AFFF
[INFO] Issuing query: SELECT * FROM DfsrReplicatedFolderConfig WHERE
Replication
GroupGuid="FA423A3D-4722-46CF-B76E-F960E519AFFF" AND
ReplicatedFolderName="home"

[INFO] Found DfsrReplicatedFolderConfig object, guid:
2F3EF3DA-8903-4224-BA42-B2
0ADC202BA6
[INFO] RootPath: D:\Shares\Home

[INFO] Execution Time: 0 seconds
Operation Succeeded


H:\>dfsrdiag propagationtest /rgname:"ocfs01 to ocfs02" /rfname:home
/member:tc0
33 /testfilename:dfsdiag.txt /cleanup:dfsdiag.txt /v
[INFO] Computer Name: tc033
[INFO] Computer DNS: tc033.ourdomain.com
[INFO] Domain Name: ourdomain
[INFO] Domain DNS: ourdomain.com
[INFO] Site Name: OC
[INFO] Connected to WMI services on computer: tc033.ourdomain.com
[INFO] Issuing query: SELECT * FROM DfsrReplicationGroupConfig WHERE
Replication
GroupName="ocfs01 to ocfs02"
[INFO] Found DfsrReplicationGroupConfig object, guid:
FA423A3D-4722-46CF-B76E-F9
60E519AFFF
[INFO] Issuing query: SELECT * FROM DfsrReplicatedFolderConfig WHERE
Replication
GroupGuid="FA423A3D-4722-46CF-B76E-F960E519AFFF" AND
ReplicatedFolderName="home"

[INFO] Found DfsrReplicatedFolderConfig object, guid:
2F3EF3DA-8903-4224-BA42-B2
0ADC202BA6
[INFO] RootPath: D:\Shares\Home
[WARNING] Failed to delete file:
\??\UNC\tc033.ourdomain.com\D$\Shares\Home\__DFSR
_DIAGNOSTICS_TEST_FOLDER__\dfsdiag.txt Err: 2 (0x2)


[INFO] Execution Time: 0 seconds
Operation Succeeded


H:\>

On Sun, 21 May 2006 19:32:51 -0400, "Ned Pyle \(MSFT\)"
Post by Ned Pyle (MSFT)
Hi Bo,
Ordinarily a pair of servers in the configuration you have described will
1. Is RDC still enabled?
2. How large are these files, on average?
3. Has user A actually saved the file and closed it at that point?
4. How long is actually taking to copy (using the dfsrdebug logs will allow
you to say precisely).
5. Does DFSRDIAG.EXE show any backlog indicating your staging area quota is
too low?
Ned
Post by Bo
I have 2 servers on the same site, and with DFS, both servers will
answer different users in the facility. This is good, it lets us load
balance. The problem is, user A creates a document and tells user B to
view it, and user B can't see it because it hasn't replicated yet. Is
there anything we can do to improve the speed? It's on a GB backbone
and set as noted in the subject - Full bandwidth 24x7.
Thanks in advance!
Bo
2006-05-22 16:56:48 UTC
Permalink
Thanks for the troubleshooting tips Ned.

I turned off RDC as a troubleshooting measure, it has not improved the
situation.

The files I'm testing with are 0kb-1kb text files. The files of daily
use would be typical spreadsheets/word documents.

Yes, the files have been saved and closed.

These servers are no longer replicating to the point that one of the
servers is displaying several thousand backlogged files. I have
adjusted the staging quota to as high as 30GB and still see the same
challenges. I have not had the error message stating that the staging
area is over 90% and files will be deleted on the replicated folders
in question.

