74-322 - Microsoft Lync Server 2010-Advanced Design and Deployment
Go back to Microsoft
The customer has implemented an SNMP based network monitoring tool and wishes to monitor changes in the cluster. The general status of the cluster is gathered via SNMP but the polling time to retrieve the status is too slow for immediate reaction to serious issues. There is a technical reason why the polling time cannot be changed. How can HACMP report changes in its status on a more timely basis?
Edit SNMP configuration file to send traps for cluster state changes.
During a cluster functionality test, an administrator performs a graceful stop for Node 'B' in a three-node cluster. What is the expected state of the cluster after Node 'B' successfully completes?
All cluster resources will stop, cluster processes halt on Node 'B', and the remaining cluster nodes take no action.
During a customer planning session, a customer presents the details of the current network design. The network design serves four separate company facilities, each on a separate subnet. As part of a disaster recovery plan, the customer is evaluating the use of HACMP. Specifically, they have reviewed IPAT via IP aliasing and are questioning the number of IP subnets required. Given the customer's current configuration, what is the recommended minimum number of IP subnets that would be required?
During a planning session, a customer expresses concern about the behavior of the nodes when HACMP is installed. One department is reliant on an older piece of equipment that is critical to operations and must communicate to the database server. Previous network changes indicated that this piece of equipment does not support gratuitous ARP. How can HACMP best incorporate this restriction into the cluster design?
HACMP can use either hardware address take over, or clinfo can update the arp cache of the clients when an adapter event occurs.
Continuous availability is required for an application in a two-node cluster. There is one rotating resource group in the cluster, and the nodes are identical machines. Client telnet connections to the application cannot be broken because it takes 20 to 30 minutes to recycle the client workstations and transactions will be lost. What configuration best provides continuous connectivity to the application?
None because continuous telnet connectivity cannot be provided with HACMP.
During the process of fallover testing, a client/server database needs to be tested. On the AIX Server, start/stop/cleanup scripts and server application functionality have been tested. Before moving onto the next step of fallover testing, what action must be completed?
Verify that the client side application is functioning properly.
Which of the following statements is true about IPAT via IP replacement?
Configuration of Hardware Address Takeover is supported.
While executing a Cluster Test Plan, the administrators realize that they will need to email the operations and network on-call staff in the event of a network outage. What can be done to alert the on-call staff of network failures?
Customize an Event Notification method.
An HACMP cluster has three nodes of unequal size. The largest node is Node 'A'. The primary application on Node 'A' has two components. One component is memory intensive, the other is IO intensive. Nodes 'B' and 'C' are running web servers and databases that access information produced by applications on Node 'A'. The customer requires the most performance from the primary application and wants all nodes available for fallover processing. Which of the following is the best solution for meeting the customer requirements?
Use dynamic node priority to manage resource groups.
A recommended PTF is released to resolve reliability issues with bos.net.tcp.server. A maintenance window has been scheduled to install the PTF on all nodes. In order to complete the maintenance, what are the next logical steps to be completed after the install of the PTF on all nodes?
Verify all necessary applications and resource groups are functioning properly and test Ethernet fallover on all nodes.
While reviewing the HACMP log file, a warning message appears stating that a custom event script could not be found. A check of the current path settings indicates that all required paths are present. What step can be taken to correct the missing paths issues during cluster event processing?
Include required paths at the start of the custom script.
A customer has installed HACMP and is configuring a three-node cluster. One of the applications uses the AIX hostname as a key for authorization to execute on the specific computer. The application vendor was contacted and consented to having multiple copies of the application but required that only one application be active at any one point in time. How can the AIX host name dependency be managed in an HACMP environment?
The application start and stop script for the application can manage the hostname.
The IP subsystem fallover behavior of a two-node cluster is being tested in a large client organization. One of the tests is to simulate a network_down event and verify that HACMP moves the single resource group to the standby node. What is the best way to simulate a network down condition?
Simultaneously unplug all network cables from the node that has the resource group.
A customer has several network adapters configured as backup paths using the Virtual IP addressing and dynamic routing of AIX. A second node needs to be added to the environment to create a cluster. What considerations must be addressed during the planning session?
The AIX Virtual Addressing mechanism is not supported by HACMP, but a similar configuration can be created in HACMP.
A customer is dependent on static routing to make connections from the nodes to the user community. What feature of HACMP can be used to address this issue?
