1. The DKT Society wishes to make a local copy of its 256 GB production LUN for testing purposes. It will be mounted to the Production UNIX host. DKT requires a solution with minimal disruption to its application.
What do you recommend?
A.Create a Snapshot of the production LUN, and start a session. Add the Snapshot to the production host’s Storage Group
B.Use an Incremental SAN Copy Session on the production LUN. Add the new LUN to the production host’s Storage Group
C.Clone the production LUN. Remove the Clone from the Clone Group. Add the new LUN to the production host’s Storage Group
D.Take the production LUN offline. Perform a full SAN Copy. Add the new copy to the production host’s Storage Group, and restart the application
Answer: C
2. ABC Corp. wants to implement an extended distance DR solution for a custom application. Part of this application is transactional with small I/O and is mission critical with an RPO of less than 30 minutes. Another component of this application consists of image repositories and is non-mission critical with an RPO of 24 hours. The link meets the minimum bandwidth requirements; however there is significant latency because of the distance. Which software solution best meets the requirements?
A.MirrorView Synchronous for the mission critical data and MirrorView/A for the image repositories updating every 12 hours
B.MirrorView/A for all data with mission critical data synchronizing every 0 minutes from the end of the last update and image repositories every 12 hours
C.MirrorView Synchronous for all data with the mission critical data constantly synchronizing and the image repositories only synchronizing during the backup windows from midnight to 6:00 AM
D.MirrorView/A for all data with mission critical data synchronizing every 0 minutes from the end of the last update and MirrorView synchronous for image repositories only synchronizing during the backup windows from midnight to 6:00 AM
Answer: B
3. When implementing SAN Copy Incremental over long-distances and the link latency is unknown, which setting will measure latency by sending test I/O to the destination?
A.Set Latency to Auto
B.Set Latency to Test
C.Set Link Bandwidth Auto
D.Set Latency to the lowest setting
Answer: A
4. Which two [2] applications are supported by Replication Manager/SE?
A.NTFS on Windows 2000
B.Oracle 8i on Windows 2000
C.Exchange 5.5 on Windows 2000
D.Exchange 2000 on Windows 2003
Answer: AC
5. CTS, Inc. has an Exchange environment with 3000 users, each generating one IOPS during peak load. CTS’ workload is a 2:1 Read / Write Ratio. Being price sensitive, CTS planned a RAID-5 solution. What should the customer expect to see in terms of IOPS at the back end?
A.3000
B.4000
C.5000
D.6000
Answer: D
6. CDF Inc. wants to purchase a CLARiiON array to add to its SAN. Path Failover and Load Balancing are requirements. CDF already owns and uses DMP on its Symmetrix. What are two [2] solutions you would recommend?
A.Dual HBAs with no cross zoning and DMP
B.Dual HBAs with no cross zoning and PowerPath
C.Multiple HBAs, Multi-path with cross zoning utilizing four FC ports per SP, and DMP
D.Dual HBAs, Multi-path with cross zoning utilizing two FC ports per SP, and PowerPath
E.Multiple HBAs, Multi-path with cross zoning utilizing four FC ports per SP, and PowerPath
Answer: DE
7. Click the Exhibit button.
Review the replication solution shown in the graphic. In the event of production data corruption, which best describes the required steps to recover production data from the remote Snapshot?
A.Unmount production LUNs, roll back Snap, restore mirror, mount production LUNs
B.Unmount production LUNs, fracture, roll back Snap, synchronize, promote mirror, mount production LUNs
C.Unmount production LUNs, promote mirror, fracture, roll back Snap, synchronize, promote mirror, mount production LUNs
D.Unmount production LUNs, roll back Snap ,promote mirror, fracture, synchronize, promote mirror, mount production LUNs
Answer: C
8. The MirrorView/A promotion script has lines which perform the following operations:
1: Flush the Primary Image buffers
2: Synchronize the Secondary Image
3: Wait until the synchronization completes
4: Promote the Secondary Image
What would you change to ensure that the script still works in the event of a disaster?
