On this page I will explain how to install and configure CCR for
Exchange 2007 Server. Incidentally, I keep wanting to call it continuous
cluster replication instead of the official Cluster Continuous Replication.
Microsoft's CCR concept involves combining the failover features provided by the Cluster
Service, with the asynchronous log-shipping technology built-in to Exchange 2007
Server. Remember to create your cluster BEFORE installing your Exchange
server.
The mission for CCR is to provide high availability for Exchange 2007
Mailbox servers.
You may have experienced a catastrophic event where a
failover to a passive node would have saved your day, for example, power failure, CPU
overheats, disk crashes, memory chip problem, or the server's motherboard is damaged.
With a clustered configuration, failover time could be as little as 3 minutes;
here are the reasons why Exchange CCR is so effective:
Eliminates single points of failure.
No shared storage requirements.
Better performance, backup from the passive node. Also reduces backup frequency and volume.
Drastically reduces the recovery time from any failure of the active node.
Exchange CCR uses the database recovery ability of Exchange 2007 server.
The cluster technology works
by continuously updating a passive copy of the mailbox database. The process is
asynchronous in the sense that the logs are not copied to the passive node until they are closed by
the active server. However, it is continuous in the sense that after an
active node finishes (closes) a log, that log is immediately copied across to
the passive node.
When you install the second (passive node) each storage group, together with its database, is
copied from the active node. This process is called 'seeding' and it
establishes a baseline in the replica. Thereafter, the logs are copied continuously from
the active to the passive node.
Once you have built the cluster, then failover only occurs when the passive
server misses 10 heartbeats. This corresponds to about 12 second of broken
connection. To anticipated the most frequently asked question, yes, you can configure the number of heartbeats
before failover occurs.
CCR Core Elements:
Microsoft's failover cluster service provides failover from Node-A to
Note-B. As the CCR technology relies on the passive node replaying the
logs, there is little performance impact on the active node.
CCR uses a cluster quorum model with a file share as the 'witness' which
controls failover, another benefit of this technology is that it prevents
'split-brain' errors.
When the logs arrive at the passive node, Exchange 2007 server replays
each new log into the mailbox database, thus when failover occurs, it will be quick.
The transport dumpster is the other key component of Exchange CCR. This dumpster
is situated away from the main action on the Hub Transport server. The
idea is to keep a copy of each message until Exchange can be very that it's
stored in the passive node. It works because the Hub Transport servers
keeps a queue of recently delivered mail.
When a failover occurs, CCR on the clustered mailbox server automatically
requests every Hub Transport server in the Active Directory site to resubmit
mail from its transport dumpster queue. The information store then
deletes the duplicates mail and delivers the remainder.
Guy Recommends: A Free Trial of the Network Performance Monitor
(NPM)
SolarWinds'
Network Performance Monitor
will help you discover what's happening on your network. This
utility will also guide you through troubleshooting; the dashboard will
indicate whether the root cause is a broken link, faulty equipment or
resource overload.
Perhaps the NPM's best feature is the way it suggests solutions to network
problems. Its
second best feature is the ability to monitor the health of individual VMware
virtual machines. If you are interested in troubleshooting, and creating network maps, then I recommend that you
give this Network Performance Monitor a try.
A reminder, create your cluster first on the Windows 2003 server. Only then actually install
the physical Exchange 2007 Server. CCR can only handle one database,
thus each mailbox storage group must have only one database. Another
pre-requisite is that you need the Enterprise Edition of the operation
system, Windows Server 2003.
Network Cards
All cluster technology requires two networks. One public for the
clients, one private for the heartbeat communication.
Because you need to
configure the public and private networks, each node needs a second physical
network card. To configure the NIC open the Network Connections
(Control Panel).
In addition to configuring the static IP address, and the DNS server IP, you
could right-click, then rename each connections. For example, change
'Local Area Connection...', to the more meaningful 'Public' or 'Private'.
One more tip, view the Advanced menu, Advanced settings, and check the
binding order, what you want is the Public network at the top.
Call for the Cluster wizard
Logon as an administrator. Create a special user for the Cluster
Service, make it a member of the Local Administrators on the Exchange
2007 Server.
Then go to the command line and type:
Cluster /create /wizard
The Cluster Administrator should launch.
Supply the wizard with the necessary information. For instance
the Active
Directory domain name, and a short memorable name for the new cluster.
The New Server Cluster wizard appears. Verify that you have the
necessary information to continue with the configuration, and then click
Next.
On the 'Select Computer' page, accept, or type the name of the
computer for the primary node.
Peruse the Analyzing Configuration page, check there are no problems.
Next is the IP Address page; type the cluster IP address, and then click
Next. The wizard automatically associates the cluster IP address with
the public network. (It susses the subnet mask to select the
correct network.)
Onward to the Cluster Service Account page, type the user name and
password for the Cluster service account. As usual, the wizard
verifies the user account and password.
Nearly finished, just the Proposed Cluster Configuration page, click
Quorum. Select Majority Node Set.
Remember to repeat the procedure on the second node.
