emc enterprise storage · emc enterprise storage emc corporation 171 south street, hopkinton, ma...

EMC Enterprise Storage

EMC Corporation 171 South Street, Hopkinton, MA 01748-9103Corporate Headquarters: (508) 435-1000, (800) 424-EMC2 Fax: (508) 435-5374 Service: (800) SVC-4EMC

EMC NavisphereAnalyzerVersion 5.X

ADMINISTRATOR’S GUIDEP/N 069001043-00

EMC Navisphere Analyzer Version 5.X Administrator’s Guideii

Copyright © EMC Corporation 2000. All rights reserved.Printed November 2000No part of this publication may be reproduced or distributed in any form or by any means, or stored in a database or retrieval system, without the prior written consent of EMC Corporation.

The information contained in this document is subject to change without notice. EMC Corporation assumes no responsibility for any errors that may appear.

All computer software programs, including but not limited to microcode, described in this document are furnished under a license, and may be used or copied only in accordance with the terms of such license.EMC either owns or has the right to license the computer software programs described in this document. EMC Corporation retains all rights, title and interest in the computer software programs.

EMC Corporation makes no warranties, expressed or implied, by operation of law or otherwise, relating to this document, the products or the computer software programs described herein. EMC CORPORATION DISCLAIMS ALL IMPLIED WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE. In no event shall EMC Corporation be liable for (a) incidental, indirect, special, or consequential damages or (b) any damages whatsoever resulting from the loss of use, data or profits, arising out of this document, even if advised of the possibility of such damages.

Trademark InformationEMC2, EMC, MOSAIC:2000, Symmetrix, and Navisphere are registered trademarks and EMC Enterprise Storage, The Enterprise Storage Company, The EMC Effect, Connectrix, EDM, SDMS, SRDF, Timefinder, PowerPath, InfoMover, FarPoint, EMC Enterprise Storage Network, EMC Enterprise Storage Specialist, EMC Storage Logic, Universal Data Tone, E-Infostructure, Celerra, SnapView, MirrorView, and EMC CLARiiON are trademarks of EMC Corporation.

All other trademarks used herein are the property of their respective owners.

EMC Navisphere Analyzer Version 5.X Administrator’s Guide iii

Preface..............................................................................................................................ix

Chapter 1 About the EMC Navisphere AnalyzerNavisphere Environments...............................................................1-2Analyzer Architecture ......................................................................1-4Analyzer Dependencies ...................................................................1-5

Chapter 2 Installing and Starting AnalyzerRemoving Analyzer..........................................................................2-2Installing Analyzer ...........................................................................2-3Starting a Navisphere Manager or Analyzer Session..................2-5Setting User Options for Navisphere.............................................2-8Selecting a Storage System to Manage (Manager Not Installed) ....2-10Selecting Storage Systems to Manage (Manager Installed) ......2-12

What Next? ...............................................................................2-15

Chapter 3 Storage System Trees and the Main WindowStorage-System Trees........................................................................3-2Components of Trees ........................................................................3-6

Accessible, Inaccessible, and Unsupported Storage Systems........................................................................................3-6Icons.............................................................................................3-7

Main Window..................................................................................3-20Application Icon.......................................................................3-20Menu Bar...................................................................................3-21Toolbar.......................................................................................3-23Workspace.................................................................................3-24

Contents

EMC Navisphere Analyzer Version 5.X Administrator’s Guideiv

Contents

Status Bar ..................................................................................3-27Window Configuration ..........................................................3-27

Chapter 4 Analyzing PerformanceAnalyzer Menu Bar and Right-Click Menu Options...................4-2Analyzer Modes................................................................................4-3Zooming and Resizing Charts ........................................................4-6Analyzer Charts ................................................................................4-8Performance Summary Chart .........................................................4-9

Performance Summary Device Menu ..................................4-11Performance Summary Chart Menu.....................................4-11Performance Summary Chart Scroll Bar ..............................4-12

Performance Detail Chart ..............................................................4-13Performance Detail Device Menu .........................................4-15Performance Detail Chart Menu ...........................................4-16Performance Detail Chart Data Popup ................................4-17

IO Size Distribution Charts (LUN Only).....................................4-20IO Size Distribution Summary Menu...................................4-21IO Size Distribution Detail Chart Menu...............................4-23

Performance Properties..................................................................4-24Exporting and Printing Charts .....................................................4-30Using the Performance Logger .....................................................4-32

Setting up the Performance Logger ......................................4-32Setting Log File Properties .....................................................4-34Setting Log File Sizes ..............................................................4-37Scheduling Logger Operation (Advanced Tab)..................4-39Starting and Stopping Logging .............................................4-41Using the Log File Size Estimator .........................................4-42

Chapter 5 Understanding Performance DataUnderstanding Workload................................................................5-2

Storage Systems Under Light Loads.......................................5-4Storage Systems With Bottlenecks ..........................................5-4

Identifying a Performance Problem...............................................5-5Eliminating Bottlenecks ...................................................................5-7

Disk-Based Factors - Raid Type and Stripe Element Size....5-8Working With the Storage-System Cache .....................................5-9

Using Analyzer to Monitor Cache Performance.................5-10Addressing Server Problems.........................................................5-14