A.Change line 1 to flush the Secondary Image buffers
B.Change line 3 to check for the Synchronized or Consistent state
C.Add a line 5 to check that the Primary Image is demoted successfully
D.Add a delay between steps 1 and 2 to allow the buffers to complete flushing
Answer: B
9. Your customer’s file system application uses small, multi-threaded, sequential I/O. The company noticed a decrease in overall performance on its server. What do you recommend to fix the problem?
A.Increase the server memory
B.Increase the number of threads
C.Set file system parameters to minimize file system coalescing
D.Set file system parameters to maximize file system coalescing
Answer: C
10. HHZ Management Inc. keeps its data on a 512GB 4+4 RAID-1/0 LUN, bound on 73GB 15Krpm drives. HHZ will use a SnapView Snapshot for backup purposes, and keep the session active for the full 24-hour period between backups. You investigate its environment and determine the following:
The application generates random 2kb I/Os, with a read/write ratio of 4:1
The average seek distance is 20 GB
The application generates a throughput of 250 IOPs for 10 hours/per day and is idle for the remaining 14 hours
You need to estimate the size of the Reserved LUN Pool and the write throughput of the Reserved LUN Pool.
What is the best initial estimate?
A.120 GB, with fifty 64KB writes/second
B.150 GB, with twenty five 128KB writes/second
C.200 GB, with fifty 2KB writes/second
D.512 GB, with two hundred fifty 2KB writes/second
Answer: A
11. BG Industries storage outgrew the capacity on a CX400. A new CX700 was ordered as a replacement for the CX400. After the data is migrated to the CX700, BG Industries intends to repurpose the CX400. Because storage on the CX400 was inefficiently allocated on some LUNs, BG Industries intends to reduce the size of these LUNs. The result is that some of the new LUNs will be the same size, and some will be smaller.
Which feature will achieve this goal?
A.SAN Copy
B.MirrorView
C.MirrorView/A
D.None of them
E.CLARiiON LUN Migration feature
Answer: D
12. MNR Industries bought two CX600s, one for New York and one for New Jersey. Each CX600 is configured with four DAEs. RAID groups span the DAEs. A new application has to be brought up in New Jersey but MNR does not have the budget to purchase additional capacity. The customer decided to reallocate a DAE from the New York site to the New Jersey site to address the capacity requirements.
What is needed to accomplish this task?
A. Discover all of the LUNs from the Storage Groups which are spread across the fourth DAE in New York and delete them
Disconnect the fourth DAE and reallocate it to New Jersey
B. Create New LUNs
Copy the LUNs which are spread over the fourth DAE to the new LUNs with SAN Copy
Modify the LUN configuration in the Storage Groups
Disconnect the 4th DAE and reallocate it to New Jersey
C. Create New LUNs
Copy the LUNs which are spread over the fourth DAE to the new LUNs with SnapView Clone
Modify the LUN configuration in the Storage Groups
Disconnect the fourth DAE and reallocate it to New Jersey
D. Discover the LUNs which are spread across the fourth DAE
Answer: D
13. ABC Corp. purchased a CX700 for eight Windows servers. The current storage requirement is 150 GB per server on RAID-5. Performance is considered important on these LUNs and each server is expected to generate around 500 IOPs. Expected growth is 50 GB every three months.
ABC Corp charges its clients a monthly fee based on allocated storage per server. The storage budget allows for the purchase of disks every six months. ABC is starting with five 36 GB disks in the vault locations for low I/O file shares, two 73 GB hot spares and forty 73 GB disks.
Which configuration will meet performance, scalability and growth requirements while considering ABC Corp’s internal billing and purchasing policies?
