HH0-440

Exam Name: Storage Architect - Performance & Virtualization Exam Type: Hitachi Exam Code: HH0-440 Total Questions 117

Page 1 of 28

Questions: 1 A customer runs two geographical dispersed data centers and plans to merge both data centers' SANs. Which three should be considered when planning the ISL installation and configuration? (Choose three.) A. Distance B. ISL trunk setup/configuration C. Estimated throughput D. Fabric trunk setup/configuration E. Buffer credits Answer: A, B, C Questions: 2 Which HDS product is used to define and report Logical Group configurations? A. Hitachi Tuning Manager B. Hitachi Device Manager C. Hitachi Tiered Storage Manager D. Hitachi Storage Services Manager Answer: B Questions: 3 When planning for capacity growth and performance for a new application, which three metrics are necessary to determine the required back-end storage configuration? (Choose three.) A. Anticipated read/write IOPS B. Cache read hit ratios C. Queue depths D. Transfer rates E. IOPS per physical disk Answer: A, B, E Questions: 4 Which three actions will help to reduce the level of Write Pending data in cache during periods of intense random update processing? (Choose three.) A. Switch to faster disk drives. B. Migrate from RAID5 to RAID1. C. Spread the load across more RAID Groups. D. Spread the load across more ports. E. Spread the load across more LUNs. Answer: A, B, C Questions: 5 Click the Exhibit button. Which three actions will help to reduce the level of Write Pending data in cache during periods of (Choose three.)


Page 2 of 28

A. Insufficient cache B. Loss of redundancy in write cache C. Heavy random writes to RAID5 D. Heavy random writes to SATA drives E. Failure to distribute workload among sufficient array groups Answer: C, D, E Questions: 6 A Microsoft Exchange administrator team is responsible for managing five mailbox servers and maintaining acceptable service level objectives for your organization's email users. Accordingly, they want to setup Windows Performance Monitor to actively monitor the Exchange servers. Which Performance Monitor counter should be monitored for the Exchange database LUNs? A. Disk write bytes/second B. Disk read bytes/second C. IO delays per sec D. Average disk read latency (ms) Answer: D Questions: 7 A customer reports erratic occurrences of concurrent slow downs involving multiple servers directly attached to a large enterprise storage system. Each server supports an independent application and uses dedicated port processors and RAID Groups. Storage replication, thin provisioning, and external storage are used. If this problem originates in the storage, what are the two most likely causes? (Choose two.) A. Highly concurrent PAIRCREATE


Page 3 of 28

B. Storage system microcode problem C. High internal path utilization D. High cache Write Pending Answer: A, D Questions: 8 A customer reports OLTP database response times as measured by Oracle have tripled from 1 ms to 3 ms. Comprehensive data collection is present throughout the incident. Which three metrics are most relevant? (Choose three.) A. LDEV response time B. LUN queue depth C. Port utilization (% Busy) D. Cache hit ratio E. BED utilization Answer: A, C, D Questions: 9 When DB2 Deferred Write Threshold (DWT) is reached, high write activity occurs, this can degrade all subsystems involved with the DB2 containers. Within the entire performance chain, which three components are critical during these occurrences? (Choose three.) A. Host CPU B. Host memory C. Storage subsystem cache D. Storage subsystem front-end processors E. Storage subsystem disks Answer: B, C, E Questions: 10 To eliminate disk related performance issues, which three Hitachi Tuning Manager reports would be most useful for analyzing Microsoft 2007 Exchange Storage Group LUN activity? (Choose three.) A. RAID Map reports B. RAID Group Configuration reports C. RAID Group Utilization reports D. Logical Device Response Time reports E. Exchange MAPI Connection reports Answer: A, C, D Questions: 11 Which two Hitachi Professional Services are available for modular and enterprise performance assessment and troubleshooting? (Choose two.) A. Storage Platform Assessment (SPA) B. Implementation Service for Hitachi Tuning Manager C. Implementation Service for Hitachi Storage Services Manager Software D. Tiered Storage Planning and Design Service Answer: A, B


Page 4 of 28

Questions: 12 In IOSTAT, which two Solaris metrics are used to determine the average read block size to a device? (Choose two.) A. Reads per second (r/s) B. Kilobytes read per second (kr/s) C. Block size (bs) D. Read block size (rbs) Answer: A, B Questions: 13 Performance monitor data indicates that CHP utilization may be causing a bottleneck in a Universal Storage Platform (USP) series system during periods of intense processing. Which two actions should be considered to increase IOPS during the bottleneck periods? (Choose two.) A. Configure the paths using high-speed mode. B. Spread the load across more CHP processors and paths. C. Spread the load across more RAID Groups. D. Spread the load across more LUNs. Answer: A, B Questions: 14 Which three statements are true regarding Hitachi storage configurations for Microsoft Exchange Server? (Choose three.) A. RAID6 is a recommended RAID level for the database RAID Groups. B. RAID1+0 is a recommended RAID level for the database RAID Groups. C. RAID5 is the recommended RAID level for the log RAID Groups.. D. LUN concatenation (LUSE) is not recommended. E. VDEV striping is a recommended configuration for database RAID Groups. Answer: B, D, E Questions: 15 Which VMware multi-pathing scheme should be used to avoid non-owning access with HDS modular storage systems with multiple VMware servers? A. Most Recently Used (MRU) B. Fixed C. Round-Robin D. Extended Round-Robin Answer: B Questions: 16 A customer is using Hitachi Dynamic Provisioning (HDP) on a Universal Storage Platform (USP) V and runs an Oracle database on HDP Volumes allocated to a HDP Pool. The customer is experiencing unexpected growth of the database and a lack of capacity in the HDP Pool. What should be done? A. Build a new HDP Pool using the same disk and RAID type and migrate the DP-Volumes to

the new pool.


