Merge Replication, on-demand PULL replication

Merge Replication, on-demand PULL replication

Post by siva » Tue, 24 Jul 2001 23:27:07



I am doing replication through internet. For this, I am
using Merge replication,   On demand-anonymous PULL
subscription.

My publisher database server will have a dedicated IP
address and it is permanently connected to internet. But
the Subscriber database server will connect to the
internet using dial-up connection (32k Modem). So I went
for anonymous subscription. I am able to pull the
publication data from the subscriber and it is working
fine.

My Problem is :-- If i change the same record's same field
value at both subscriber and publisher then the
replication is NOT working properly. If a record's same
column value is changed at both publisher and subscriber
then it should be marked as Conflict and SQL server should
solve it. But in anonymous subscription it is NOT working
properly.

I got the error "The Process could not log the Conflict"

Any help greatly appreciated.

Regards,
Siva.

 
 
 

Merge Replication, on-demand PULL replication

Post by siva » Tue, 24 Jul 2001 23:27:08


I am doing replication through internet. For this, I am
using Merge replication,   On demand-anonymous PULL
subscription.

My publisher database server will have a dedicated IP
address and it is permanently connected to internet. But
the Subscriber database server will connect to the
internet using dial-up connection (32k Modem). So I went
for anonymous subscription. I am able to pull the
publication data from the subscriber and it is working
fine.

My Problem is :-- If i change the same record's same field
value at both subscriber and publisher then the
replication is NOT working properly. If a record's same
column value is changed at both publisher and subscriber
then it should be marked as Conflict and SQL server should
solve it. But in anonymous subscription it is NOT working
properly.

I got the error "The Process could not log the Conflict"

Any help greatly appreciated.

Regards,
Siva.

 
 
 

Merge Replication, on-demand PULL replication

Post by siva » Tue, 24 Jul 2001 23:27:01


I am doing replication through internet. For this, I am
using Merge replication,   On demand-anonymous PULL
subscription.

My publisher database server will have a dedicated IP
address and it is permanently connected to internet. But
the Subscriber database server will connect to the
internet using dial-up connection (32k Modem). So I went
for anonymous subscription. I am able to pull the
publication data from the subscriber and it is working
fine.

My Problem is :-- If i change the same record's same field
value at both subscriber and publisher then the
replication is NOT working properly. If a record's same
column value is changed at both publisher and subscriber
then it should be marked as Conflict and SQL server should
solve it. But in anonymous subscription it is NOT working
properly.

I got the error "The Process could not log the Conflict"

Any help greatly appreciated.

Regards,
Siva.

 
 
 

Merge Replication, on-demand PULL replication

Post by siva » Tue, 24 Jul 2001 23:27:11


I am doing replication through internet. For this, I am
using Merge replication,   On demand-anonymous PULL
subscription.

My publisher database server will have a dedicated IP
address and it is permanently connected to internet. But
the Subscriber database server will connect to the
internet using dial-up connection (32k Modem). So I went
for anonymous subscription. I am able to pull the
publication data from the subscriber and it is working
fine.

My Problem is :-- If i change the same record's same field
value at both subscriber and publisher then the
replication is NOT working properly. If a record's same
column value is changed at both publisher and subscriber
then it should be marked as Conflict and SQL server should
solve it. But in anonymous subscription it is NOT working
properly.

I got the error "The Process could not log the Conflict"

Any help greatly appreciated.

Regards,
Siva.

 
 
 

Merge Replication, on-demand PULL replication

Post by TSVK » Wed, 25 Jul 2001 02:56:08


Where is the conflict table supposed to be in your case?  On publisher or
subscriber?  ( in other words is it centralized or decentralized logging
. ).  The default is centralized publisher.

Check to see if the anonymous subscription -PublisherLogin has permissions
to write to conflict table.

Also check to see if the sysmergearticles table has the correct entry for
conflict_table name.  If the entry is incorrect the problem must be with the
snapshot agent and object owner qualification of some sort.
Let me know.


Quote:> I am doing replication through internet. For this, I am
> using Merge replication,   On demand-anonymous PULL
> subscription.

