HP-UX High Availability

1.Use the information provided in the Cluster Configuration Diagram to answer the following question. Which commands can be used to build a cluster configuration file under /etc/cmcluster called cluster.ascii. Select TWO.

A.cmquerycl -C /etc/cmcluster/cluster.ascii -n 15.20.30.6 -n 15.20.30.7
B.cmquerycl -C /etc/cmcluster/cluster.ascii -n 15.20.30.6 -n 15.20.30.7 -v
C.cmmakecl -C /etc/cmcluster/cluster.ascii -n shrek -n fiona
D.cmmakecl -C /etc/cmcluster/cluster.ascii -n shrek -n fiona -v
E.cmquerycl -C /etc/cmcluster/cluster.ascii -n shrek -n fiona
F.cmquerycl -C /etc/cmcluster/cluster.ascii -n shrek -n fiona -v
Correct:E F
2.Click the Cluster Configuration Diagram button to answer this item. Complete the following task using the information in the Cluster Configuration Diagram and the command outputs and files under the Exhibit button below. You may not need all exhibits. Fill in the indicated disk device files on the Cluster Configuration Diagram using c?t?d? format.

Correct:
3.Click the Task button to begin. Configure a Serviceguard package by dragging and dropping the appropriate options into the empty field locations. There is more than one possible answer. The package will be called 'missionspace' and will have an IP address of 10.20.30.10 on the 10.20.30.0 network. A failure of the 10.20.30.0 network will cause the package to fail over to the adoptive node. The package will run on FIONA as the primary node and fail back from SHREK manually. A service process called space_monitor will be used to monitor the health of the package. The monitor is started by running the script /etc/cmcluster/EPCOT/spaceman and passing it the argument A failure of this process 2 times will cause the package to fail to the alternate node without affecting the node it is running on. The package mounts the logical volume /dev/vg04/moonvol at the directory /opt/moon. The filesystem type should be the Veritas file system and mounted with a delaylog option. The application is started by the command: /opt/moon/missionspace/takeoff.sh and stopped by the command: /opt/moon/missionspace/touchdown.sh.

Correct:
Green choice14---->Yellow Choice1
Green choice3---->Yellow Choice2
Green choice4---->Yellow Choice3
Green choice7---->Yellow Choice5
Green choice8---->Yellow Choice4
Green choice1---->Yellow Choice6
Green choice2---->Yellow Choice7
Green choice18---->Yellow Choice8
Green choice18---->Yellow Choice9
Green choice12---->Yellow Choice10
Green choice2---->Yellow Choice11
Green choice11---->Yellow Choice12
4.Click the 'Configure a Serviceguard Task' button and use drag and drop to place the appropriate options into the excerpts from the file. Use the Exhibit button at the bottom to determine how to employ this script in the package files. The package has the following characteristics: The package will be called 'toystory' and will have an IP address of 15.31.54.64 on the 15.31.54.0 network. A failure of this network will NOT affect the package running on the node. The package will run on WOODY as the primary node and fail back from BUZZ manually. The package will mount the following veritas logical volumes with the delaylog option: /opt/vgtoys/lvo11 /opt/bin/toystory /opt/vgtoys/lvo12 /opt/bin/toystory/logs /opt/vg06/datavo1 /toydata A service process called 'andy_room_mon' will be used to monitor the health of the package. The service process will have no restarts. The starting, stopping, and monitoring is run by using the script /etc/cmcluster/TOYS/andy_room.

Correct:
Green choice3---->Yellow Choice1
Green choice5---->Yellow Choice2
Green choice13---->Yellow Choice4
Green choice13---->Yellow Choice5
Green choice7---->Yellow Choice3
Green choice9---->Yellow Choice6
Green choice6---->Yellow Choice7
5.Click the Task button. Using drag and drop, match each cluster type to its description.

Correct:
Green choice1---->Yellow Choice1
Green choice2---->Yellow Choice3
Green choice3---->Yellow Choice2
6.Click the Task button. Using drag and drop, match the correct solution with each single point of failure.

