onbarcode.barcode.winforms.dll free download Used for database server clusters in VB.NET

Drawing Data Matrix 2d barcode in VB.NET Used for database server clusters

7-52

Page 7-28

7

You are the administrator of a Network Load Balancing cluster consisting of six Web servers running in unicast mode, with a single network interface adapter in each server. You are using the Network Load Balancing Manager application on one of the cluster servers to try to shut down the NLB service on one of the other servers, so that you can upgrade its hardware. Why is the Manager not letting you do this

Because when cluster servers with one network interface adapter are running in unicast mode, the adapters on all the host servers are using the same cluster MAC address, making ordinary communications between the cluster servers impossible.

Which of the following Nlb.exe commands do you use to shut down NLB opera tions on a cluster server without interrupting transactions currently in progress a. Nlb drain b. Nlb params c. Nlb drainstop d. Nlb queryport

How long does it take a Network Load Balancing cluster to begin the convergence process after one of the servers in the cluster fails

Page 7-45

Server capacity planning is the process of estimating the resources needed by each node in a cluster during a worst-case scenario. When one or more nodes in a cluster malfunction, their applications fail over to other nodes, and those nodes must be capable of running them effi ciently for the applications to remain available. Depending on the failover policies you configure for the cluster, you might have to use servers that are capable of running several applications at once in the event of a node failure, and the servers must have sufficient memory, processing power, and disk space to do so.

Which shared storage hardware configuration is the only possible solution for a server cluster with more than two nodes

7-53

Which of the following failover policies provides the best compensation for mul tiple node failures a. Failover pairs b. Hot-standby servers c. N+I d. Failover ring

Page 7-46

Given the information provided in the Case Scenario Exercise, answer the following questions: 1.

Each of the six Web servers on the perimeter network has a single network interface adapter installed. What additional hardware is absolutely required to imple ment the NLB cluster on the Web servers

You want to be able to use Network Load Balancing Manager, running on one of the Web servers, to configure all the servers in the NLB cluster. Other than this, very little noncluster communication between the Web servers is required. Which of the following communication models should you use to make this possible with the greatest economy Explain your answer. a. Single network interface adapter in unicast mode b. Single network interface adapter in multicast mode c. Multiple network interface adapters in unicast mode d. Multiple network interface adapters in multicast mode

b. Running the cluster in multicast mode enables the servers to communicate with each other without having to install a second network interface adapter in each server.

Which of the following storage hardware configurations should you use for the 4-node database server cluster Explain your answer. a. Install a SCSI host adapter in each server and connect them all to a single SCSI bus b. Install a SCSI host adapter in each server and connect them all to a SCSI hub c.

d. The Fibre Channel switched fabric solution is the only one that Windows Server 2003 supports for a 4-node server cluster.

7-54

7

You have decided to partition your database server application to spread the load among the servers in the cluster. Which of the following failover policies will ensure that the entire database is constantly available without any server running multiple partitions, even if two servers fail Explain your answer. a.

Failover pairs. Split the database into two partitions and assign each one to an active server. Then, configure each of the active servers to fail over to one of the two remaining servers. b.

Hot-standby server. Split the database into three partitions and assign each one to an active server. Then, configure each of the active servers to fail over to the one remaining server. c.

N+I. Split the database into two partitions and assign each one to an active server. Then, configure each of the active servers to fail over to either one of the two remaining servers. d.

Failover ring. Split the database into four partitions and assign each one to an active server. Then, configure each of the four servers to failover to the next server.

c. To guarantee that no server ever has to run more than one partition, you must have two standby servers, so answers b and d are incorrect. Answer a is incorrect because the failure of an active server and its designated standby would cause one of the partitions to go offline.