The diagnostic reports generated from the DFS Management tool say both
servers are healthy. When I try to run DFSRDiag from a command line,
it errors out stating that:
"C:\WINDOWS\system32>dfsrdiag backlog /receivingmember:tc033
/sendingmember:tc104
/rgname:"ocfs01 to ocfs02" /rfname:home
[ERROR] Failed to connect to WMI services on computer:
tc033.the-can.org

Operation Failed"




On Sun, 21 May 2006 19:32:51 -0400, "Ned Pyle \(MSFT\)"
Post by Ned Pyle (MSFT)
Hi Bo,
Ordinarily a pair of servers in the configuration you have described will
1. Is RDC still enabled?
2. How large are these files, on average?
3. Has user A actually saved the file and closed it at that point?
4. How long is actually taking to copy (using the dfsrdebug logs will allow
you to say precisely).
5. Does DFSRDIAG.EXE show any backlog indicating your staging area quota is
too low?
Ned
Post by Bo
I have 2 servers on the same site, and with DFS, both servers will
answer different users in the facility. This is good, it lets us load
balance. The problem is, user A creates a document and tells user B to
view it, and user B can't see it because it hasn't replicated yet. Is
there anything we can do to improve the speed? It's on a GB backbone
and set as noted in the subject - Full bandwidth 24x7.
Thanks in advance!
Ned Pyle (MSFT)
2006-05-23 01:01:44 UTC
Permalink
Whoa nelly! Thousands of backlogged files - there should be 0 backlogged
files unless the server is having big general communication problems, based
on your config. The fact that you cannot connect to WMI either speaks to
some larger issues.

If you open Perfmon from start | run and try to add counters, are all the
counters just random numbers or do they say things like process, processor,
etc?
Post by Bo
Thanks for the troubleshooting tips Ned.
I turned off RDC as a troubleshooting measure, it has not improved the
situation.
The files I'm testing with are 0kb-1kb text files. The files of daily
use would be typical spreadsheets/word documents.
Yes, the files have been saved and closed.
These servers are no longer replicating to the point that one of the
servers is displaying several thousand backlogged files. I have
adjusted the staging quota to as high as 30GB and still see the same
challenges. I have not had the error message stating that the staging
area is over 90% and files will be deleted on the replicated folders
in question.
The diagnostic reports generated from the DFS Management tool say both
servers are healthy. When I try to run DFSRDiag from a command line,
"C:\WINDOWS\system32>dfsrdiag backlog /receivingmember:tc033
/sendingmember:tc104
/rgname:"ocfs01 to ocfs02" /rfname:home
tc033.the-can.org
Operation Failed"
On Sun, 21 May 2006 19:32:51 -0400, "Ned Pyle \(MSFT\)"
Post by Ned Pyle (MSFT)
Hi Bo,
Ordinarily a pair of servers in the configuration you have described will
1. Is RDC still enabled?
2. How large are these files, on average?
3. Has user A actually saved the file and closed it at that point?
4. How long is actually taking to copy (using the dfsrdebug logs will allow
you to say precisely).
5. Does DFSRDIAG.EXE show any backlog indicating your staging area quota is
too low?
Ned
Post by Bo
I have 2 servers on the same site, and with DFS, both servers will
answer different users in the facility. This is good, it lets us load
balance. The problem is, user A creates a document and tells user B to
view it, and user B can't see it because it hasn't replicated yet. Is
there anything we can do to improve the speed? It's on a GB backbone
and set as noted in the subject - Full bandwidth 24x7.
Thanks in advance!
Bo
2006-05-23 20:27:43 UTC
Permalink
If you note in my follow-up, the inability to connect via WMI was only
from my workstation while using psexec (Mark Russinovich's port from
Unix) to execute a command prompt on TC104. When I used terminal
services, I did not receive the WMI errors. That being said, here are
the results of your question, in case it's still helpful in resolving
the challenge.

When using on my perfmon on my workstation, I am able to see the
counters for TC033 without a problem.

When using on my perfmon on my workstation, I am able to see the
counters for TC104 without a problem.