A. Configure (4+1) R-5 RAID Groups with metaLUNs consisting of two 75 GB components striped across two RAID Groups
Future growth will involve expanding the metaLUNs in 50 GB increments from one new RAID Group
B. Configure (9+1) R-5 RAID Groups with two LUNs sharing one RAID Group
Future growth will involve creating two new LUNs on a new RAID Group and using Windows Dynamic Disk to create a striped volume on the server
C. Configure (4+1) R-5 RAID Groups, each containing a single max LUN on a dedicated RAID Group
Future growth will involve expanding the LUN to a metaLUN by striping on another max LUN from a new dedicated RAID Group
D. Configure (4+1) R-5 RAID Groups with LUNs of 50 GB
Answer: D
14. FND Bank has CLARiiON CX-700 and 30 Windows Servers, 15 HP UX 11i Servers and 18 IBM AIX 5.1 Servers connected to it. All the servers have two HBAs and are zoned to two different Cisco 9509 Switches (IOS 2.x) for HA. FND Bank is adding more Windows Servers to its Cisco SAN.
Which precaution should FND Bank take before adding any new server zones to the SAN to ensure there is no interruption to its production environment?
A.Take necessary steps to set Domain_ID and do not set Persistence FCID
B.Take necessary steps to set Domain_ID and persistence FCID after new zoning
C.Take necessary steps to set Domain_ID and persistence FCID before new zoning
D.Take one SAN switch off line at a time, do zoning applied changes and bring back online
Answer: C
15. ACME Corporation requested that you design a CLARiiON solution to meet its performance requirements. ACME has a Windows application with random access characteristics. ACME provided you with the following current performance statistics using PerfMon on one production host:
Disk reads/sec: 300 (avg)
Disk writes/sec 229 (avg)
Avg Disk bytes/read: 15585
Avg Disk bytes/write: 17377
Here are the design considerations:
Connect three hosts with identical performance
Requires 3X its current performance
Configure using RAID-5
Configure using 146GB 10K FC drives
How many spindles are required to meet the performance requirements?
A.19
B.31
C.57
D.92
Answer: D
16. HS Special Effects Inc uses a Windows-based application to capture and edit large graphics files with application I/O sizes of 256 KB. The customer standardized on 4+1 RAID-5 LUNs. You wish to achieve maximum performance on the LUNs by correctly aligning the customer’s data. HS will be using SnapView Snapshots for backup purposes.
What is the best solution?
A.Use diskpar with a reserved block count of 128
B.Use diskpar with a reserved block count of 512
C.Bind each LUN with a CLARiiON LUN Offset value of 63
D.Bind each LUN with a CLARiiON LUN Offset value of 128
Answer: B
17. You are planning a SnapView implementation for your customer, FDF Manufacturers. Its production application is write-intensive and FDF plans to use the Snapshot of its production LUN for write-intensive testing. The access pattern for the production application and the testing will be largely random, with a mix of I/O sizes.
Which drive type and RAID type is the best choice to use for the Reserved LUN Pool?
A.Writes to the Reserved LUN Pool are random, of various sizes, as are reads. FC drives with RAID-5 are a good choice
B.Writes to the Reserved LUN Pool are random, of various sizes, as are reads. FC drives with RAID 1/0 are a good choice
C.Writes to the Reserved LUN Pool are random 64 KB, and reads are sequential 64 KB. This makes ATA drives with RAID-3 a good choice
D.Reads from the Reserved LUN Pool are random 64 KB, and writes are sequential 64 KB. ATA drives with RAID-5 are therefore the best choice
Answer: B
18. When a CLARiiON is doing a write to a RAID-5 LUN and the write data is reasonably sequential, what happens when there is not enough data in cache to fill the stripe?
A.Flare will bypass MR3 writes for this operation and write directly to disk.
B.Flare will read data from the physical disk to fill in the ‘holes’, calculate the parity, then write the data and parity to disk in a single operation.
C.Flare will write the data up to the first ‘hole’. It will read data from the physical disk to fill that hole and all remaining ‘holes’, calculate the parity for the entire stripe, then write the remaining data and parity to disk during the second operation.