Correct:
Green choice5---->Yellow Choice4
Green choice4---->Yellow Choice1
Green choice3---->Yellow Choice3
Green choice1---->Yellow Choice2
7.In the context of high availability, which phrase describes mean time to failure?
A.the amount of time to indefinitely restart a failed service
B.the amount of time a system can provide service without failure
C.the amount of time to stop a service after a failure has been experienced
D.the amount of time to resume service after a failure has been experienced
Correct:B
8.In the context of high availability, what is the affect of split-brain syndrome?
A.different applications accessing the same disks simultaneously
B.two clusters initiate an application that accesses the same disks
C.different users of an application accessing the same disks consecutively
D.two sub-clusters initiated from a single cluster without knowledge of each other
Correct:D
9.While performing a rolling Serviceguard software upgrade, which statements are true? Select TWO.
A.You need at least two nodes running the same Serviceguard release.
B.You must change the cluster configuration as each node is updated.
C.You cannot change the cluster configuration files.
D.You may have up to three versions of Serviceguard running.
E.You cannot use the new features of the new Serviceguard release.
Correct:C E
10.Click the Exhibit button. Based on the cmviewcl output, how many times can the package service oracle9i be restarted for package ora_pkg?

A.0
B.1
C.2
D.3
E.unlimited
Correct:B
11.A server node in an Serviceguard cluster contains three LAN interface cards; two 100baseT Ethernet NICs and one 1000baseT gigabit Ethernet NIC, with the intent of providing a high speed data LAN with redundancy. What is a consequence of this LAN configuration?
A.The NICs for the data LAN and failover LAN must be identical.
B.The gigabit Ethernet and the 100baseT NICs cannot share the same subnet.
C.Three NICs are insufficient to provide redundancy for both the data and the heartbeat LAN.
D.The data packet size on the gigabit Ethernet is larger than the 1000baseT NIC and therefore will cause additional network latency.
Correct:A
12.What is the correct way to shutdown node_a when it is part of a two node cluster? Current snapshot of the cluster: node_a - pkg1 running node_b - pkg2 running
A.[node_a]#: cmhaltpkg pkg1 [node_b]#: cmhaltnode -f [node_b]#: cmrunpkg -n node_b pkg1 [node_a]#: shutdown -h 0
B.[node_a]#: cmhaltpkg pkg1 [node_b]#: cmrunpkg -n node_b pkg1 [node_a]#: cmhaltnode [node_a]#: shutdown -h 0
C.[node_b]#: cmhaltpkg pkg2 [node_a]#: cmhaltcl -f [node_b]#: cmrunpkg -n node_b pkg1 [node_a]#: shutdown -h 0
D.[node_b]#: cmhaltpkg pkg1 [node_a]#: cmhaltcl -f [node_ b]#: cmrunpkg -n node_b pkg1 [node_a]#: shutdown -h 0
Correct:B
13.Which commands move a package called pkg to the alternate nodeB?
A.cmhaltpkg; cmrunpkg pkg
B.cmhaltpkg; cmmvpkg -n nodeB pkg
C.cmhaltpkg; cmrunpkg pkg; cmmodpkg -e pkg
D.cmhaltpkg; cmrunpkg -n nodeB pkg; cmmodpkg -e pkg
Correct:D
14.Click the Task button. Using drag and drop, order the steps that Serviceguard performs during a package halt sequence. One step will not be used.

Correct:
Green choice5---->Yellow Choice1
Green choice4---->Yellow Choice2
Green choice6---->Yellow Choice3
Green choice2---->Yellow Choice4
Green choice1---->Yellow Choice5
15.Click the Task button. Using drag and drop, order the steps to add a package to a running two-node cluster. There is more than one correct answer.

Correct:
Green choice3---->Yellow Choice5
Green choice6---->Yellow Choice6
Green choice5---->Yellow Choice7
Green choice4---->Yellow Choice8
Green choice1---->Yellow Choice4
Green choice2---->Yellow Choice3
Green choice7---->Yellow Choice2
Green choice8---->Yellow Choice1
16.Which file would you edit to enable each node in the cluster to automatically start or join the cluster at boot time?
A.cmcld
B.cmlvmd
C.cmcluster
D.cmclconfig
Correct:C
17.What are the results of the cmapplyconf command?
A.creates the cmcld file and distributes it to all nodes
B.creates the cmclconfig file and distributes it to all nodes
C.creates the cmclconfig file and distributes the ascii file and the cmclconfig file to all nodes
D.creates the cmclcuster file and distributes the ascii file and the cmclconfig file to all nodes
Correct:B
18.What is the purpose of a lock disk in a cluster?
A.prevents activation of unclustered disks
B.keeps track of who is activating each volume group
C.prevents split-brain syndrome in the cluster
D.locks non-root users from running Serviceguard commands
Correct:C
19.What is the function of the lock disk in a Serviceguard cluster?
A.ensures that Quorum occurs for a cluster formation
B.determines which node becomes the cluster coordinator
C.keeps track of which volume groups are activated on each node
D.avoids the cluster splitting into two running clusters and corrupting data
Correct:D
20.Click the Cluster Diagram buttom above to view the cluster diagram graphic. Use the Exhibit button below to see four cluster ascii file excerpts. Which cluster ascii file excerpt is correct for the cluster diagram graphic shown?

