11.6.13 Redundancy and High Availability
You manage your company's website, which uses a cluster of two servers with a single shared storage device. The shared storage device uses a RAID 1 configuration. Each server has a single connection to the shared storage and a single connection to your ISP. You want to provide redundancy so that a failure on a single component doesn't cause the website to become unavailable. What should you add to your configuration to accomplish this?
Connect one server to the internet through a different ISP. In this scenario, the ISP is the single point of failure. If the ISP connection goes down, the website will be unavailable. Connecting one server to a different ISP or both servers to two ISPs provides redundancy for the connection.
You have purchased a solar backup power device to provide temporary electrical power to critical systems in your data center should the power provided by the electrical utility company go out. The solar panel array captures sunlight, converts it into direct current (DC), and stores it in large batteries. The power supplies on the servers, switches, and routers in your data center require alternating current (AC) to operate. Which electrical device should you implement to convert the DC power stored in the batteries into AC power that can be used in the data center?
Inverter A power inverter changes direct current (DC) power to alternating current (AC) power. In this scenario, you can use a power inverter to convert the DC power stored in the batteries to AC power that your servers, switches, and routers can use in an emergency.
For a true fault-tolerant strategy, you must consider all system components. Ethernet bonding (also called adapter teaming) is a fault-tolerant strategy that uses multiple network adapters configured on a failover solution. In the event of a NIC failure, other adapters automatically provide link redundancy. Multiple adapters can also increase performance by distributing the network load between adapters.
Ethernet bonding Ethernet bonding (also called NIC teaming) logically groups two or more physical connections to the same network. If one NIC fails, the second NIC with a connection to the same network can still be used.
Which of the following devices accepts incoming client requests and distributes those requests to specific servers?
Load balancer A load balancer is a device that accepts incoming client requests and distributes those requests to multiple servers. One goal of load balancing is to distribute client requests evenly between multiple servers to improve performance.
You are adding a new rack to your data center, which will house two new blade servers and a new switch. The new servers will be used for virtualization. The only space you have available in the data center is on the opposite side of the room from your existing rack, which already houses several servers, a switch, and a router. You plan to configure a trunk port on each switch and connect them with a straight-through UTP cable that will run across the floor of the data center. To protect equipment from power failures, you also plan to install a UPS on the rack along with redundant power supplies for the server. Will this configuration work?
No, you should not run a cable across the data center floor. In this scenario, running a cable across the data center floor represents a tripping hazard. It also represents a point of failure, as the cable will be walked on constantly, resulting in it being kicked out of one or both jacks. It will also likely fail prematurely due to the excessive wear. A better option would be to run the cable through the ceiling plenum.
What is the purpose of using Ethernet bonding? (Select two.)
Provides a failover solution for network adapters Increases network performance For a true fault-tolerant strategy, you must consider all system components. Ethernet bonding (also called adapter teaming) is a fault-tolerant strategy that uses multiple network adapters configured on a failover solution. In the event of a NIC failure, other adapters automatically provide link redundancy. Multiple adapters can also increase performance by distributing the network load between adapters.
In addition to performing regular backups, what must you do to protect your system from data loss?
Regularly test restoration procedures. The only way to ensure that you have protection against data loss is to regularly test your restoration procedures. This activity reveals whether or not your backup process functions properly and your restoration and recovery procedures are accurate.
Which of the following is the least effective power loss protection for computer systems?
Surge protector A surge protector provides no power loss protection.
Why should you store backup media off site?
To prevent the same disaster from affecting both the network and the backup media Backup media should be stored off site to prevent the same disaster from affecting the network and the backup media. If your primary facility is destroyed, your only hope of recovery is off site data storage.
You are adding a new rack to your data center, which will house two new blade servers and a new switch. The new servers will be used for file storage and a database server. The only space you have available in the data center is on the opposite side of the room from your existing rack, which already houses several servers, a switch, and a router. You plan to configure a trunk port on each switch and connect them with a crossover UTP plenum cable that will run through the suspended tile ceiling in the data center. To provide power for the new devices, you had an electrician install several new 20-amp wall outlets near the new rack. Each device on the rack will be plugged directly into one of these new wall outlets. What is wrong with this configuration? (Select two.)
You should implement a UPS between the wall outlet and the network devices. You should implement redundant power supplies for the network devices. In this scenario, all the devices on the new rack will go down if the power from the wall outlet fails for some reason (such as a power outage). To prevent this from happening, you should implement a UPS between the wall outlets and the network devices. In addition, the power supplies used by computing equipment have finite life spans and fail frequently. Because these are mission-critical devices, you should consider implementing redundant power supplies.