Questions and Answers


Lesson 1 Review

Page 6-16

1.

You are a senior network administrator for a fast-growing company that wants to implement a clustering technology in order to ensure the continued availability of a Web application if one of the Web servers fails and, more importantly, to handle the increased traffic that is expected in the future. You are asked whether you should deploy Network Load Balancing or the Cluster Service for this application, and why you would choose one over the other.

you should choose to deploy network load balancing in this situation. the cluster service is designed to provide high availability of application resources, but it does not load balance in any way. this makes it less suitable than network load balancing for scaling applications to handle increased traffic. while network load balancing cannot monitor at the service level like the cluster service, the primary goal is to handle the increased traffic.

2.

You are setting up an Exchange Server 2003 virtual server on the Cluster Service, using a two-node active/active configuration. As you read through a planning checklist, you identify how the services will failover. Which Exchange Server 2003 service cannot be set to an active/active configuration like the other services?

  1. Microsoft Exchange Information Store

  2. Microsoft Exchange System Attendant

  3. Microsoft Exchange Message Transfer Agent

  4. Microsoft Exchange Routing Service

the correct answer is c.

3.

You are the administrator of an Exchange Server 2003 cluster that consists of Server01 and Server02 in a two-node active/passive configuration. The servers have the following TCP/IP configuration:

Server01 (active)

  • Local Area Connection 1 (Public) — IP Address — 10.1.1.250, DHCP

  • Local Area Connection 2 (Private) — IP Address — 192.168.254.1, static

Server02 (passive)

  • Local Area Connection 1 (Public) — IP Address — 10.1.1.251, DHCP

  • Local Area Connection 2 (Private) — IP Address — 192.168.254.2, static

Over a holiday weekend, a Windows Server 2003 domain controller fails due to a faulty hard drive. On Tuesday morning, you order a hard drive for the server, and it is scheduled for delivery on Wednesday. Tuesday afternoon, users begin to report that e-mail is down. You check the cluster, and it appears at first glance to be functioning properly internally. You manually failover the cluster to the second node, but the problem is not corrected. Where should you look next to troubleshoot the problem?

the next place to check is actually the first place you should check, which is basic network connectivity. all too often, network administrators make a high-level change such as failing over a cluster node without first determining where the problem really lies. in this instance, the domain controller that failed was likely the dhcp server, and the cluster was unable to renew its ip addresses and therefore released them. while clusters can use dynamic ip addressing on the public network, it is recommended to use static addressing to avoid this type of problem. to fix the problem, you can either assign static ip addresses to the network adapters connected to the public network or bring the dhcp server back online. since you're waiting on a hard drive, you should configure static ip addressing and then register those addresses in dns. then clients will be able to connect again.

Answers

1.

You should choose to deploy Network Load Balancing in this situation. The Cluster Service is designed to provide high availability of application resources, but it does not load balance in any way. This makes it less suitable than Network Load Balancing for scaling applications to handle increased traffic. While Network Load Balancing cannot monitor at the service level like the Cluster Service, the primary goal is to handle the increased traffic.

2.

The correct answer is c.

3.

The next place to check is actually the first place you should check, which is basic network connectivity. All too often, network administrators make a high-level change such as failing over a cluster node without first determining where the problem really lies. In this instance, the domain controller that failed was likely the DHCP server, and the cluster was unable to renew its IP addresses and therefore released them. While clusters can use dynamic IP addressing on the public network, it is recommended to use static addressing to avoid this type of problem. To fix the problem, you can either assign static IP addresses to the network adapters connected to the public network or bring the DHCP server back online. Since you're waiting on a hard drive, you should configure static IP addressing and then register those addresses in DNS. Then clients will be able to connect again.

Lesson 2 Review

Page 6-27

1.