Page 5 of 28

B. Build a new HDP Pool using any disk and any RAID type and migrate the DP-Volumes to the new pool.

C. Extend the existing HDP Pool with Pool Volumes using the same RAID and disk type. D. Extend the existing HDP Pool with Pool Volumes using any RAID and disk type. Answer: C Questions: 17 Which two statements are true about the Hitachi Tuning Manager Agent for RAID Product Interval (PI) record type? (Choose two.) A. Performance data is collected by the Agent Collector at specified intervals and summarized

for every hour/day/week/month/year. B. Configuration data is collected by the Agent Collector at specified intervals and summarized

for every hour/day/week/month/year. C. Performance Reporter can directly access Product Interval records through Hitachi Tuning

Manager. D. Product Interval. Answer: A, C Questions: 18 A customer, which runs a large Web server farm as an Internet service and hosting provider, wants to consolidate the data from all these servers from the internal server disks to an iSCSI SAN solution which uses a Universal Storage Platform (USP) V as the storage system. The IO is generally random read in small block size. Which two RAID Group configurations would you recommend for the USP V? (Choose two.) A. RAID1 2+2 B. RAID1 4+4 C. RAID5 7+1 D. RAID6 6+2 Answer: C, D Questions: 19 What happens when the Extended Round-Robin load balancing algorithm is used with an LDEV that is executing only random type IOs? A. Performance will be severely impacted. B. Load balancing will be disabled for this LDEV. C. Round-Robin will be used. D. HGLAM will be notified of this inconsistency. Answer: C Questions: 20

A. Which three metrics aid in calculating how long a block of data remains available in cache after its last access? (Choose three.)

A. IOPS rate into the cache partition or storage system B. Number of available cache segments C. The cache hit rate D. Average response time E. Average LUN queue size


Page 6 of 28

Answer: A, B, C Questions: 21 What are two benefits of placing transaction log files on RAID1+0 disks for Microsoft SQL Server databases rather than RAID5 disks under normal operating conditions and during a parity group rebuild process? (Choose two.) A. Improved capacity utilization rates B. Enhanced write performance C. Minimized performance impact of disk hardware failure D. Enabled support for multiple transaction log files Answer: B, C Questions: 22 Which three statements are true about migrating objects within a CLPR/SLPR environment? (Choose three.) A. Only Target ports on the same channel adapter can be migrated to another storage logical

partition. B. LUNs in parity groups can reside within multiple SLPRs. C. Resources can only be migrated within the same CU. D. Only Initiator ports, RCU Target ports and External ports can be migrated to another storage

logical partition. E. RAID Groups that are associated with a port in a particular SLPR must stay within that SLPR. Answer: A, C, E Questions: 23 You enabled Hitachi Tiered Storage Manager to use performance metrics to control and manage migration tasks. Which performance metric(s) should be monitored? A. Only the migration source B. Only the migration target C. The migration target and the migration source D. The migration target and the migration source when the migration target is located on

external storage volumes Answer: C Questions: 24 An application has full cold backups occurring in a nightly maintenance window that consists of files used daily by an engineering firm. What is the maximum recommended utilization level range, measured in minute intervals, for each port microprocessor in an active-active pair servicing a batch workload? A. 30-40% B. 40-50% C. 50-60% D. 80% or higher Answer: D Questions: 25


Page 7 of 28

Hitachi Tuning Manager Agent Level alerts allow which three methods of notification when pre-defined alert thresholds are exceeded? (Choose three.) A. SNMP action notification B. CLI action to produce a notification C. Email notification D. Phone notification E. Screen alert notification Answer: A, B, C Questions: 26 You need to replace an existing system which is breaking a response time service level agreement. Which statement describes the most appropriate plan? A. Build the new configuration capable of exceeding the IOPS rate of the current system during

one minute intervals. B. Build a configuration capable of matching the IOPS rate of the current system during the

busiest one hour period. C. Propose a configuration capable of sustaining 70% of the peak IOPS measured in the current

system over one hour. D. Build a configuration capable of matching the IOPS rate of the current system averaged over

a period defined by the customer. Answer: A Questions: 27 The performance of an audio streaming modular storage system has deteriorated as more streams have been added. Audio streaming is the only workload on the array and the store consists of many thousands of separate audio files. Cache hit rates drop quickly as several streams are started. Which two will improve the prefetch for the audio but prevent prefetch taking place for the directory accesses? (Choose two.) A. Multi-stream mode B. Count of sequential judgment C. High speed write mode D. Disable verify after write Answer: A, B Questions: 28 Which two actions should you apply to the disk configuration of ShadowImage secondary volumes (S-VOLs) to limit the impact of the initial PAIRCREATE and subsequent PAIRRESYNC operations on the primary volumes (P-VOLs)? (Choose two.) A. Create the S-VOL LUNs in the same RAID Groups as the P-VOL LUNs. B. Create the S-VOL LUNs in separate RAID Groups from the P-VOL LUNs. C. Ensure the S-VOL LUNs use separate BEDs from the P-VOL LUNs. D. Ensure the S-VOL LUNs use the same BEDs as the P-VOL LUNs. Answer: B, C Questions: 29


Page 8 of 28

Which two techniques can be used in an HDS modular array to ensure maximum throughput (MB/s) in a large block application such as oil exploration analysis, medical, or video file storage? (Choose two.) A. Build the RAID Groups so that all the back-end loops have the same number of disks. B. Build the RAID Groups vertically so that all disks in the same slot position are in the same