On Mon, 22 May 2006 21:01:44 -0400, "Ned Pyle \(MSFT\)"
Post by Ned Pyle (MSFT)
Whoa nelly! Thousands of backlogged files - there should be 0 backlogged
files unless the server is having big general communication problems, based
on your config. The fact that you cannot connect to WMI either speaks to
some larger issues.
If you open Perfmon from start | run and try to add counters, are all the
counters just random numbers or do they say things like process, processor,
etc?
Post by Bo
Thanks for the troubleshooting tips Ned.
I turned off RDC as a troubleshooting measure, it has not improved the
situation.
The files I'm testing with are 0kb-1kb text files. The files of daily
use would be typical spreadsheets/word documents.
Yes, the files have been saved and closed.
These servers are no longer replicating to the point that one of the
servers is displaying several thousand backlogged files. I have
adjusted the staging quota to as high as 30GB and still see the same
challenges. I have not had the error message stating that the staging
area is over 90% and files will be deleted on the replicated folders
in question.
The diagnostic reports generated from the DFS Management tool say both
servers are healthy. When I try to run DFSRDiag from a command line,
"C:\WINDOWS\system32>dfsrdiag backlog /receivingmember:tc033
/sendingmember:tc104
/rgname:"ocfs01 to ocfs02" /rfname:home
tc033.the-can.org
Operation Failed"
On Sun, 21 May 2006 19:32:51 -0400, "Ned Pyle \(MSFT\)"
Post by Ned Pyle (MSFT)
Hi Bo,
Ordinarily a pair of servers in the configuration you have described will
1. Is RDC still enabled?
2. How large are these files, on average?
3. Has user A actually saved the file and closed it at that point?
4. How long is actually taking to copy (using the dfsrdebug logs will allow
you to say precisely).
5. Does DFSRDIAG.EXE show any backlog indicating your staging area quota is
too low?
Ned
Post by Bo
I have 2 servers on the same site, and with DFS, both servers will
answer different users in the facility. This is good, it lets us load
balance. The problem is, user A creates a document and tells user B to
view it, and user B can't see it because it hasn't replicated yet. Is
there anything we can do to improve the speed? It's on a GB backbone
and set as noted in the subject - Full bandwidth 24x7.
Thanks in advance!
Ned Pyle (MSFT)
2006-05-24 00:04:48 UTC
Permalink
Sorry about that, I missed the other thread where you explained the psexec
piece.

Do you have KB908521 installed (if not, your DFSR Healthlogs will be
complaining about it). Without it, we could be seeing backlogs based on RPC
connectivity issues. Also, do you have MS05-019 (KB913446) installed - if
not then we may be running into a TCPIP bug that was introduced in SP1?
Post by Bo
If you note in my follow-up, the inability to connect via WMI was only
from my workstation while using psexec (Mark Russinovich's port from
Unix) to execute a command prompt on TC104. When I used terminal
services, I did not receive the WMI errors. That being said, here are
the results of your question, in case it's still helpful in resolving
the challenge.
When using on my perfmon on my workstation, I am able to see the
counters for TC033 without a problem.
When using on my perfmon on my workstation, I am able to see the
counters for TC104 without a problem.
On Mon, 22 May 2006 21:01:44 -0400, "Ned Pyle \(MSFT\)"
Post by Ned Pyle (MSFT)
Whoa nelly! Thousands of backlogged files - there should be 0 backlogged
files unless the server is having big general communication problems, based
on your config. The fact that you cannot connect to WMI either speaks to
some larger issues.
If you open Perfmon from start | run and try to add counters, are all the
counters just random numbers or do they say things like process, processor,
etc?
Post by Bo
Thanks for the troubleshooting tips Ned.
I turned off RDC as a troubleshooting measure, it has not improved the
situation.
The files I'm testing with are 0kb-1kb text files. The files of daily
use would be typical spreadsheets/word documents.
Yes, the files have been saved and closed.
These servers are no longer replicating to the point that one of the
servers is displaying several thousand backlogged files. I have
adjusted the staging quota to as high as 30GB and still see the same
challenges. I have not had the error message stating that the staging
area is over 90% and files will be deleted on the replicated folders
in question.
The diagnostic reports generated from the DFS Management tool say both
servers are healthy. When I try to run DFSRDiag from a command line,
"C:\WINDOWS\system32>dfsrdiag backlog /receivingmember:tc033
/sendingmember:tc104
/rgname:"ocfs01 to ocfs02" /rfname:home
tc033.the-can.org
Operation Failed"
On Sun, 21 May 2006 19:32:51 -0400, "Ned Pyle \(MSFT\)"
Post by Ned Pyle (MSFT)
Hi Bo,
Ordinarily a pair of servers in the configuration you have described will
1. Is RDC still enabled?
2. How large are these files, on average?
3. Has user A actually saved the file and closed it at that point?
4. How long is actually taking to copy (using the dfsrdebug logs will allow
you to say precisely).
5. Does DFSRDIAG.EXE show any backlog indicating your staging area quota is
too low?
Ned
Post by Bo
I have 2 servers on the same site, and with DFS, both servers will
answer different users in the facility. This is good, it lets us load
balance. The problem is, user A creates a document and tells user B to
view it, and user B can't see it because it hasn't replicated yet. Is
there anything we can do to improve the speed? It's on a GB backbone
and set as noted in the subject - Full bandwidth 24x7.
Thanks in advance!
Bo
2006-05-24 03:15:28 UTC
Permalink
No problem.