Guy Recommends:
The SolarWinds Exchange Monitor
Here is a
free tool to monitor your Exchange Server. Download and
install the utility, then inspect your mail queues, monitor the Exchange
server's memory, confirm there is enough disk space and check the CPU
utilization.
This is the real deal - there is no catch. SolarWinds provides this
fully-functioning freebie, as part of their commitment to supporting
the network management community.
Microsoft always say that every Windows Server 2008 feature will be easier than before; in the case
of setting up clustering this statement is true, largely due to the Failover
Cluster Tool. What you also need is a good grasp of the concepts
starting with the two networks: Public (normal), Private (heartbeat).
You also need to know the difference between the active and passive Mailbox
nodes. It won't take long to understand more unusual concepts such as
'File Share Majority quorum*'. Lastly, appreciate the role of the
'Transport Dumpster' which is important should a failover actually occur.
*The quorum keeps the definitive cluster configuration. This quorum
is a resource which can only have one owner, the active node. The File
Share Majority quorum is designed to avoid the 'split-brain' nightmare
scenario where each server thinks the other server has failed and it is in
charge of the active Mailbox node.
Incidentally, there are command line tools to configure Exchange CCR, however, I
prefer the Failover Cluster GUI because it provides a better view of all the
components and options. Remember that to install CCR on Windows Server
2008 it must be the Enterprise edition of the operating system.
Plan carefully. Create your cluster FIRST, only then build
your Exchange 2007 Server.
Setup the cluster networks. Plan, then configure the static IP
addresses for Public and Private networks. (See instructions for Network
Connections in the above Windows Server 2003 section.)
Let us verify that conditions are suitable for installing a cluster.
Launch the Failover Cluster Management tool, select 'Validate a
configuration'. Check: Inventory, Network, and System
Configuration. Guy says ignore Storage as it means shared storage
which we are not using here.
Use the Failover Cluster Management tool again, but this time
select: Create a cluster (before installing Exchange).
Form the cluster, beginning with the first node and then the second.
Configure the cluster networks for missed cluster heartbeats.
Create a shared folder, secure it with the correct permissions for
the file share witness.
Install the active Mailbox Server roles into the cluster. The wizard
will install the CMS (Clustered Mailbox Server) for you. Finally,
install the passive Mailbox Server into your cluster.
Tuning failover control settings.
Check the configuration of the transport dumpster.
Try a test move of the CMS between the nodes in the cluster.
Once you have created your cluster, then it's time to install the
physical Exchange 2007 servers.
In conclusion, once you get started with the Failover
Cluster Management tool, the wizards will guide you through the Exchange
2007 CCR setup
and configuration.
»
Hidden Benefits of Cluster Continuous Replication
CCR improves the performance of the active Exchange 2007 server because
you can backup from the passive node. The continuous cluster
replication removes the need for a SAN (storage area network). Another
hidden benefit of CCR is that there is no need for specialist cluster
hardware. In Guy's experience, the new Failover Cluster Management
tool lives up to the hype that CCR, especially with Windows Server 2008,
really is easier to configure than previous Microsoft cluster systems.
CCR and PowerShell
As with all Exchange Server 2007 configuration, there are PowerShell cmdlets to verify the health and status of Exchange CCR:
For example: get-ClusteredMailboxServerStatus -Identity:Node-A
If you need to repair a clustered server, there is also:
move-ClusteredMailboxServer
For example: move-ClusteredMailboxServer -Identity:Node-A -TargetMachine:Node-B
-Identity is the source -TargetMachine as the name suggests is the
destination.
In Exchange 2007 LLR you can recover an Exchange 2007 database even
though one of the transaction log files is missing. LLR enables a
mailbox database to mount even when recently generated log files are
unavailable.
While LLR is enabled by default on all Exchange 2007 mailbox servers, you
can configure the maximum number of logs that can be lost before the
database is prevented from mounting. The key is to set the
AutoDatabaseMountDial parameter. This parameter, has three values: Best
Availability 6 logs lost (Default), Good Availability, with 3 logs lost,
and, Lossless with 0 logs lost.
Guy Recommends: SolarWinds' Free Bulk Mailbox Import Tool
Import users from a spreadsheet, complete with their mailbox. Just provide a list of the
users with the fields in the
top row, and save as .csv file. Then launch this FREE utility, match
your Exchange fields with AD's
attributes, click and import the users. Optionally, you can
provide the name of the OU where the new mailboxes will be born.
There are also two bonus tools in the free download, and all 3 have been approved by Microsoft:
Bulk-import new users and mailboxes into Active Directory.
Cluster Continuous Replication gives administrator's peace of mind that their
mailbox servers are protected from a single point of failure. To master
this technology, I advise investigating the components, and playing with the
Cluster Tools such as the Failover Cluster Management tool on Windows Server
2008. Then develop a CCR plan to suit your network, only then are you
ready to install the Exchange Server 2007 mailbox servers.
Credit and acknowledgement Alain Laventure provided the screenshots, the detailed steps and the
background for this CCR article.
If you like this page then please share it with your friends
Here is a
free tool to monitor your Exchange Server.
Download the utility, then inspect your mail queues, monitor Exchange server's
memory, confirm there is enough disk space and check the CPU utilization.