RAID Group. C. Use LUSE to build volumes that are spread across multiple RAID Groups. D. Use host based striping to distribute the load across both controllers. Answer: A, D Questions: 30 You are implementing a Hitachi High Performance NAS (HNAS) cluster system with a Universal Storage Platform (USP) V as the storage system. Currently you start planning the FC SAN infrastructure environment. Which two actions should you perform when planning the FC SAN structure? (Choose two.) A. Connect HNAS FC Port 1 from Node 1 and Node 2 to two separate USP V Clusters. B. Connect HNAS FC Port 1 and FC Port 3 to separate FED MPs on a USP V FED. C. Connect HNAS FC Port 1 from Node 1 and Node 2 to the same USP V Cluster. D. Connect HNAS FC Port 1 and FC Port 3 to the same FED MPs on a USP V FED. Answer: A, B Questions: 31 You have a 100% random read application. It is running on two RAID5 7D+1P parity groups with 146 GB drives @ 15K RPM and two 4 GB fiber ports to access your host. The application profile calls for a 50% read hit ratio. From a capacity perspective, what is the approximate maximum IO rate that you can achieve with this configuration? A. 1000 B. 3000 C. 6000 D. 10000 Answer: C Questions: 32 In the Universal Storage Platform (USP) family, what is the maximum number of external ports that one external LDEV can be attached to? A. 2 B. 4 C. 8 D. 16 Answer: C Questions: 33 How long does Storage Navigator Performance Monitor store the collected short range data of a Universal Storage Platform (USP) V with less then 64 CUs? A. Up to 15 days B. Up to 1 month


Page 9 of 28

C. Up to 2 months D. Up to 3 months Answer: A Questions: 34 When planning for an upgrade, which three Hitachi Tuning Manager sets of metrics should be collected prior to configuring and sizing storage systems? (Choose three.) A. SAN B. Cache C. Host bus adapter D. RAID Group E. Port Answer: B, D, E Questions: 35 During a storage assessment, a customer experiences that just 45% of the allocated storage is used by the file system of connected servers. Which Hitachi Dynamic Provisioning (HDP) functionality will allow the customer to optimize the usage of the allocated storage? A. Thin Provisioning B. VDEV Striping C. VDEV Provisioning D. Thin Striping Answer: A Questions: 36 Which statement applies to both Hitachi modular and enterprise storage families? A. Active-active load balancing is recommended. B. A high level of random writes benefit from selecting RAID1. C. A logical device should be accessed by only one processor at a time. D. A write pending problem can exist even when write pending is reporting as low as 20%. Answer: B Questions: 37 Which three types of data collected by Hitachi Tuning Manager Agents are stored in the Store Database? (Choose three.) A. Real time performance data B. Data which can be used to analyze changes and trends in system status over time C. Aggregated data at hour, daily, weekly, and monthly intervals D. Logs and messages from the system and applications E. System status of resources at specified times, such as detailed data of currently active

processes Answer: B, D, E Questions: 38 Which two events must occur in Hitachi Tuning Manager (HTnM) to start monitoring a newly provisioned LDEV? (Choose two.)


Page 10 of 28

A. HTnM refresh must be performed. B. Two HTnM polling cycles must be completed. C. The RAID agent must be stopped and started. D. The "monitored logical disks" list in HTnM must be updated with the new LDEVs. Answer: B, D Questions: 39 Which three actions provide a way of increasing the number of Tags available to a database volume in an enterprise storage system? (Choose three.) A. Build the volume across multiple LUNs with host-based logical volume manager striping. B. Construct the LUN using LUSE with LDEVS from different RAID Groups. C. Use a VDEV LUN from a VDEV striping pool (LDEV concatenation). D. Map the LUN across more host paths. E. Increase the HBA LUN Queue Depth value. Answer: A, D, E Questions: 40 Solaris IOSTAT shows that response times are degrading for a number of LUNs and that these LUNs are 100% busy. The active queue values are in a range up to, but not exceeding 8, and the wait queue values are in the range from 0 to 5 when the active queue is 7 or 8. Which condition is implied by these values? A. These are normal values for busy LUNs with a LUN queue depth of 8. B. There are no more Tags available on this storage port due to heavy IO load. C. There are no more Tags available for these LUNs at the port due to heavy IO load. D. The HBA LUN queue depth is set incorrectly. Answer: A Questions: 41 Your customer wants to implement Hitachi High Performance NAS (HNAS) within their company. You are in the planning phase for the Storage Pool layout and configuration. What are three critical considerations that you should verify with the customer before proceeding with the implementation? (Choose three.) A. Application IO behavior B. Fibre Channel properties/settings C. Data/file system structure D. User quotas E. Ethernet Channel properties/settings Answer: A, C, D Questions: 42 When using Hitachi Tiered Storage Manager, which two metrics are available as filtering options? (Choose two.) A. Cache % Busy B. Array Group % Busy C. Processor % Busy D. Max Array Group % Busy