Both the hotfix, and the patch, are installed. The servers should be
up to date, and appears to be, based on Windows Server Update Server,
as well as MS Windows Update site.

I spent about 3 and a half hours with MS Second level support today
and the best they have so far is looking at my debug logs for hints.
This is a real puzzler.

On Tue, 23 May 2006 20:04:48 -0400, "Ned Pyle \(MSFT\)"
Post by Ned Pyle (MSFT)
Sorry about that, I missed the other thread where you explained the psexec
piece.
Do you have KB908521 installed (if not, your DFSR Healthlogs will be
complaining about it). Without it, we could be seeing backlogs based on RPC
connectivity issues. Also, do you have MS05-019 (KB913446) installed - if
not then we may be running into a TCPIP bug that was introduced in SP1?
Post by Bo
If you note in my follow-up, the inability to connect via WMI was only
from my workstation while using psexec (Mark Russinovich's port from
Unix) to execute a command prompt on TC104. When I used terminal
services, I did not receive the WMI errors. That being said, here are
the results of your question, in case it's still helpful in resolving
the challenge.
When using on my perfmon on my workstation, I am able to see the
counters for TC033 without a problem.
When using on my perfmon on my workstation, I am able to see the
counters for TC104 without a problem.
On Mon, 22 May 2006 21:01:44 -0400, "Ned Pyle \(MSFT\)"
Post by Ned Pyle (MSFT)
Whoa nelly! Thousands of backlogged files - there should be 0 backlogged
files unless the server is having big general communication problems, based
on your config. The fact that you cannot connect to WMI either speaks to
some larger issues.
If you open Perfmon from start | run and try to add counters, are all the
counters just random numbers or do they say things like process, processor,
etc?
Post by Bo
Thanks for the troubleshooting tips Ned.
I turned off RDC as a troubleshooting measure, it has not improved the
situation.
The files I'm testing with are 0kb-1kb text files. The files of daily
use would be typical spreadsheets/word documents.
Yes, the files have been saved and closed.
These servers are no longer replicating to the point that one of the
servers is displaying several thousand backlogged files. I have
adjusted the staging quota to as high as 30GB and still see the same
challenges. I have not had the error message stating that the staging
area is over 90% and files will be deleted on the replicated folders
in question.
The diagnostic reports generated from the DFS Management tool say both
servers are healthy. When I try to run DFSRDiag from a command line,
"C:\WINDOWS\system32>dfsrdiag backlog /receivingmember:tc033
/sendingmember:tc104
/rgname:"ocfs01 to ocfs02" /rfname:home
tc033.the-can.org
Operation Failed"
On Sun, 21 May 2006 19:32:51 -0400, "Ned Pyle \(MSFT\)"
Post by Ned Pyle (MSFT)
Hi Bo,
Ordinarily a pair of servers in the configuration you have described will
1. Is RDC still enabled?
2. How large are these files, on average?
3. Has user A actually saved the file and closed it at that point?
4. How long is actually taking to copy (using the dfsrdebug logs will allow
you to say precisely).
5. Does DFSRDIAG.EXE show any backlog indicating your staging area quota is
too low?
Ned
Post by Bo
I have 2 servers on the same site, and with DFS, both servers will
answer different users in the facility. This is good, it lets us load
balance. The problem is, user A creates a document and tells user B to
view it, and user B can't see it because it hasn't replicated yet. Is
there anything we can do to improve the speed? It's on a GB backbone
and set as noted in the subject - Full bandwidth 24x7.
Thanks in advance!
Ned Pyle (MSFT)
2006-05-24 13:07:33 UTC
Permalink
If you have a case number or engineer alias I can look over their shoulder
