Technical Resources > Tips & Tricks > Tips & Tricks Library
Administration Tip: Increasing
NNTP Reliability Using a
Master/Subordinate
Arrangement


Posted: October 02, 2001

Because Exchange 2000 clustering does not support Network News Transfer Protocol
(NNTP) components, consider using a master/subordinate (previously known as
master/slave) arrangement to increase NNTP reliability. A master/subordinate
arrangement involves keeping newsgroup information on more than one server. Keeping
newsgroup information on more than one server reduces the impact in situations where a
server becomes unavailable. One possible configuration would be to use one master
server and one or more subordinate servers. With such a configuration, you can configure
your newsgroup DNS in two ways:


Configure DNS so that users access newsgroups through a specific subordinate
server.
Configure DNS using round robin DNS so that users access newsgroups through
any available subordinate server. Configuring round robin DNS offers additional
reliability because users can continue to use all newsgroup functionality even if one of
the subordinate servers becomes unavailable.

Note: When you set up your NNTP master/subordinate arrangement, configure a feed
from each subordinate server to the master server. Do not configure any feeds between
subordinate servers.


For example, Contoso, Ltd., a fictitious company that previously used only one server to
store newsgroup information, decides to use a master/subordinate arrangement to
increase reliability. Contoso configures its master news server
(mynewsserver.contoso.com) to point to three subordinate news servers
(subordinate1.contoso.com, subordinate2.contoso.com, and subordinate3.contoso.com).
Contoso also configures round robin DNS so that users will gain access to newsgroups
through either subordinate1, subordinate2, or subordinate 3. With this configuration,
when a Contoso user connects to the newsgroups from his or her browser, he or she is
routed to one of the subordinate servers (for example, subordinate1). When the user posts
a message to subordinate1, the message replicates to the master server, and then the
master server replicates the message to subordinate2 and subordinate3. That new message
post is not available to any users (that is, the message cannot be seen by any of Contoso's
users), until it successfully replicates to the master server. After the replication occurs,each subordinate server contains the same content.


Master Server or Subordinate Server Failure


If the master server becomes unavailable, users can still view the messages within the
newsgroups. They can even create and submit new posts. However, any new posts
created on any subordinate server while the master server is unavailable will not be
available to any users. The new posts are not available because the master server has not
replicated them to the subordinate servers. Each subordinate server holds any pending
posts until the master server is available. After the master is available, it accepts any
pending posts, and then replicates those messages to the subordinate servers. After
replication occurs, each subordinate server contains the same up-to-date content. In this
scenario, the only temporary interruption of service to users is that newly posted
messages cannot be viewed.


If one of your subordinate servers becomes unavailable, and if you configured your
NNTP to use round robin DNS, your users have full newsgroup functionality because
they can connect to another available subordinate server. In this scenario, remove the IP
address of the unavailable subordinate server from your round robin DNS configuration.
After you remove the IP address of the unavailable subordinate server from DNS, users
who connect to the newsgroups will be pointed to one of the available subordinate
servers. If you do not remove the IP address of the unavailable subordinate server, users
will continue to be routed to the unavailable subordinate server. If any of your users were
connected to the subordinate server when it became unavailable, they must restart their
browsers to continue using the newsgroups because they may have cached information
specifying the now unavailable subordinate server. When the subordinate server becomes
available again, you must reconfigure the NNTP round robin DNS to include the IP
address of the subordinate server.


If one of your subordinate servers becomes unavailable and your NNTP does not use
round robin DNS, your users who gain access to newsgroups through that unavailable
subordinate server will lose all NNTP functionality. In this scenario, you must change
your configuration so that users who access newsgroup information on the unavailable
subordinate server can access newsgroups through one of your other subordinate servers.
After the subordinate server is once again available, reconfigure your NNTP access to the
subordinate server so that users can access newsgroups through their original subordinate
server.


Master Server and Subordinate Server Newsgroup Hierarchy
Considerations


When you use a master/subordinate arrangement, keep in mind the following
considerations:


When creating newsgroup hierarchies, only newsgroups that are created using
control messages will be automatically created on the subordinate servers. For more information about using control messages, see the Exchange online documentation.
Newsgroups created by using System Manager or by using the Rgroup.vbs
command at a command prompt do not replicate to other subordinate servers until a
message is posted to the group on the master server. For more information about
using Rgroup.vbs, see the Exchange online documentation.
Users cannot post messages on a subordinate server until the newsgroup to which
they want to post has been replicated to that subordinate server.


After you establish your newsgroup hierarchy, you should replicate it to the subordinate
servers by doing one of the following:

Post a test message in the group on the master server,
Create an identical hierarchy on each subordinate server, or
Create your newsgroups using control messages.

For more information about NNTP master/subordinate arrangements, see the
Exchange 2000 online documentation.

This tip is provided by the Microsoft Exchange User Education team.
For additional information, see the Microsoft TechNet Exchange site.

 
 

EXCHANGE DATA RECOVERY SUPPORT AVAILABLE 24/7!!!!! CALL NOW WITH YOUR MICROSOFT EXCHANGE SERVER QUESTIONS!!!

EXCHANGE RECOVERY SUPPORT CONTACT NUMBER: 800-364-9838

 

Compro Technologies, Inc.
Exchange Server Recovery through a Rebuild of the Database by using only the Log Files in Exchange 2000

Border Environment Cooperation
Exchange Recovery of a 14 GB Priv.edb File on LTO Tapes Microsoft Exchange 2000 Exchange Server Tape Recovery


Click Here For Our Client List

 

"I am taking the time out of a very busy schedule to make the Microsoft
Corporation aware of a recent Exchange 2000 Server outage in my
non-profit organization, and of the extraordinary support we received
from Innovative Technology Concepts.......
Click Here to read the rest

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
Copyright (c) 2003 - 2007 Innovative Technology Concepts, Inc. All rights reserved. www.InnovativeTechnologyConcepts.Com