Page 11 of 28

Answer: B, D Questions: 43 Which three key metrics are needed for sizing industry standard applications? (Choose three.) A. IO read/write ratios B. Port speed C. IO rates and data transfer rates D. Cache read hit ratios E. Number of ACP/BED pairs Answer: A, C, D Questions: 44 A customer wants to migrate an existing application which generates heavy random write IO to a new Universal Storage Platform (USP) V and is concerned about write delays and ease of management. The customer plans to use sixty-four 146 GB 15K RPM HDDs. Which back-end/RAID Group configuration should you setup? A. 2 x 4 concatenated RAID1 4+4 RAID Groups B. 2 x 4 concatenated RAID5 7+1 RAID Groups C. HDP Pool built out of 8 RAID1 4+4 RAID Groups D. HDP Pool built out of 8 RAID5 7+1 RAID Groups Answer: C Questions: 45 Which three components of an enterprise storage array can cause performance bottlenecks with a small block intensive random OLTP workload? (Choose three.) A. Port Fibre Channel DX interface B. Busy RAID Group C. LVM striped LDEVs D. Channel Processor E. Cache Write Pending level Answer: B, D, E Questions: 46 Which two reports are used to correlate CHA microprocessor utilization to port IO? (Choose two.) A. Tuning Manager Port IO B. Device Manager Port Configuration C. Tuning Manager RAID Configuration D. Tuning Manager CHA Microprocessor Utilization Answer: A, D Questions: 47 A large enterprise wants to implement a multi-tier environment as follows: Tier 1 on Internal - 15K RPM Tier 2 on Internal - 10K RPM Tier 3 on External - 10K RPM


Page 12 of 28

Tier 4 on External - 7.2K RPM They are primarily concerned about their OLTP environment, data warehouse, low utilization, and data archiving application. Which configuration is typical for this environment? A. OLTP on Tier 1, DW on Tier 3, Archiving on Tier 4 B. OLTP on Tier 2, DW on Tier 3, Archiving on Tier 4 C. OLTP on Tier 1, DW on Tier 2, Archiving on Tier 3 D. OLTP on Tier 1, DW on Tier 2, Archiving on Tier 4 Answer: D Questions: 48 Which observation, made on a modular storage system running batch jobs, clearly represents a problem? A. Utilization of both controllers is sometimes 80%. B. LUNs have activity from both controllers. C. Aggregate throughput is 300 MB/s. D. Aggregate throughput is 3000 IOPS. Answer: B Questions: 49 A customer seeks to double the throughput of a 3000 IOPS batch workload from an Oracle database dominated by 8 KB sequential reads. It is currently deployed on a port processor pair and a 7+1 RAID5 array group. Hitachi Dynamic Link Manager (HDLM) is used with the Extended Round-Robin path balancing. Port utilization is high and RAID Group utilization is low. The application is heavily IO bound. Which method would achieve this objective? A. Add more cache. B. Configure Oracle to use 64 KB physical reads. C. Use Round-Robin path balancing. D. Move from RAID5 to RAID1. Answer: B Questions: 50 An older configuration is achieving a peak of 8,900 IOPs with 112 (one hundred and twelve) 10K RPM drives. How many 15K RPM drives will be needed in a new storage system that should achieve 50% more peak IOPS (13,350 IOPS)? A. 84 B. 112 C. 140 D. 168 Answer: B Questions: 51 What are three initial metrics that should be considered when diagnosing a response time problem? (Choose three.)


Page 13 of 28

A. Application transfer rate B. Application IO rate C. Port processor utilization D. Cache and BED utilization E. Array group utilization Answer: C, D, E Questions: 52 Which two cases are good candidates for using a dedicated Cache Logical Partitioning (CLPR) on a Universal Storage Platform (USP)? (Choose two.) A. Parity groups associated with a service level response time objective B. Parity groups with LDEVs running OLTP workloads with low IOPS C. Parity groups with LDEVs running high write activity on external devices D. Parity groups with LDEVs running intensive video streaming Answer: A, C Questions: 53 What is the maximum recommended number of port processors servicing a LUN? A. 2 B. 4 C. 6 D. 8 Answer: B Questions: 54 Which three statements are true regarding available features in Hitachi Server Priority Manager? (Choose three.) A. Constrain IO rate by HBA WWN. B. Constrain read IO rate by HBA WWN. C. Constrain data transfer rate (MB/s) by HBA WWN. D. Constrain read hit ratio by HBA WWN. E. Constrain IOPS on a physical front-end Fibre Channel port. Answer: A, C, E Questions: 55 Which component in an enterprise back-end director (BED) assists performance with write and data recovery operations to a 7D+1P RAID5 array group? A. DRR RAID processors B. DKP/DKA processors C. Back-end loops D. Dedicated data paths from the BED to the cache switch Answer: A Questions: 56 A Universal Storage Platform (USP) has four external paths connected to a Lightning 9900V storage system and another four external paths connected to an active/passive modular array. All


Page 14 of 28

the external LUNs are accessed through partitions in the USP, and the external LUNs must deliver the highest possible levels of performance. Which Universal Volume Manager parameters you would select for the LUNs on the two external systems? A. "Cache Mode Enabled = On" for both systems, 9900V LUNs mapped to all four ports with

default path priority, and modular LUNs mapped to all four ports with path priority set manually to match controller ownership.

B. "Cache Mode Enabled = On" for the 9900V and "= Off" for the modular system, default path priority for both external systems.

C. "Cache Mode Enabled = On" for both systems and default path priority for both external systems.

D. "Cache Mode Enabled = On" for both systems, 9900V LUNs mapped in two sets to two pairs of ports with default path priority, and modular LUNs mapped to all four ports with path priority set manually to match controller ownership.