and assist as well.
Post by Bo
No problem.
Both the hotfix, and the patch, are installed. The servers should be
up to date, and appears to be, based on Windows Server Update Server,
as well as MS Windows Update site.
I spent about 3 and a half hours with MS Second level support today
and the best they have so far is looking at my debug logs for hints.
This is a real puzzler.
On Tue, 23 May 2006 20:04:48 -0400, "Ned Pyle \(MSFT\)"
Post by Ned Pyle (MSFT)
Sorry about that, I missed the other thread where you explained the psexec
piece.
Do you have KB908521 installed (if not, your DFSR Healthlogs will be
complaining about it). Without it, we could be seeing backlogs based on RPC
connectivity issues. Also, do you have MS05-019 (KB913446) installed - if
not then we may be running into a TCPIP bug that was introduced in SP1?
Post by Bo
If you note in my follow-up, the inability to connect via WMI was only
from my workstation while using psexec (Mark Russinovich's port from
Unix) to execute a command prompt on TC104. When I used terminal
services, I did not receive the WMI errors. That being said, here are
the results of your question, in case it's still helpful in resolving
the challenge.
When using on my perfmon on my workstation, I am able to see the
counters for TC033 without a problem.
When using on my perfmon on my workstation, I am able to see the
counters for TC104 without a problem.
On Mon, 22 May 2006 21:01:44 -0400, "Ned Pyle \(MSFT\)"
Post by Ned Pyle (MSFT)
Whoa nelly! Thousands of backlogged files - there should be 0 backlogged
files unless the server is having big general communication problems, based
on your config. The fact that you cannot connect to WMI either speaks to
some larger issues.
If you open Perfmon from start | run and try to add counters, are all the
counters just random numbers or do they say things like process, processor,
etc?
Post by Bo
Thanks for the troubleshooting tips Ned.
I turned off RDC as a troubleshooting measure, it has not improved the
situation.
The files I'm testing with are 0kb-1kb text files. The files of daily
use would be typical spreadsheets/word documents.
Yes, the files have been saved and closed.
These servers are no longer replicating to the point that one of the
servers is displaying several thousand backlogged files. I have
adjusted the staging quota to as high as 30GB and still see the same
challenges. I have not had the error message stating that the staging
area is over 90% and files will be deleted on the replicated folders
in question.
The diagnostic reports generated from the DFS Management tool say both
servers are healthy. When I try to run DFSRDiag from a command line,
"C:\WINDOWS\system32>dfsrdiag backlog /receivingmember:tc033
/sendingmember:tc104
/rgname:"ocfs01 to ocfs02" /rfname:home
tc033.the-can.org
Operation Failed"
On Sun, 21 May 2006 19:32:51 -0400, "Ned Pyle \(MSFT\)"
Post by Ned Pyle (MSFT)
Hi Bo,
Ordinarily a pair of servers in the configuration you have described will
1. Is RDC still enabled?
2. How large are these files, on average?
3. Has user A actually saved the file and closed it at that point?
4. How long is actually taking to copy (using the dfsrdebug logs will allow
you to say precisely).
5. Does DFSRDIAG.EXE show any backlog indicating your staging area
quota
is
too low?
Ned
Post by Bo
I have 2 servers on the same site, and with DFS, both servers will
answer different users in the facility. This is good, it lets us load
balance. The problem is, user A creates a document and tells user B to
view it, and user B can't see it because it hasn't replicated yet. Is
there anything we can do to improve the speed? It's on a GB backbone
and set as noted in the subject - Full bandwidth 24x7.
Thanks in advance!
Bo
2006-05-24 18:23:50 UTC
Permalink
MS Case#SRX060523604907