You are the administrator of a two-node active/passive Exchange Server 2003 cluster that is currently not in production but is being tested prior to being put into production. You are testing your failover policies, so you initiate failures on several resources until the group failure threshold is reached. The server successfully fails over the group to the formerly passive node, and the previously active node becomes passive. You want the group to fail back to the original service, but after an hour of waiting, the group is still running on the second node. It is late in the day so you decide to go home and deal with it the next day, since the cluster is not yet in production. The next morning, you notice that the cluster is still running on the second node, and you can't find a command in the Cluster Administrator GUI to initiate a failback. What can you do to get the group back on the original server node?

failback is controlled through a policy that is defined at the group level. you need to right-click the group, then click properties, and then click the failback tab. the default setting is to not fail back, but you can configure the group to fail back immediately when the original node returns to service or to fail back only during certain hours. in this situation, you could set the policy to immediately, which would cause the group to fail back after you click ok.

2.

You are setting up a test server to practice with the cluster service as a prelude to a later cluster deployment. You don't have any real servers available, but you know that you can use a single system and a local quorum disk resource for testing purposes. You set up a desktop workstation class machine that has a SCSI C drive and an IDE D drive and 512 MB of RAM with Windows Server 2003, Enterprise Edition, and then you install the clustering service and Exchange Server 2003. You create the IP address, the network name, and the physical disk resources. However, when you attempt to bring the group online, the process fails. Why?

when you create a physical disk resource, it will allow you to select an available ide hard drive; however, you won't be able to bring the resource online. only scsi disks are supported as physical disk resources. in this case, the scsi c drive is already in use as the local quorum disk, so it cannot be used as the physical disk resource for the exchange server 2003 virtual server. you must add another scsi drive in order to proceed.

3.

You are the administrator of an Exchange Server 2003 virtual server on a two-node active/active cluster running Windows Server 2003, Enterprise Edition, on each node. You know that Windows Server 2003, Enterprise Edition, supports eight-node clustering, so you decide to add two additional active nodes to your Exchange Server cluster. You install Windows Server 2003, Enterprise Edition, on two additional servers and successfully add them to the cluster. You then install Exchange Server 2003 successfully on the new servers. However, when you attempt to create the virtual servers, you find that you are unable to, even though you have the correct permissions to do so. What is the problem, and how can you solve it?

exchange server 2003 supports only active/active configurations on two-node clusters. in this situation, you cannot add two additional active nodes without changing the configuration to an active/passive model with at least one passive node. you could configure three active nodes and one passive node, or add a fifth server and have four active nodes and one passive node. exchange server 2003, enterprise edition, supports the full eight-node clustering of windows server 2003, enterprise edition, but only with at least one passive node.

Answers

1.

Failback is controlled through a policy that is defined at the group level. You need to right-click the group, then click Properties, and then click the Failback tab. The default setting is to not fail back, but you can configure the group to fail back immediately when the original node returns to service or to fail back only during certain hours. In this situation, you could set the policy to Immediately, which would cause the group to fail back after you click OK.

2.

When you create a physical disk resource, it will allow you to select an available IDE hard drive; however, you won't be able to bring the resource online. Only SCSI disks are supported as physical disk resources. In this case, the SCSI C drive is already in use as the local quorum disk, so it cannot be used as the physical disk resource for the Exchange Server 2003 virtual server. You must add another SCSI drive in order to proceed.

3.

Exchange Server 2003 supports only active/active configurations on two-node clusters. In this situation, you cannot add two additional active nodes without changing the configuration to an active/passive model with at least one passive node. You could configure three active nodes and one passive node, or add a fifth server and have four active nodes and one passive node. Exchange Server 2003, Enterprise Edition, supports the full eight-node clustering of Windows Server 2003, Enterprise Edition, but only with at least one passive node.

Lesson 3 Review

Page 6-33

1.