Answer: D Questions: 57 How many back-end disk operations are required to complete a small block random write operation to a RAID5 LUN where the operation does not overlap a RAID Chunk or Stripe or benefit from cache-write grouping? A. 1 B. 2 C. 3 D. 4 Answer: D Questions: 58 Which three operations are performed using Hitachi Tuning Manager Main Console as opposed to Performance Reporter? (Choose three.) A. Display data collected on hourly intervals. B. Display data collected on minute intervals. C. Display an overview of the system, perform comparative data analysis, and understand trends. D. Display customized reports. E. Display general data collected by Agents. Answer: A, C, E Questions: 59 Which three actions must be performed to have a host agent active in Hitachi Tuning Manager (HTnM)? (Choose three.) A. The agent service on the host server must be started successfully. B. A refresh of HTnM must occur. C. One HTnM polling cycle must be completed. D. The host agent should point to the hostname of the HTnM server. E. The host agent should be connected in HTnM. Answer: A, D, E Questions: 60


Page 15 of 28

In which two situations should Hitachi Tuning Manager Performance Reporter be used instead of the Main Console? (Choose two.) A. When real time monitoring is required. B. When you do not know where the problem is located. C. When you want to identify current top ten IO rate LDEVs. D. When correlation between LDEVs and drive letters (Windows) is required. Answer: A, C Questions: 61 Which two can cause high cache Write Pending on the Universal Storage Platform (USP) family? (Choose two.) A. High port utilization B. Attempting to write to a RAID Group at a rate beyond its ability C. Oversubscribed BEDs D. Inadequate WAN bandwidth for True Copy Asynchronous Answer: B, C Questions: 62 A customer wants to migrate an existing application which generates heavy random read IO to a new Universal Storage Platform (USP) V and is concerned about read delays and ease of management. The customer plans to use sixty-four 300 GB 15K RPM HDDs. Which back-end/RAID Group configuration should you setup? A. 2 x 4 concatenated RAID1 4+4 RAID Groups B. 2 x 4 concatenated RAID5 7+1 RAID Groups C. LUSE LDEV built out of 8 RAID5 7+1 RAID Groups D. HDP Pool built out of 8 RAID5 7+1 RAID Groups Answer: D Questions: 63 Which three Windows Performance Monitor counters should be included in diagnostic activities of potential disk-related issues for a Microsoft SQL Server database? (Choose three.) A. Average disk queue length B. Average disk read latency (ms) C. IO delays per sec D. SCSI command rejects/second E. Disk write bytes/second Answer: A, B, E Questions: 64 Which Solaris MPXIO multi-pathing scheme should be used to avoid non-owning access with HDS modular storage systems? A. Load_balance_none B. Load_balance_rr C. Load_balance_lba D. Load_balance_fixed


Page 16 of 28

Answer: A Questions: 65 Which three statements are true for Hitachi Global Link Availability Manager (HGLAM)? (Choose three.) A. HGLAM manages multiple instances of Hitachi Dynamic Link Manager (HDLM). B. HGLAM assigns load balancing algorithms only at the server level. C. HGLAM can alert which server is impacted when a fabric switch goes down. D. HGLAM reports IO rate at the HBA level. E. HGLAM can report a USP cache failure. Answer: A, C, D Questions: 66 Which two HDS modules should be used to extract information related to an application when planning systems for capacity growth and for performance? (Choose two.) A. StorageNav Performance Monitor B. Hitachi Tuning Manager (HTnM) Main Console C. Hitachi Storage Services Manager (HSSM) D. Hitachi Tuning Manager (HTnM) Performance Reporter Answer: B, D Questions: 67 You have a 100% random read application. It is running on two RAID5 7D+1P parity groups with 146 GB drives @ 15K RPM and two 4 GB fiber ports to access your host. The application profile calls for a 50% read hit ratio. From a capacity perspective, what is the approximate maximum IO rate that you can achieve with this configuration? A. 1000 B. 3000 C. 6000 D. 10000 Answer: C Questions: 68 Which three statements are true about Virtual Partition Manager? (Choose three.) A. There is no maximum or minimum to the number of SLPRS that can be created. B. A storage partition administrator only has authority within the assigned storage logical

partition(s). C. There is no maximum or minimum size for a SLPR. D. You can create up to 32 storage logical partitions in one storage system, including the default

SLPR 0. E. There is a maximum of 8 cache logical partitions in one storage subsystem, including the

default CLPR 0. Answer: B, C, D Questions: 69 Which three statements describe the capabilities provided with Hitachi Tuning Manager to assist with the analysis of Exchange Storage Group usage? (Choose three.)


Page 17 of 28

A. A dynamically created mapping that correlates Exchange Storage groups to the LUNs on the

HDS Storage array. B. Customized reports that focus on the RAID Groups, LDEVS, and ports that are specifically

used by Exchange Storage groups. C. Customized alarms that generate alerts with one minute granularity when Exchange users

are added to the system. D. Customized alarms that generate alerts with one minute granularity when RAID Groups that

contain Exchange Storage groups exceed pre-defined utilization thresholds. E. A dynamically created mapping that correlates LUNs to the physical disks defined on the

Exchange Server. Answer: B, D, E Questions: 70 The performance of a response critical application generating a majority of 4 KB IOs has progressively deteriorated on a new modular storage system as unplanned development and general file serving workloads have been added. The system administrator is convinced that this is due to falling cache hit rates. Which three actions will improve the cache hit rates of the response critical application? (Choose three.) A. Place the response critical application in its own cache partition. B. Upgrade the cache from 8 GB to 16 GB. C. Reduce the application cache partition segment size from 16 KB to 8 KB. D. Move the development and general file server workloads to another storage system. E. Move the development and general file server workloads to lower performance tier RAID

Groups. Answer: A, B, C Questions: 71 What is the maximum recommended utilization level range for each port microprocessor in a highly available, response critical, active-active pair servicing the workload? A. 15-25% B. 30-40% C. 45-55% D. 60-65% Answer: B Questions: 72 Which two workloads gain a performance improvement by using Tagged Command Queuing? (Choose two.) A. Multi-user transaction database access B. Sequential, single-threaded scan C. Archival writes to tape drive D. Corporate mail and notes server Answer: A, D Questions: 73 Which three statements are true regarding the JPCAlarm feature included in Hitachi Tuning Manager (HTnM)? (Choose three.)