A.Excerpt 1
B.Excerpt 2
C.Excerpt 3
D.Excerpt 4
Correct:A
21.Which file controls the default behavior of the boot process to activate all known volume groups at boot time?
A.lvmrc
B.lvmtab
C.inetd.conf
D.rc.config
Correct:A
22.Consider the following excerpt from the cluster ascii file: NODE_NAME Zappa NETWORK_INTERFACE lan0 HEARTBEAT_IP 1.2.3.5 NETWORK_INTERFACE lan2 STATIONARY_IP 6.7.3.4 NETWORK_INTERFACE lan6 HEARTBEAT_IP 6.7.3.17 NETWORK_INTERFACE lan4 Which statement is true?
A.lan4 is a standby interface; lan2 does not carry heartbeat packets.
B.lan4 is a standby interface for lan2 and lan6; lan6 carries only heartbeat packets.
C.lan4 is a standby interface for lan6; lan6 carries only heartbeat packets.
D.lan4 is a standby interface; lan6 carries heartbeat and all data packets.
Correct:A
23.Click the Exhibit button. On this two node cluster, what happens if the heartbeat LAN on system_a fails?

A.The package ora_pkg will failover to node system_b.
B.Node system_a will TOC and node system_b will reform a one node cluster.
C.On system_a the standby LAN card will be configured as heartbeat LAN card and the cluster status will not change.
D.Both systems will try to grab the lock_disk and whoever activates the lock_vg first will reform the cluster and the other node will TOC.
Correct:D
24.Click the Exhibit button. Based on the cmviewcl output, how many package services are configured for package ora_pkg?

A.2
B.3
C.4
D.5
Correct:A
25.Which parameter in the /etc/lvmrc file should be modified to ensure that LVM disks for Serviceguard data do not activate at boot time?
A.AUTOSTART_VG
B.AUTO_ACTIVATE
C.AUTOSTART_LVMD
D.AUTO_VG_ACTIVATE
Correct:D
26.In Serviceguard, the term "rolling upgrade" refers to which actions? Select TWO.
A.migrating to new disk storage
B.applying operating system patches
C.upgrading network switch equipment
D.introducing a new operating system revision
Correct:B D
27.In a two node cluster, two identical servers (same CPU, memory, and I/O card configurations) are designed to host the application and database servers of a 3-tier application. Under normal operations the database is hosted on server A and the application server is hosted on server B; each server with load at 90% of capacity. What is likely to occur in the event of a single server failure, after both applications have migrated to the surviving node?
A.Nothing; both application will perform as usual.
B.Both applications run in degraded mode.
C.The database application runs in degraded mode.
D.The application server runs in degraded mode.
Correct:B
28.Which high availability solutions are provided by Serviceguard? Select THREE.
A.application switching
B.software redundancy
C.hardware redundancy
D.power supply redundancy
E.bridge network redundancy
Correct:A B C
29.In a two node HA cluster, two identical servers (same CPU, memory, and I/O card configurations) are designed to host the database tier of a 3-tier application. They are connected to a large external disk array through two fibre channel links. Under normal operations the production database is hosted on server A and a batch reporting tool is hosted on server B. What are possible remedies for slow performance of the production database during non-failure operation? Select TWO.
A.Implement Pvlinks on both servers.
B.Activate available ICOD CPUs on server A.
C.Terminate batch reporting on server B to reduce disk array load.
D.Terminate WLM and Perfview middleware to free up additional CPU.
Correct:B C
30.In a two-node HA cluster, two servers are specified to host the application and database servers of a 3-tier application. Client systems connect through an external network. Which techniques are employed to provide a degree of high availability to the clients that connect to the application tier? Select TWO.
A.Implement load balancing on all three tiers.
B.Implement alias user accounts to provide alternate access.
C.Utilize redundant client IP networks for different groups.
D.Implement network address translations in a round robin algorithm.
E.Implement redundant routers between clients and the HA cluster network.
Correct:C E