On Wed, 24 May 2006 09:07:33 -0400, "Ned Pyle \(MSFT\)"
Post by Ned Pyle (MSFT)
If you have a case number or engineer alias I can look over their shoulder
and assist as well.
Post by Bo
No problem.
Both the hotfix, and the patch, are installed. The servers should be
up to date, and appears to be, based on Windows Server Update Server,
as well as MS Windows Update site.
I spent about 3 and a half hours with MS Second level support today
and the best they have so far is looking at my debug logs for hints.
This is a real puzzler.
On Tue, 23 May 2006 20:04:48 -0400, "Ned Pyle \(MSFT\)"
Post by Ned Pyle (MSFT)
Sorry about that, I missed the other thread where you explained the psexec
piece.
Do you have KB908521 installed (if not, your DFSR Healthlogs will be
complaining about it). Without it, we could be seeing backlogs based on RPC
connectivity issues. Also, do you have MS05-019 (KB913446) installed - if
not then we may be running into a TCPIP bug that was introduced in SP1?
Post by Bo
If you note in my follow-up, the inability to connect via WMI was only
from my workstation while using psexec (Mark Russinovich's port from
Unix) to execute a command prompt on TC104. When I used terminal
services, I did not receive the WMI errors. That being said, here are
the results of your question, in case it's still helpful in resolving
the challenge.
When using on my perfmon on my workstation, I am able to see the
counters for TC033 without a problem.
When using on my perfmon on my workstation, I am able to see the
counters for TC104 without a problem.
On Mon, 22 May 2006 21:01:44 -0400, "Ned Pyle \(MSFT\)"
Post by Ned Pyle (MSFT)
Whoa nelly! Thousands of backlogged files - there should be 0 backlogged
files unless the server is having big general communication problems, based
on your config. The fact that you cannot connect to WMI either speaks to
some larger issues.
If you open Perfmon from start | run and try to add counters, are all the
counters just random numbers or do they say things like process, processor,
etc?
Post by Bo
Thanks for the troubleshooting tips Ned.
I turned off RDC as a troubleshooting measure, it has not improved the
situation.
The files I'm testing with are 0kb-1kb text files. The files of daily
use would be typical spreadsheets/word documents.
Yes, the files have been saved and closed.
These servers are no longer replicating to the point that one of the
servers is displaying several thousand backlogged files. I have
adjusted the staging quota to as high as 30GB and still see the same
challenges. I have not had the error message stating that the staging
area is over 90% and files will be deleted on the replicated folders
in question.
The diagnostic reports generated from the DFS Management tool say both
servers are healthy. When I try to run DFSRDiag from a command line,
"C:\WINDOWS\system32>dfsrdiag backlog /receivingmember:tc033
/sendingmember:tc104
/rgname:"ocfs01 to ocfs02" /rfname:home
tc033.the-can.org
Operation Failed"
On Sun, 21 May 2006 19:32:51 -0400, "Ned Pyle \(MSFT\)"
Post by Ned Pyle (MSFT)
Hi Bo,
Ordinarily a pair of servers in the configuration you have described will
1. Is RDC still enabled?
2. How large are these files, on average?
3. Has user A actually saved the file and closed it at that point?
4. How long is actually taking to copy (using the dfsrdebug logs will allow
you to say precisely).
5. Does DFSRDIAG.EXE show any backlog indicating your staging area
quota
is
too low?
Ned
Post by Bo
I have 2 servers on the same site, and with DFS, both servers will
answer different users in the facility. This is good, it lets us load
balance. The problem is, user A creates a document and tells user B to
view it, and user B can't see it because it hasn't replicated yet. Is
there anything we can do to improve the speed? It's on a GB backbone
and set as noted in the subject - Full bandwidth 24x7.
Thanks in advance!
Ned Pyle (MSFT)
2006-05-24 20:42:14 UTC
Permalink
Hi Bo,

Thanks, I'm trakcing it. You spoke with Wes and Gary today it looks like -
Gary is one of my engineers. I'll keep my eyes on it and if you need
anything feel free to ping me directly at ***@microsoft.com
(remove the .REMOVE :) ).