Page 18 of 28

A. Alarms are triggered hourly. B. Alarms can be activated using Command Line Interface (CLI). C. A damping function is available to apply rules. D. Alarms can be triggered as often as every minute. E. Metrics used by JPCAlarm are retrieved from the HTnM HiRDB database. Answer: B, C, D Questions: 74 Click the Exhibit button. In the exhibit regarding modular storage cache, which condition does this graphic indicate?

A. Insufficient physical cache B. Owning controller conflict C. Sporadic intense sequential reads D. Sporadic intense sequential writes Answer: B Questions: 75 On an enterprise storage system, what describes how the cache is configured by Virtual Partition Manager (VPM)? A. Minimum CLPR size = 2 GB and increment size = 2 GB B. Minimum CLPR size = 4 GB and increment size = 2 GB C. Minimum CLPR size = 8 GB and increment size = 2 GB D. Minimum CLPR size = 8 GB and increment size = 4 GB


Page 19 of 28

Answer: B Questions: 76 Which two storage array components should you separate for the production and target backup LUNs in a host-based disk-to-disk (D2D) or a disk-to-disk-tape (D2D2T) backup scenario? (Choose two.) A. Front-end ports B. RAID Groups C. Cache partitions D. Back-end directors Answer: A, B Questions: 77 You are implementing a Hitachi High Performance NAS (HNAS) cluster system with a Universal Storage Platform (USP) V as the storage system. Which two factors for designing Storage Pools should be considered when the same performance attributes are required for both pools? (Choose two.) A. Balance LDEVs from different FEDs and BEDs for each Storage Pool. B. Use LDEVs from the same FEDs and the same BEDs for each Storage Pool. C. Use different LDEV sizes for each Storage Pool. D. Use the same LDEV size for each Storage Pool. Answer: A, D Questions: 78 Which two Hitachi Tuning Manager metrics are analyzed to determine whether a RAID Group is configured appropriately for an application? (Choose two.) A. Random Write Level B. Cache Write Pending C. CHA microprocessor utilization D. RAID Group utilization Answer: A, D Questions: 79 Your customer wants to implement Hitachi High Performance NAS (HNAS) within their company. The application creates mostly small random IO with a block size of 4 KB. Which HNAS setting will improve the IO read performance? A. Read ahead chunk size is set to 0 B. Read ahead chunk size is set to 4 KB C. Read ahead chunk size is set to 8 KB D. Read ahead chunk size is set to 16 KB Answer: A Questions: 80 Which Hitachi product combination should be used for collecting configuration and performance metrics prior to configuring and sizing storage systems for performance?


Page 20 of 28

A. Hitachi Device Manager and Hitachi Tiered Storage Manager B. Hitachi Tuning Manager and Hitachi Tiered Storage Manager C. Hitachi Device Manager and Hitachi Tuning Manager D. Hitachi Tuning Manager and Storage Services Manager Answer: C Questions: 81 A customer is experiencing SCSI timeouts with a new cluster of three servers each with four ports accessing 170 LUNs mapped to all four paths. The HBA LUN queue depth is set to 32 and you suspect that the storage port Tag count of 1024 is being exceeded. Which configuration will eliminate the risk of SCSI timeouts due to Tag exhaustion? A. Set the HBA LUN queue depth to 6. B. Divide the LUNs into two path pairs and set the LUN queue depth to 4. C. Divide the LUNs into two path pairs and set the LUN queue depth to 12. D. Set the HBA LUN queue depth to 12. Answer: B Questions: 82 You want to enable Hitachi Tiered Storage Manager to use performance metrics to control and manage migration tasks. Which prerequisite is required to do this? A. Set up Single Sign On for Hitachi Tiered Storage Manager. B. Set up Single Sign On for Hitachi Device Manager. C. Set up Single Sign On for Hitachi Tuning Manager. D. Set up Single Sign On for all HiCommand products. Answer: C Questions: 83 Which three types of data collected by Hitachi Tuning Manager Agents are stored in the Store Database? (Choose three.) A. Real time performance data B. Data which can be used to analyze changes and trends in system status over time C. Aggregated data at hour, daily, weekly, and monthly intervals D. Logs and messages from the system and applications E. System status of resources at specified times, such as detailed data of currently active

processes Answer: B, D, E Questions: 84 From a capacity planning perspective, what information should the customer collect to implement a data warehouse in a Very Large Data Base (VLDB) environment? (Choose three.) A. Number of records in the fact table at full deployment B. Record size in the fact table C. Number of registered users D. Number of table partitions in the parallel process E. Total log size Answer: A, B, D


Page 21 of 28

Questions: 85 Which two metrics indicate that a modular storage system is receiving write IOPS much faster than they can be destaged or written to disk, and that inflow control has occurred? (Choose two.) A. Cache Write Pending is at a high threshold limit. B. Port Write Hit Rate % is less than 100. C. The number of Write IOPS is higher than the number of Read IOPS. D. The Write MB/s throughput is greater than the Read MB/s throughput. Answer: A, B Questions: 86 Which three techniques can be used to ensure maximum write throughput to SATA LUNs in a fully populated HDS modular array? (Choose three.) A. Build the RAID Groups with HDD roaming across disk trays. B. Use the largest possible RAID Group size, e.g. 14D+2P. C. Configure many LUNs on each RAID Group and divide their access equally across both

controllers. D. Map all the LUNs on each SATA RAID Group through the same controller. E. Ensure that both controllers share the write workload. Answer: A, D, E Questions: 87 What is the maximum recommended utilization level range for a BED processor in a highly available, response critical, environment for OLTP transactions? A. 30-40% B. 50-60% C. 60-70% D. 80% or higher Answer: A Questions: 88 Which three factors should be considered when sizing for performance instead of capacity in a Microsoft Exchange Server environment? (Choose three.) A. Number of disks B. Mailbox quotas C. RAID type D. Size of the disks E. Disk utilization percentage Answer: A, C, E Questions: 89 Which feature optimizes the IO workload when using Hitachi Dynamic Provisioning (HDP) on a Universal Storage Platform (USP) V? A. Data is striped across the pool volumes in a fixed size that is optimized to achieve both

performance and storage area saving.