You are the Exchange Server administrator for TailspinToys, which recently acquired WingtipToys. TailspinToys uses a front-end server outside of its firewall that connects to a back-end server cluster. You have migrated the mailboxes for WingtipToys employees onto the Exchange Server 2003 cluster. In order to support Internet-based client computers connecting to WingtipToys resources, you configure a new HTTP virtual server and virtual directory on the front-end server for wingtiptoys.com and mimic it on the Exchange Server 2003 virtual server. However, mail traffic fails to flow properly, even though it works fine for TailspinToys. You verify the network settings, and they are all correct. What else might you be missing?

since you are running the back-end server as an exchange server 2003 virtual server on a cluster, you must create an http virtual server resource in the exchange server cluster group in addition to creating the virtual servers and virtual directories. once you set this up, mail traffic should flow properly. if you were not using a cluster, this additional step would not be required.

2.

You install Exchange Server 2003 clean onto a new server, with the plan to make it a front-end server in an existing Exchange Server 2003 organization. After a successful installation, you start Exchange System Manager and expand the Servers container, then right-click the server and then click Properties. You select the check box for This Is A Front-End Server and click OK, but you receive an error. What did you do wrong?

prior to configuring a server as a front-end server, you must remove any services from it that are not supported on front-end servers. by default, exchange server 2003 configures servers with a recipient update service and a default offline address list. neither are supported on front-end servers and must be removed before exchange system manager will allow you to designate a server as a front-end server.

3.

You are the Exchange Server administrator for an organization that has an existing Exchange 2000 Server front-end and back-end server architecture. There are four non-load-balanced front-end servers servicing 15 non-clustered back-end servers. You want to upgrade to Exchange Server 2003 on all servers. In what order would you upgrade?

  1. Upgrade a front-end server and then each of its corresponding back-end servers, and then repeat the process with the next front-end server.

  2. Upgrade all the back-end servers, and then upgrade all the front-end servers.

  3. Upgrade back-end servers, then upgrade the corresponding front-end server, and then repeat the process with the next set of back-end servers.

  4. Upgrade all the front-end servers, and then upgrade all the back-end servers.

the correct answer is d.

Answers

1.

Since you are running the back-end server as an Exchange Server 2003 virtual server on a cluster, you must create an HTTP virtual server resource in the Exchange Server cluster group in addition to creating the virtual servers and virtual directories. Once you set this up, mail traffic should flow properly. If you were not using a cluster, this additional step would not be required.

2.

Prior to configuring a server as a front-end server, you must remove any services from it that are not supported on front-end servers. By default, Exchange Server 2003 configures servers with a Recipient Update Service and a Default Offline Address List. Neither are supported on front-end servers and must be removed before Exchange System Manager will allow you to designate a server as a front-end server.

3.

The correct answer is d.

Case Scenario Exercise: Requirement 1

Page 6-35

1.

You determine that you will need to use the Cluster Service rather than Network Load Balancing to ensure the availability of the Exchange Server services. You plan to start with a two-node cluster, given the amount of e-mail traffic you expect the cluster to handle. Should you use an active/active model or an active/passive model? Why?

microsoft recommends an active/passive configuration for most applications, which is appropriate for this situation. the e-mail traffic is less of a concern than the web traffic with respect to quantity, and it can easily be handled by a higher-end server. by configuring a cluster with an active node and a passive node, you can ensure that if the active node fails, the passive node will come online to service client requests. in addition, the active/passive model is scalable up to eight nodes, so you can add additional active nodes later, if necessary.

2.

Describe the process that you will use to set up this cluster using Windows Server 2003 and Exchange Server 2003, Enterprise Edition.

first you will install the windows server 2003 operating system on the server that will become the first node in the cluster. after the operating system is configured and a service account for the cluster is created, you start cluster administrator and choose to create a new cluster using the new server cluster wizard. then, you will install exchange server 2003 on the server, joining the existing organization. after the installation, you will delegate the cluster service account the exchange full administrator role in exchange system manager. then you will use cluster administrator to create a cluster group for the exchange server 2003 virtual server, and then create cluster resources for ip address, network name, physical disk, and exchange system attendant. the exchange system attendant will install the other exchange server resources, which you will bring online when the cluster is complete.

Answers

1.