Pre and post events
A three-node cluster experienced a node failure. The resource groups were acquired as expected. The node has been repaired, and HACMP has been restarted. How does the operator determine that the node has successfully re-integrated into the cluster?
Use the clstat command.
During a customer planning session, the LVM requirements for HACMP are provided. The administrator explains that there are five volume groups in the current production environment because filesystems constantly change due to the nature of the business. Maintenance outages are limited to one Saturday per month. There is some concern about the amount of time that will be required for outages once HACMP is installed. With HACMP, what will be the simplest method of managing LVM changes while minimizing required downtime?
Use the Cluster Single Point Of Control to manage LVM changes.
A cluster administrator is performing fallover testing during a scheduled system outage. Node 'A' was forced down and eventually rebooted. When Node 'A' reintegrates into the cluster, the resource group state is offline. No errors are found in reviewing the hacmp.out file. What could have been done to ensure the resource group was acquired when cluster processing was restarted on Node 'A'?
Select reacquire resources after forced down on start menu.
A Cluster Administrator is implementing a custom resource group for a new web application. The goal is to ensure the resource group is always active. Preference should be given to the priority node at startup time, but fallback to the priority node should be scheduled during a nonbusy time. Using custom resource group options, what combination best achieves the desired resource group behavior?
Online On First Available Node, using both settling and fallback timers
An operator just completed a repair to a disk subsystem that contains a disk heartbeat. Before returning the system back into production mode, there is a requirement to verify the disk heartbeat is functioning. How should the operator determine if the disk heartbeat network is functioning?
Use the clstat command to verify the appropriate network connections are available.
A customer has two sites that are 3072 kilometers apart and wants to implement HACMP between the two locations. The nodes at each location must be capable of running the application in the event of a failure. Currently the customer's disk technology is SCSI. How can HACMP be used in this environment?
HACMP/XD:HAGeo has unlimited distance mirroring that may be appropriate for this environment.
A customer has a cluster that is administered remotely. It is desirable to have the nodes automatically power up and start HACMP in the event of a power outage or operating system dump. What configuration parameter can be set to perform this function for HACMP?
The "on restart" option in clstart
A customer has a resource group that contains a database that is accessed by a web client based on another system. The communication to the database is TCP/IP. Occasionally after a failure, the end users get unusual error messages that indicate the database server is not available. It has been determined that the web server is trying to access the database server's information before it has completed file system recovery after a failure. How can HACMP affect this problem ?
HACMP can delay making the network label available until after the filesystem checks have completed with the "filesystem mounted before IP is configured".
A customer wishes to use IPAT via IP replacement in a two-node mutual cluster. Each node has two Ethernet 10/100/1000 adapters. Throughout the planning session, the customer has asked questions about subnet and VLAN requirements for the intended network topology. Which of the following best describes the network requirements?
One non-service and one service IP address must be on the same logical subnet. All interfaces must be on the same VLAN.
A customer is implementing an HACMP solution. The application has a license key based on the hostname. The hostname is also the name used for one of the TCP/IP interfaces. The application vendor has worked with HACMP systems before and is aware of the issue. The preferred method for resolving this issue is to:
obtain a second license key for a "standby" system from the application vendor.
A customer has decided to implement the "Fast Disk Takeover" and "Disk Heart Beat" features of HACMP. What IBM distribution media will be required for a default installation?
AIX and HACMP media
A customer has a four-node cluster (Nodes 'A', 'B', 'C', and 'D'). The nodes are set up in resource group pairs. Node 'A' falls over to Node 'B', and Node 'C' falls over to Node 'D'. During a maintenance window, it was discovered that the six resource groups on Node 'A' are too much to be adequately handled by Node 'B'. What could be done to better utilize the available resources during fallover?
Implement Dynamic Node Priorities in the resource groups in order to find a node with suitable system resources.
What is required during the initial creation of a cluster?
The cluster name and a TCP/IP path to each of the nodes
A HACMP/XD:HAGeo cluster has been installed with four nodes, two nodes at each site. The cluster is protected from site isolation by the DBFS facility in HACMP/XD:HAGeo. In the event of a site failure, what happens?
The non-dominate site initiates a call to the dominate site to verify if any of the nodes are functional.
A customer has an application that requires a serial connection for Electronic Data Interchange. What must be done to integrate the serial connection into an HACMP cluster?
Use a network based terminal server for the serial connection.