Page 22 of 28

B. Data is striped across the pool volumes in a variable size that is optimized to achieve both performance and storage area saving.

C. Data is striped across the pool volumes in a fixed size that is optimized to achieve storage area saving.

D. Data is striped across the pool volumes in a variable size that is optimized to achieve performance.

Answer: A Questions: 90 A large bank is completing a capacity planning exercise and wants to upgrade their current Universal Storage Platform (USP) Vs to properly support their online banking application for the next 24 months. The IT department wants to know in which specific component they should invest to achieve the transactional Service Level Objective (SLO). What should have priority? A. More cache and shared memory B. More and faster fiber channel ports C. More and faster disks D. More and faster fabric directors Answer: C Questions: 91 You are implementing a Hitachi High Performance NAS (HNAS) system ith a Universal Storage Platform (USP) V as the storage system. Which LDEV configuration would you reate in the Universal Storage Platform (USP) V? A. Mny LDEVs per RAID Group B. Oe LDEV per RAID Group C. Ay LDEVs per RAID Group and build LUSE volumes D. LDEVs smaller than 500 GB Answer: B Questions: 92 Which three characteristics describe an OLTP database? (Choose three.) A. R/W ratio: from 1:1 to 4:1 B. R/W ratio: from 4:1 to 40:1 C. IO type: random access D. IO type: sequential read E. Bock size: from 4 KB to 8 KB Answer: A, C, E Questions: 93 Your customer is concerned about "bad" storage performance and asks you how to resolve it. Which two steps should be performed first? (Choose two.) A. Measure cache hits and processor utilization. B. Clarify the customer's concerns and expectations. C. Establish comprehensive data collection. D. Determine LUN response times. Answer: B, C


Page 23 of 28

Questions: 94 In which two environments are shared resources recommended? (Choose two.) A. Resources shared among multiple servers running the same application B. Resources shared among intense ad hoc workloads and scheduled production workloads C. Resources shared among large numbers of small to medium sized applications D. Resources shared among workloads having conflicting optimization objectives, for example

interactive and batch Answer: A, C Questions: 95 A RAID5 array group in a Universal Storage Platform (USP) is over utilized. The IOPS IO profile is 50% 8 KB random reads, 30% 8 KB random writes, 10% 64 KB sequential reads, and 10% 64 KB sequential writes. Which two actions would likely resolve this problem? (Choose two.) A. Add more Universal Storage Platform (USP) cache. B. Redistribute the workload over a larger number of disk drives. C. Move the workload to similar disk resources on external storage. D. Migrate to RAID1. Answer: B, D Questions: 96 Which two statements apply to a many-to-many configuration of Hitachi Hitachi Server Priority Manager? (Choose two.) A. One threshold value cannot be established for the entire storage system. B. IO operations are specified for each port. C. One threshold value can be established for the entire storage system. D. Upper limit values are set for each HBA WWN. Answer: C, D Questions: 97 Which two workloads are appropriate for operation with RAID6 LUNs? (Choose two.) A. Transaction database access B. Extremely high available archive C. Low intensity streaming workload D. Tempdb and log volumes Answer: B, C Questions: 98 Which three parameters will help to calculate a server's maximum LUN queue depth to avoid the risk of SCSI timeouts? (Choose three.) A. The number of active LUNs mapped to a storage port. B. The number of active servers in the cluster accessing the active LUNs. C. The number of command tags supported by the storage port. D. The number of disk drives supporting the LUN configuration. E. The total number of LUNs mapped to the storage port.


Page 24 of 28

Answer: A, B, C Questions: 99 How many back-end disk operations are generated to the physical disks when a Random Write to a RAID6 LUN overlaps a RAID stripe? A. 4 B. 8 C. 12 D. 16 Answer: C Questions: 100 A television company wants to use an HDS Adaptable Modular Storage AMS1000 system to both store and stream audio visual data . Which two RAID Group configurations should you configure on the AMS1000? (Choose two.) A. RAID1+0 4+4 B. RAID1+0 8+8 C. RAID5 4+1 D. RAID5 8+1 Answer: C, D Questions: 101 An unidentified bottleneck in an Adaptable Modular Storage (AMS) system is causing a regular performance degradation some time between 9:00 a.m. and 10:00 a.m. every day. What information should you ask the user to collect to assist with the initial troubleshooting? (Choose two.) A. RAID Group/Logical Unit constitute file B. A series of 60 PFM Output Files with 1 minute interval C. A single PFM Output File for the hour using a 60 minute interval D. Simple traces repeated continuously during the 60 minute period Answer: A, B Questions: 102 On a fully configured Universal Storage Platform (USP) V, how many concurrent internal memory operations can be performed? A. 128 B. 256 C. 320 D. 512 Answer: C Questions: 103 A customer wants to set up TrueCopy Synchronous on Universal Storage Platform (USP) V storage systems. The systems are located in two different data centers on two different sites. In the SAN, the customer uses 4 GB long wave ISLs to connect the data centers. The customer wants to use the existing SAN infrastructure to establish the TrueCopy links. Which two should be


