sql server replication: While connected to the Subscriber, the Merge Agent connects to the distribution database at the Distributor and then to the publication database at the Publisher. The Merge Agent also reads from the snapshot folder when applying the snapshot files, so getting the correct level of security is really important. If you right-click on your pull subscription and go to properties on the security tab, choose to impersonate the SQL Agent on the distributor and on the publisher. Ideally, you have the SQL Agent's login on the Subscriber as the same SQL Agent's login on the Publisher/Distributor. This way if it is the usual recommended setup, it is bound to work as this account will be in the sysadmin role and will be a local administrator on the publisher (so can easily access the snapshot files). Please let us know if this impersonation works
Regards

Budiono
I just thought of something elae. This may be the case if the server has been renamed. You may alternatively see these cryptic errors :
18483 - could not connect to server 'Newname' because distributor_admin is not defined as a remote login at the server.
18482 - could not connect to server 'Newname' because 'Oldname' is not defined as a remote server
Try:
Use Master
go
Select @@Servername

This should return your current server name but if it
returns NULL then try:
Use Master
go
Sp_DropServer 'Server1'
GO
Use Master
go
Sp_Addserver 'Server1', 'local'
GO
Regards
Paul Ibiso

hi, i'm new in replication of database.
i've 2 cpu. cpu A as main sql server publisher/distributor and cpu B as
subscriber.
i already publish 2 table on cpu A and set cpu B as subscriber to pull 2
tables on cpu A and insert into cpu B. but it failed with error message 'the
proces could not connect to distributor A'.
what's wrong with this?
i already try open database on cpu B using enterprise manager on cpu A and
it success. also i open database on cpu A using enterprise manager on cpu B
and it success too. all the process, i use 'sa' login account.
anybody could help solve this problem?

thanks :)

Budiono
I can't really go through all the options on a newsgroup - there are too many! To set up transactional replication have a look at BOL

"replication, overview
"replication, properties

To set up the security

"Agent Login Security
"Connecting to the Distributor
"Replication Security
"Security Considerations (Location: Replication)

Alternatively Alexander Chigrik did an article on setting it up - http://www.mssqlcity.com/Articles/Replic/SetupTR/SetupTR.htm (other setup guides referred to on www.replicationanswers.com, links page)

However, I think that your particular error message is because the agent is not set to use impersonation, or the sql login is not correct if you are using sql authentication. Have a look at my first post and hopefully this will help

Regards
Paul Ibiso

thanks for reply
i never change the server name. the server name and sql server name is the
same.
i've execute select @@severname from CPU 'A' and it's returns 'A'. also on
CPU 'B'
the error message on subscriber when want to connect to publisher on CPU 'A'
is 'the process could not connect to distributor CPU 'B'.
i can't drop the publisher coz it's production database. so, i've the
production database 'DB1' on CPU 'A'. and have 2 table published. named
'TBL_A' and 'TBL_B'. and want to make CPU 'B' as subscriber to pull the 2
table into 'DB2' on CPU 'B'.
would you like teach me step by step how to do it? and i want it as
'transactional publications'.
:D
thanks alot.

Can both yr servers see each other. Do a ping from server
1 to server 2 and vice versa.

Also it may help if the subscriber is set up using the IP
Address rather than server name.


[quoted text, click to view]