D.Flare will use a modified MR3, which will write the parity and data up to the first ‘hole’ then will read data from the physical disk to fill that and all remaining ‘holes’, calculate the parity for the entire stripe, then write the remaining data and parity to disk in a second operation.
Answer: B
19. Volume Managers allow the use of plaids on the CLARiiON. The primary performance impact is the manner in which CLARiiON LUNs are striped. Which two [2] points must be considered?
A.Plaids are only useful when the overlying applications are multi-threaded
B.Use host based RAID implementations that do not require parity calculation
C.Use host based RAID implementations that require parity calculation (i.e. RAID-3, RAID-5)
D.LUNs should not be plaid together from RAID groups with different RAID types or stripe sizes
E.The host manager stripe element size has to be equal to, or smaller than the CLARiiON LUN stripe size
Answer: BD
20. Your customer plans to use a local copy of its 512 GB database for testing. The test host will use 4 KB random I/Os with a read/write ratio of 1:1. The LUN activity will be constant at 750 IOPs, and tests will run for 24 hours. In an attempt to save money, the customer wishes to use a Snapshot for testing rather than a Clone.
Which reason do you use to persuade them to use a Clone?
A.The synchronization of the Clone will have a larger performance impact than COFW
B.The synchronization of the Clone will have a smaller performance impact than COFW
C.The Reserved LUN Pool will need to be approximately 1.1 times the size of the Production LUN
D.The Reserved LUN Pool will need to be approximately 2.1 times the size of the Production LUN
Answer: D
21. ABC Corp is having performance problems on a LUN. It has been determined that there is spindle contention in the RAID Group. The host cannot tolerate any downtime. A new LUN has been created on dedicated spindles.
How should the data be moved?
A.Use DRU to copy the data to a new LUN
B.Use SAN Copy to copy the data to a new LUN
C.Use RoboCopy to copy the data to a new LUN
D.Use CLARiiON LUN Migration to copy the data to a new LUN
Answer: D
22. ABC Corp has an existing production application that has been experiencing poor performance. EMC Engineering determined that the stripe element size should be 32KB. The current production LUNs have been configured with the default stripe element size.
How can the default stripe element size be changed while minimizing the impact to the production application?
A.Use LUN Migration to copy the data to a new LUN with a stripe element size set to 64
B.Use the Microsoft diskpar utility to change the stripe element size on the file system to 32
C.Enter engineering mode and change the stripe element size to 64 in the LUN properties for the production LUN
D.Use SAN Copy to copy the data to a new LUN then enter engineering mode and change the new LUNs stripe element size to 32
Answer: A
23. Your customer is running an internet e-mail service on Solaris using Veritas Volume Manager, connected to a CLARiiON CX700. Due to growth the CX700 initial configuration is at its limits and some LUNs show disk contention due to heavy mail-users. A disk upgrade has been sold.
Which steps should be taken to improve performance on overloaded LUNs and also solve the need for more disk space? There should be as little downtime as possible on the host.
A.Create new and larger LUNs on new drives and establish a new mirror using Veritas Volume Manager
B.Create new and larger LUNs on new drives and migrate data using any copy software (i.e. Robocopy)
C.Create new larger LUNs which reside on new drives using the LUN migration feature provided by FLARE16
D.After having created the new and larger LUNs use SAN Copy incremental to migrate the data from the old LUNs
Answer: A
24. Which restriction has to be taken in account when planning a LUN migration?
A.The destination LUN cannot be in an expansion process during a LUN migration
B.The destination LUN cannot be part of layered applications like MirrorView or SnapView
C.An expanding source LUN can only be migrated when the destination LUN is in expansion as well
D.The source and the destination LUN must not be part of layered applications such as MirrorView or SnapView