Ned
Post by Bo
MS Case#SRX060523604907
On Wed, 24 May 2006 09:07:33 -0400, "Ned Pyle \(MSFT\)"
Post by Ned Pyle (MSFT)
If you have a case number or engineer alias I can look over their shoulder
and assist as well.
Post by Bo
No problem.
Both the hotfix, and the patch, are installed. The servers should be
up to date, and appears to be, based on Windows Server Update Server,
as well as MS Windows Update site.
I spent about 3 and a half hours with MS Second level support today
and the best they have so far is looking at my debug logs for hints.
This is a real puzzler.
On Tue, 23 May 2006 20:04:48 -0400, "Ned Pyle \(MSFT\)"
Post by Ned Pyle (MSFT)
Sorry about that, I missed the other thread where you explained the psexec
piece.
Do you have KB908521 installed (if not, your DFSR Healthlogs will be
complaining about it). Without it, we could be seeing backlogs based on RPC
connectivity issues. Also, do you have MS05-019 (KB913446) installed - if
not then we may be running into a TCPIP bug that was introduced in SP1?
Post by Bo
If you note in my follow-up, the inability to connect via WMI was only
from my workstation while using psexec (Mark Russinovich's port from
Unix) to execute a command prompt on TC104. When I used terminal
services, I did not receive the WMI errors. That being said, here are
the results of your question, in case it's still helpful in resolving
the challenge.
When using on my perfmon on my workstation, I am able to see the
counters for TC033 without a problem.
When using on my perfmon on my workstation, I am able to see the
counters for TC104 without a problem.
On Mon, 22 May 2006 21:01:44 -0400, "Ned Pyle \(MSFT\)"
Post by Ned Pyle (MSFT)
Whoa nelly! Thousands of backlogged files - there should be 0 backlogged
files unless the server is having big general communication problems, based
on your config. The fact that you cannot connect to WMI either speaks to
some larger issues.
If you open Perfmon from start | run and try to add counters, are all the
counters just random numbers or do they say things like process, processor,
etc?
Post by Bo
Thanks for the troubleshooting tips Ned.
I turned off RDC as a troubleshooting measure, it has not improved the
situation.
The files I'm testing with are 0kb-1kb text files. The files of daily
use would be typical spreadsheets/word documents.
Yes, the files have been saved and closed.
These servers are no longer replicating to the point that one of the
servers is displaying several thousand backlogged files. I have
adjusted the staging quota to as high as 30GB and still see the same
challenges. I have not had the error message stating that the staging
area is over 90% and files will be deleted on the replicated folders
in question.
The diagnostic reports generated from the DFS Management tool say both
servers are healthy. When I try to run DFSRDiag from a command line,
"C:\WINDOWS\system32>dfsrdiag backlog /receivingmember:tc033
/sendingmember:tc104
/rgname:"ocfs01 to ocfs02" /rfname:home
tc033.the-can.org
Operation Failed"
On Sun, 21 May 2006 19:32:51 -0400, "Ned Pyle \(MSFT\)"
Post by Ned Pyle (MSFT)
Hi Bo,
Ordinarily a pair of servers in the configuration you have described will
1. Is RDC still enabled?
2. How large are these files, on average?
3. Has user A actually saved the file and closed it at that point?
4. How long is actually taking to copy (using the dfsrdebug logs
will
allow
you to say precisely).
5. Does DFSRDIAG.EXE show any backlog indicating your staging area
quota
is
too low?
Ned
Post by Bo
I have 2 servers on the same site, and with DFS, both servers will
answer different users in the facility. This is good, it lets us load
balance. The problem is, user A creates a document and tells user
B
to
view it, and user B can't see it because it hasn't replicated yet. Is
there anything we can do to improve the speed? It's on a GB backbone
and set as noted in the subject - Full bandwidth 24x7.
Thanks in advance!
Loading...