A two-node cluster has both a single TCP/IP heartbeat and a single non-IP serial heartbeat. The Service Level Agreement regarding the speed of cluster fallover has been updated, and Fast Disk Takeover will be implemented to meet the new Service Level Agreement. Which of the following should be implemented to increase the reliability of Fast Disk Takeover?
Disk heartbeats should be added over multiple disk adapters.
Following a successful migration, it is highly recommended that concurrent Volume Groups be converted to enhanced concurrent capable Volume Groups. In order to follow this recommendation, what step must be completed first?
The Volume Groups must be a member of a resource group and must be varied off all nodes of the cluster.
The Cluster Administrator wants to use disk heartbeats as a non-IP network. What Volume Group consideration must be accounted for prior to configuring disk heartbeats?
The Volume Group must be created as Enhanced Concurrent Capable.
A cluster node has three network interfaces that support three resource groups with three service IP labels. Each IP label uses IPAT via IP aliasing. One of the adapters fails, and all of the IP labels on it are moved to other interfaces as expected. The system supports the "hot swap" replacement of the network interfaces, and the system engineer has corrected the problem. What will happen to the service IP labels when the repaired adapter is noticed by HACMP?
They will not move; the interface will join the cluster as an available non-service adapter.
The snmpinfo command can be used to obtain information about the current state of a running cluster. Which of the following will provide an administrator with a complete list of information from the cluster MIB?
snmpinfo -m dump -o /usr/es/sbin/cluster/hacmp.defs -v cluster
A two-node cluster provides NFS exports to a variety of UNIX clients. During fallover testing, it is determined that some clients are unable to mount the exported filesystems after fallover, and before node reintegration. It is verified that the filesystems belong to a resource group and are correct in /usr/es/sbin/cluster/etc/exports. Which of the following may provide a possible solution to this problem?
Use the "Filesystem Mounted before IP Configured" flag to provide more predictable NFS serving.
A customer is migrating from a third party managed disk subsystem to an IBM TotalStorage Enterprise Storage Server. Connection of the new storage server is complete, and it has been tested on all of the nodes in the three node HACMP cluster. How does the customer migrate to the new storage server with minimum down time?
Use C-SPOC to add the new disks to the cluster and migrate the information off the existing disks, then remove the old disks with C-SPOC.
A customer has created a process application monitor for an application that is initiated by a "c" shell script. The application monitor is not functioning as designed. What could be the issue with the application monitor?
A process monitor cannot monitor an application started by a shell.
What is a prerequisite for using the application availability analysis tool in HACMP?
An application monitor must be associated with the application to monitor.
A cluster with 10 adapters per node is converted to use IPAT address takeover via IP aliasing. During the conversion, the customer ran out of IP subnets in the allocated range. What feature in HACMP can help resolve this issue?
Heart beating via IP aliasing
Automatic error notification can be configured to manage which type of disk error?
Using the C-SPOC utilities, an administrator has created a new volume group and filesystems. The administrator performs an lsvg -l command on the volume group and notices that the log logical volume created in C-SPOC is closed/syncd and loglv01 has been created and is in use. What step was required in order for the logfile created by C-SPOC to be used?
Manually perform a logform outside of C-SPOC prior to creating any filesystems.
When defining resources for a concurrent resource group, which of the following is NOT supported?
Definition of Multiple IP aliases
During cluster fallover testing, the administrator determines that additional work is required to stop existing test applications on the standby node, prior to starting the production applications on the surviving node. To manage this type of requirement, which step should be performed first?
Add a custom user defined event.
While reviewing cluster logs, an administrator notices that an event, network_down rs232a, occurred on the previous day. After verifying the physical network connections and device states, what command can be used to verify the cluster state?
While performing LVM maintenance, the volume group datavg1 was exported and re-imported. The volume group is varied online, and the filesystems have been mounted. However, the database will not start. What action would most likely resolve this problem?
Perform a chown command on the VG device files.
Node 'C' has just completed reintegration in a three-node cluster. The hacmp.out logfile displays the event node_up_complete for Node 'C'. The cluster administrator wishes to verify that the cluster is stable. In addition to reviewing the cluster log files, what utility should be run to verify the cluster state?
When a problem occurs in an HACMP cluster, which log file will show the high level overview of the cluster operation?
If a kill -9 is performed on the clstrmgrES process, the default action is to halt the system. To change the default action to do something other then halting the system, which of the following files must be modified?
As part of fallover testing of a corporate database server, client application connectivity must be verified. Which of the following would best verify the application functionality?
Perform successful reads and writes to the application.