Answer: B
25. Which restrictions are associated with LUN Migration and MirrorView/S?
A.Primary and Secondary images may be migrated as long as the mirror is out of sync on the same size LUN
B.Primary or Secondary images may be migrated as long as the mirror is not in a consistent state on the same size LUN
C.Primary and Secondary images may be migrated as long as the mirror is in a synchronized state on the same size LUN
D.Primary or Secondary images may be migrated as long as the mirror is not in a synchronized state on the same size LUN
Answer: D
26. The SHA Company will be using MirrorView/S to replicate 40 LUNs from a CX700 to another CX700 at a remote site. It has 10 LUNs in the 100 MB to 10 GB range, 15 LUNs in the 10 GB to 100 GB range, and 15 LUNs larger than 400 GB. All LUNs have a random, small-block I/O profile, with read:write ratios close to 60:40. Recoverability in the event of a disaster is important, but the customer wishes to optimize performance.
Based on these criteria, what is your recommendation for Write Intent Log use and why?
A.Use the Write Intent Log on all LUNs. 40 is less than the allowable limit for the CX700.
B.Use the Write Intent Log on the 25 largest LUNs only. The CX700 allows only 25 LUNs to use the Write Intent Log.
C.Use the Write Intent Log for all LUNs over 10 GB in size. The performance impact is greater than the benefit for small LUNs.
D.Use the Write Intent Log only for LUNs under 10 GB in size. The performance impact is greater than the benefit for large LUNs.
Answer: C
27. Click the Exhibit button.
Your customer uses CX500 and CX700 arrays with existing mirrors. They want to add an application on SITE B and configure an additional five Mirrors using write intent logs. They want to enable WIL on all primary mirrored LUNs on the CX500. In the event that all the secondary images are promoted to primary on the CX500 they want to be able to enable WIL on these LUNs as well.
Will this configuration meet the customer’s requirements?
A.No, it will surpass the WIL maximum on the CX500.
B.Yes, it will meet the customer requirement as a valid configuration.
C.No, it will surpass the maximum images allowed on CX500 even if WIL is not enabled.
D.Yes, it will work but only if four WIL LUNs are added to avoid surpassing the maximum number of LUNs allowed.
Answer: A
28. Click the Exhibit button.
REW You need to analyze a customer’s NAR file. Pay particular attention to the LUN shown in the graphic. How do you describe what’s going on to your customer on this LUN?
A.Everything is normal: dirty pages are below the high watermark. You are not concerned about forced flush/s.
B.Everything is normal: dirty pages are above the watermark limit set to 80% and the forced flush/s ratio is high. However, LUN performance is still good.
C.Everything is normal. Dirty pages are above the watermark limit set to 80%. However, the forced flush/s ratio is normal which still provides relatively good performance.
D.Everything is going badly because both dirty pages and forced flush ratios are too high. If the dirty pages in the cache and the forced flushing on the array are reduced, this LUN will perform better.
Answer: D
29. The customer has asked for your help using Navisphere Analyzer to better understand the randomness of the I/O profile. You launch the disk component of Analyzer. Which parameter would be most helpful?
A.Write Size
B.Queue Length
C.Response Time
D.Average Seek Distance
Answer: D
30. Your customer is running an environment with two sites, one CX500 each, MirrorView/S to meet the company’s RTO requirements for its call data records. The full capacity of 25 mirrored LUNs is configured using the WIL.
Due to the unexpected growth of the company in the last few months, the bill run at the end of the month suffers from a response time increased by 50% due to the CPU utilization of the CLARiiON. This brings the OLTP application to the limits of the available batch window.
Management realized that they have to migrate to a CX700 platform. Which actions can be taken to improve the next bill run performance temporarily?
A.Disable the WIL to offload the CPU utilization
B.Disable the Fracture Log to offload the CPU utilization
C.Disable Write Cache of affected LUNs to decrease the response time
D.Disable Read Cache of affected LUNs to decrease the response time
Answer: A
Link : http://www.killtest.co.kr/TechnologyFoundations/e20-820.asp