Microsoft recommends an active/passive configuration for most applications, which is appropriate for this situation. The e-mail traffic is less of a concern than the Web traffic with respect to quantity, and it can easily be handled by a higher-end server. By configuring a cluster with an active node and a passive node, you can ensure that if the active node fails, the passive node will come online to service client requests. In addition, the active/passive model is scalable up to eight nodes, so you can add additional active nodes later, if necessary.

2.

First you will install the Windows Server 2003 operating system on the server that will become the first node in the cluster. After the operating system is configured and a service account for the cluster is created, you start Cluster Administrator and choose to create a new cluster using the New Server Cluster Wizard. Then, you will install Exchange Server 2003 on the server, joining the existing organization. After the installation, you will delegate the cluster service account the Exchange Full Administrator role in Exchange System Manager. Then you will use Cluster Administrator to create a cluster group for the Exchange Server 2003 virtual server, and then create cluster resources for IP address, network name, physical disk, and Exchange System Attendant. The Exchange System Attendant will install the other Exchange Server resources, which you will bring online when the cluster is complete.

Case Scenario Exercise: Requirement 2

Page 6-35

1.

For the back-end server, you used the Cluster Service. Should you also use it on the front-end server? Why or why not?

the cluster service is inappropriate for the front-end server cluster. the cluster service does not load balance, which makes it less suited for use on front-end servers. you do want clustering technology in place but would be better off using network load balancing for this part of the configuration. network load balancing can load balance client connections and it can be used to provide a unified namespace to internet-based clients. these users and web applications can use a single uniform resource locator (url) for each e-mail domain, regardless of which back-end server hosts the mailbox or public folder that is being accessed.

2.

How will the server nodes in the cluster using Network Load Balancing handle one of the servers going offline?

network load balancing cannot monitor at the application service level, but if a server goes offline, the other servers in the cluster will perform convergence to assume the load previously handled by the offline server and mark it as offline.

3.

After configuring the Exchange Server 2003 two-node active/passive cluster and setting a two-node network load balanced cluster as front-end servers in Exchange System Manager, what additional step must you take so that the tailspintoys.com domain can also receive e-mail on the cluster through the front-end servers for wingtiptoys.com?

in order to support multiple domains, you must configure a second http virtual server for the wingtiptoys.com domain. you must create this virtual server on both front-end servers and on the active node of the back-end cluster (since only the active node has exchange server 2003 installed on it), and then create an http virtual server cluster resource for the wingtiptoys.com domain in the exchange server cluster group in cluster administrator. default http virtual servers already exist in these three places, so the primary domain tailspintoys.com is already configured.

Answers

1.

The Cluster Service is inappropriate for the front-end server cluster. The Cluster Service does not load balance, which makes it less suited for use on front-end servers. You do want clustering technology in place but would be better off using Network Load Balancing for this part of the configuration. Network Load Balancing can load balance client connections and it can be used to provide a unified namespace to Internet-based clients. These users and Web applications can use a single Uniform Resource Locator (URL) for each e-mail domain, regardless of which back-end server hosts the mailbox or public folder that is being accessed.

2.

Network Load Balancing cannot monitor at the application service level, but if a server goes offline, the other servers in the cluster will perform convergence to assume the load previously handled by the offline server and mark it as offline.

3.

In order to support multiple domains, you must configure a second HTTP virtual server for the wingtiptoys.com domain. You must create this virtual server on both front-end servers and on the active node of the back-end cluster (since only the active node has Exchange Server 2003 installed on it), and then create an HTTP virtual server cluster resource for the wingtiptoys.com domain in the Exchange Server cluster group in Cluster Administrator. Default HTTP virtual servers already exist in these three places, so the primary domain tailspintoys.com is already configured.




MCSA/MCSE Self-Paced Training Kit (Exam 70-284(c) Implementing and Managing Microsoft Exchange Server 2003)
MCSA/MCSE Self-Paced Training Kit (Exam 70-284): Implementing and Managing MicrosoftВ® Exchange Server 2003 (Pro-Certification)
ISBN: 0735618992
EAN: 2147483647
Year: 2003
Pages: 221

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net