> My publisher database server will have a dedicated IP
> address and it is permanently connected to internet. But
> the Subscriber database server will connect to the
> internet using dial-up connection (32k Modem). So I went
> for anonymous subscription. I am able to pull the
> publication data from the subscriber and it is working
> fine.

> My Problem is :-- If i change the same record's same field
> value at both subscriber and publisher then the
> replication is NOT working properly. If a record's same
> column value is changed at both publisher and subscriber
> then it should be marked as Conflict and SQL server should
> solve it. But in anonymous subscription it is NOT working
> properly.

> I got the error "The Process could not log the Conflict"

> Any help greatly appreciated.

> Regards,
> Siva.

 
 
 

Merge Replication, on-demand PULL replication

Post by siva_baskar siva_baska » Wed, 25 Jul 2001 12:19:15


Hi,

Thanks you very much for the info. Yeah. the problem lies with the conflict table. In my case the Conflict_<usertableName>, conflict table, should be in the publisher.

The problem is, when I am trying to PULL the data from the subscriber (one with dynamic IP address--connecting to internet through modem) using on demand, PULL based anonymous subscription, the conflict_<usertableName> table is NOT created in the publisher database. So the sysMergeArticles table is having NULL value for Conflict_table column.

I am working on the test database with sa as username with NO password. This account is valid for both (test) publisher database server and (test)subscriber database server. Since this account is having all the permissions in both publisher and subscriber, i guess there is NO permission problem.

The other thing I noticed is:- if I create the SAME type pf subscription (on the publisher database server) for the SAME publication then the Conflict_table is created in the Publisher database.

Interestingly, there after the dynamic IP address subscriber (one pulling the data thorough internet) is also working smoothly (WITHOUT that error: "The Process could not log the Conflict"). This is obvious, because now the publiher database is having the conflict table.

My Question: What is the general procedure for creating the conflict table in anonymous PULL based subscription, Merge replication?

PS: I guess that the conflict table will be created, the first time when the conflict records are occuring.

Please advice me..

Awaiting your valuable suggestion..

Regards,
Siva.

*** Sent via Developersdex http://www.developersdex.com ***
Don't just participate in USENET...get rewarded for it!

 
 
 

Merge Replication, on-demand PULL replication

Post by TSVK » Fri, 27 Jul 2001 03:08:50


When you are working with dynamic IP addresses the rule of thumb is try to
mask the IP address from the replication agents.  Have some kind of mapping
so that the replication agent neeeds to know only about the server name
rather than the IP address. This will avoid a lot of problems.
This is besides the problem you just had.

Once the conflict table is created there should not be a problem. Conflict
table is created by the snapshot agent for centralized logging in general.
But snapshot agent can be run on publisher or subscriber.
But if the table wasnt created during snapshot for some reason (error)  it
is created during merge.  You must have not had permission to create table
at several points of time.

If you had run the snapshot agent from publisher with a account that has
permissions to create table, there wont be such problems.


Hi,

Thanks you very much for the info. Yeah. the problem lies with the conflict
table. In my case the Conflict_<usertableName>, conflict table, should be in
the publisher.

The problem is, when I am trying to PULL the data from the subscriber (one
with dynamic IP address--connecting to internet through modem) using on
demand, PULL based anonymous subscription, the conflict_<usertableName>
table is NOT created in the publisher database. So the sysMergeArticles
table is having NULL value for Conflict_table column.

I am working on the test database with sa as username with NO password. This
account is valid for both (test) publisher database server and
(test)subscriber database server. Since this account is having all the
permissions in both publisher and subscriber, i guess there is NO permission
problem.

The other thing I noticed is:- if I create the SAME type pf subscription (on
the publisher database server) for the SAME publication then the
Conflict_table is created in the Publisher database.

Interestingly, there after the dynamic IP address subscriber (one pulling
the data thorough internet) is also working smoothly (WITHOUT that error:
"The Process could not log the Conflict"). This is obvious, because now the
publiher database is having the conflict table.

My Question: What is the general procedure for creating the conflict table
in anonymous PULL based subscription, Merge replication?

PS: I guess that the conflict table will be created, the first time when the
conflict records are occuring.

Please advice me..

Awaiting your valuable suggestion..

Regards,
Siva.

*** Sent via Developersdex http://www.developersdex.com ***
Don't just participate in USENET...get rewarded for it!