The
thinking behind the Exchange Data centre infrastructures has evolved over the
years. The focus has lead to a consolidation approach to Mail server’s
technology and design plans. This
document is intended for the Network
Administrators that will be managing all of the functionalities for the
Exchange 2003 servers. The document is an overview of the main features for the
management of the Exchange 2003 servers in Native
Mode implementation. The main idea is for business to improvement TCO over the short and long term.
The document is for the fine-tuning of the Exchange
2003 servers in Native Mode. The idea is to deliver a high performance critical application infrastructure for
all of the business units. Making sure that application downtime is done
through change control which will
help to eliminate unplanned outages. Progressively, no changes on the Exchange
2003 servers should be implemented or actualized without being technically
fully documented. Downtimes should
not affect business continuity, application delivery, data integrity, and the
architectural availability of the service. For the management of Exchange there
are two modes preventive or reactive. This document will focus upon helping the
Exchange Network Administrator to lean how to coordinate preventive measures in
order to be prepared for reactive manifestations. This is a level 300 or 400
documentation for the Exchange Administrator because it requires the
manipulation of the Registry of the Exchange 2003 servers.
It is
understood that the new Exchange 2003
servers are inter-connected to the functioning of the infrastructure of the
Active Directory with a Kerberos Trust hierarchy. Active
Directory is a single depository of information about the enterprise. But, the
Exchange 2003 servers can also retrieve information from the Mailbox Store, the
Microsoft Windows registry, and the Exadmin virtual directory. Of course you
can manage Exchange through Active Directory Users and Computers (ADUC) or
Exchange system manager (ESM) but it may also be managed through Remote
Desktop, Terminal Server, or a dedicated management station.
- Enhanced security with Windows Server 2003 and Active Directory
- Improved and simplified administration with enhanced server tools
- Enhanced control over the desktop environment through the use of Group Policies
- Enhanced auditing and monitoring of clients and server environments
- Improved and more cost-effective storage using storage groups and multiple databases
- Improved anti-virus and anti-spam functionality
- Enhanced recovery options with features such as Recovery Storage Groups
- Routing groups can consist of servers from multiple administrative groups
- Servers can be moved between routing
groups
- Mailboxes can be moved across administrative
groups
- SMTP is the Default routing protocol
- Routing Bridgehead server
pairs use can use 8BITMIME data
transfers to provide significant bandwidth improvement over routing group
connectors
- The
IS service in Exchange Server 2003 automatically ignores and removes "zombie" Access Control Entries (ACEs) from
the Exchange Server 5.5 computers that were previously used in an
organization
- You can only create InetOrgPerson
objects if you are running a Windows Server 2003 domain controller object
class can be mailbox-enabled or mail-enabled
- This Active
Directory objects can be mail-enabled or mailbox-enabled only in a
native Exchange 2003 topology
- When you right-click this object it does
not always have an Exchange tasks option
- Query-based
distribution groups: Exchange queries Active Directory for all
recipients that match the filter to match the condition specified. A query-based distribution group allows you to use an LDAP query to dynamically build
membership in the distribution group. When the user belongs to a
query-based DG group the categorizer sends an LDAP request to a global
catalog server. The server will return matched addresses for the
categorizer to generate a recipient list for the routing of the message.
- Eliminate the need for Active Directory connector agreements.
- The possible use of universal security groups can
contain user accounts, global groups, or universal groups from any domain
in the forest – using nested ACLs groups
- What must not be forgotten is
the Exchange 2003 collaborative features that
will help you to share information quickly and efficiently
- For the Active Directory Windows Server 2003 supports renaming
and repositioning Domains after you configure the forest at the Windows
Server 2003 forest functional level. It is possible to rename the DCs without demoting it.
- Use of new Volume Shadow Copy Service (VSS) along with the new VSS API
calls for creating backups for
normal or copy backups only.
- Configure and apply WMI filters with GOP settings
- IMF intelligent message filtering of
emails
Each
Exchange 2003 server has by Default one storage group, named First Storage Group,
that contains one mailbox store and one public folder store. Every store within
the storage group shares a single transaction log. Exchange Server databases
can grow over time and become fragmented. Therefore the recovery plan must
specify the company's restore time requirements. Exchange Server 2003 Standard
Edition has one storage group that
contains one mailbox store and one public folder store. Exchange Server
2003 Enterprise Edition has up to four storage groups, each of which contains
as many as five databases. Exchange 2003 mailbox stores and public folder
stores are also called databases and are organized into storage groups.
All of the databases in a
storage group share a single set of transaction log files, a single backup
schedule, and a single set of logging and backup-related settings. Remember
that the default First Storage Group is used for the
transaction log configuration allowing Administrators to recover data if the
stores are damaged. Therefore a storage group includes in between 1 to 5
Databases. On any particular server you want to have more than five Databases.
The Database tab for a mailbox store includes several
different backup options most specifically for scheduled maintenance. Remember
that a Mailbox will inherit the properties
of the Mailbox stores. Exchange will allow only for one public folder tree
but each server has a public folder store. Verify public folder store integrity
likewise you can do online defragmentations or off life defragmentations of the
database.
The Transaction Log (EDB) logs database updates first before
they are applied to the database in background operations. It is important to
Monitor the size, traffic, I/O operations across disks, of the
IS Private and IS Public databases of each Exchange Server. Generally, a
knowledge script will be issued if there is a correction to be taken regarding
the overflow or corruption.
The IS database holds
all messages and documents.
IS Private — Stores all private mailboxes.
IS Public — Stores all public folders.
By default, the failure of one group resource may affect the whole group.
So - if a resource fails a specific number of times (default is 4 times in 15 minutes), the whole group will be moved to
the other node. When thinking about the expansion of the clusters within the
Exchange 2003 infrastructure. It is to be remembered that the Front-End servers are stateless and should be use only for Load Balancing (NLB) purposes. While
for Back-end servers you can have a clustering
environment that span up to eight manageable mail stores. In this type of
infrastructure each server becomes a node and each node is connected to a shared storage (SAN, fibre channel and
scsi). All of the nodes are lined to the shared storage resource but only one
server can talk at any one time to the storage system. The connected node is
considered as being the Active node while the other node is passive. It is also
a failed-over practise. The
service requires similar machines to share a common data storage device eg a
SAN or external drive array. Each of the servers will run a local set of drives
for the OS and uses the external device for the database stores and transaction
logs. If 1 server should fail, the other server can fail over and load the
databases and continue to services clients. The switch over from one node to
another can take a few seconds to several minutes (15/4). It is important to
understand that you also have two replications topologies asynchronous and
synchronous. It all depends upon the routing cost to render the data
available.
DATABASE
MAINTENANCE TASKS:
ESEutil: It can be used to verify the integrity of the
Exchange database stores to perform hard and soft recoveries of the store, and
copy repair and defrag the database stores.
ISinteg: Will be used to browse the Exchange store
tables and indexes for inconsistencies