Page 25 of 28

verified before setting up the TrueCopy links to prevent performance issues caused by the SAN? (Choose wo.) A. ISL usage level B. Nmber of fabrics C. Nmber of shared SAN FC ports D. Nmber of ISLs Answer: A, D Questions: 104 What are three metrics collected by Hitachi Tuning Manager (HTnM) for modular storage systems? (Choose three.) A. IO rate (IOPS) B. HDD to parity group correlation C. Data transfer rate (MB/s) D. IO Usage (%) at the RAID Group level E. Response time (ms) Answer: A, C, E Questions: 105 Which three offerings are included in the Storage Platform Assessment service (SPA)? (Choose three.) A. Storage Assessment B. Storage Scorecard C. Storage Design D. Storage Deployment Consulting E. SAN Assessment Answer: A, B, D Questions: 106 Which performance metric is used by Hitachi Tiered Storage Manager to manage migration tasks? A. Cache Write Pending B. MP utilization C. LDEV response time D. Array Group Usage Answer: D Questions: 107 Which AIX host tool will provide the most comprehensive disk metrics from a host perspective? A. Topas B. Nmon C. Perfmon D. Hdskmon Answer: B


Page 26 of 28

Questions: 108 Which two statements describe the behaviors of the Universal Storage Platform (USP) cache algorithms with an external device when "cache mode = disable" under normal mixed workload? (Choose two.) A. Caching algorithms are used for all read IO operations. B. Caching algorithms are used for all IO operations. C. Caching algorithms are not used for write IO operations. D. Only the "sequential detect" algorithm is used for write IO operations. Answer: A, C Questions: 109 A Universal Storage Platform (USP) is experiencing high Cache Write Pending. TrueCopy is not involved and the level of BED utilization is not high. What are the first two steps in resolving this issue? (Choose two.) A. Add more cache. B. Identify the array groups or LDEVs involved. C. Identify the workload causing the issue. D. Add more shared memory. Answer: B, C Questions: 110 Which two disk drive attributes allow you to calculate the number of IOPS that a disk drive will be able to achieve? (Choose two.) A. Rotational latency B. Seek time C. Number of sectors per track D. Drive loop interface speed Answer: A, B Questions: 111 In which two software products would you find array group utilization metrics? (Choose two.) A. Hitachi Tuning Manager B. Storage Navigator Performance Monitor C. Hitachi Dynamic Link Manager D. Hitachi Dynamic Provisioning Answer: A, B Questions: 112 Which three statements accurately describe the capabilities that Hitachi Tuning Manager provides? (Choose three.) A. Monitors storage capacity and performance metrics. B. Dynamically migrates RAID Groups. C. Maps and reports on storage resources with a focus on applications, servers, and storage. D. Maps and reports on HBA performance. E. Provides configurable alerts to provide notification when performance or capacity thresholds

are exceeded.


Page 27 of 28

Answer: A, C, E Questions: 113 Which property should you consider when analyzing long range data using Universal Storage Platform (USP) V Storage Navigator Performance Monitor? A. Fixed 15-minute interval for 3 months B. Fixed 15-minute interval for 2 months C. Fixed 1-minute interval for 2 months D. Fixed 5-minutes interval for 2 months Answer: A Questions: 114 A large customer wants to have three copies of a 15 TB database created with ShadowImage during a three week cycle (60 TB total capacity). They want to have fast disk-based recovery option from any of the preceding three weeks but without the expense of using Tier1 or Tier2 storage. Some performance impact is acceptable during the recovery as disk-based recovery is much faster than the previous tape-based scheme. What is the appropriate solution? A. Install 45 TB of 10K RPM drives in the Universal Storage Platform (USP). B. Install 45 TB of the largest available capacity 15K RPM drives in the Universal Storage

Platform (USP). C. Install 45 TB of Fibre Channel drives in external Adaptable Modular Storage (AMS) systems. D. Install 45 TB of SATA drives in external Workgroup Modular Storage (WMS) systems. Answer: C Questions: 115 Which three parameters indicate that LUN ownership access or changes may be causing a performance problem in an HDS modular storage system? (Choose three.) A. The total cache queue drops below 100% for any 1 minute measurement period. B. The cache free queue drops below 100% for any 1 minute measurement period. C. IOPS are recorded to the same LUN from different controllers over several measurement

periods. D. The LUN C-CTL value does not match the D-CTL value. E. The Cache Write Pending level is at a high threshold level. Answer: A, C, D Questions: 116 Which level of Write Pending is normal for a busy OLTP Universal Storage Platform (USP) system adequately managing the destaging of writes? A. 0% to 20% B. 20% to 30% C. 30% to 40% D. 40% to 50% Answer: B


Page 28 of 28

Questions: 117 The back-end processor utilization in an existing Universal Storage Platform (USP) series system running OLTP is occasionally exceeding 50%. You want to add 50% more workload to the same without experiencing any further increase in latency. There are no free slots for BED expansion but plenty of room for disk and FED expansion. Which two solutions will solve this problem? (Choose two.) A. Add FEDs and employ external Fibre Channel storage to meet the additional IOPS demand. B. Install more disks and migrate from RAID5 to RAID1 C. Add FEDs and map the LUNs across more processors. D. Increase Shared Memory to allow more concurrent operations. Answer: A, B

End of Document

HH0-440

Documents

b questions

d questions

c questions

total questions

e questions

hitachi exam code

storage replication

external storage