Server Throttles........................................................................5-14Data Alignment .......................................................................5-14

EMC Navisphere Analyzer Version 5.X Administrator’s Guide v

1-1 Sample Analyzer Environment .................................................................. 1-31-2 Architectural Components of an Analyzer Environment ...................... 1-43-1 Sample Partially Expanded Equipment Tree ........................................... 3-33-2 Sample Partially Expanded Storage Tree .................................................. 3-43-3 Sample Partially Expanded Hosts Tree ..................................................... 3-54-1 Display Before Zoom ................................................................................... 4-64-2 Chart With Zoom Area Selected (Left Mouse Button) ............................ 4-74-3 Chart After Zoom (Release of Left Mouse Button) .................................. 4-74-4 Sample Performance Summary Chart ..................................................... 4-104-5 Sample Performance Detail Chart Showing Utilization of SP ............. 4-154-6 Sample Performance Detail Chart Data Popup ..................................... 4-174-7 Sample IO Size Distribution Summary Chart ........................................ 4-214-8 Sample IO Size Distribution Detail Chart ............................................... 4-224-9 Performance Logger Agents Dialog Box ................................................. 4-335-1 Performance Analysis Flowchart ............................................................... 5-6

Figures

vi EMC Navisphere Analyzer Version 5.X Administrator’s Guide

Figures

EMC Navisphere Analyzer Version 5.X Administrator’s Guide vii

2-1 Default Settings for User Options .............................................................. 2-83-1 Accessible, Inaccessible, and Unsupported Storage Systems ................ 3-63-2 Icon colors ...................................................................................................... 3-73-3 Icon Images and Descriptions for Servers and Server HBAs ................. 3-83-4 Menu Option on the Menu for Single or Multiple Servers .................... 3-83-5 Storage-System Icon Images ....................................................................... 3-93-6 Multiple Storage-System Icon Image ....................................................... 3-103-7 Storage-System Menu Options for a Single Storage System.................. 3-123-8 Options on the Storage-System Menus for Multiple Storage System . 3-133-9 Basic Storage Component Icons ............................................................... 3-143-10 MirrorView and SnapView Storage Component Icons ....................... 3-173-11 Application Icon Image ............................................................................. 3-213-12 Toolbar Buttons ........................................................................................... 3-243-13 Filters for Displaying Managed Storage Systems .................................. 3-264-1 Analyzer Menu Options .............................................................................. 4-24-2 Analyzer Main Window Right-click Options for SPs, LUNs, and

Disks ................................................................................................................ 4-24-3 Performance Summary Device Menu Options ...................................... 4-114-4 Performance Summary Chart Options .................................................... 4-124-5 Performance Detail Device Menu Options ............................................. 4-164-6 Performance Detail Chart Options .......................................................... 4-164-7 IO Size Distribution Size Summary Chart Menu Options ................... 4-224-8 IO Size Distribution Detail Chart Options ............................................. 4-234-9 Performance Properties ............................................................................. 4-25

Tables

viii EMC Navisphere Analyzer Version 5.X Administrator’s Guide

Tables

EMC Navisphere Analyzer Version 5.X Administrator’s Guide ix

Preface

This manual describes installing and using the EMC Navisphere Analyzer disk-array performance monitor, which this manual calls Analyzer.

Audience and ObjectivesThis manual is for people who want to monitor and tune disk-array storage systems. You should be familiar with the operating system running on the servers and with the concepts and operation of the storage systems. After reading this manual, you will be able to install and use Analyzer.

This manual assumes that the disk-array storage systems have been connected to servers, and that each server is using the storage-system disks.

Organization of the Manual Chapter 1 Introduces Navisphere Analyzer.

Chapter 2 Explains installing and starting Analyzer.

Chapter 3 Explains using the Navisphere Main Window, which lets you select equipment, server, and storage display trees.

Chapter 4 Explains using Analyzer.

Chapter 5 Explains interpreting Analyzer performance data.

x EMC Navisphere Analyzer Version 5.X Administrator’s Guide

Preface

Format Conventions We use the following format conventions in this manual:

ConventionMeaning

this type Indicates text that you type verbatim. All commands, pathnames, and names of files and directories also appear in this typeface.

this type Represents variables for which you supply values; for example, the name of a directory or file and explicit arguments to commands.

x -> y Represents a menu path. For example, Views → Summary means select the Views item from the menu that appears, and then select the Summary item from the next menu that appears.

↵ Represents the Enter key. (On some keyboards this key is called Return or New Line.)

About the EMC Navisphere Analyzer 1-1

1

The EMC Navisphere® Analyzer performance monitor program (called Analyzer) measures the performance of disk-array storage systems.

This chapter describes the following topics:

• Navisphere Environments ................................................................1-2• Analyzer Architecture .......................................................................1-4• Architectural Components of an Analyzer Environment ............1-4

About the EMCNavisphere Analyzer

1

1-2 EMC Navisphere Analyzer Version 5.X Administrator’s Guide

About the EMC Navisphere Analyzer

Navisphere EnvironmentsAnalyzer runs on a Microsoft Windows host called a management station. A management station can manage storage systems connected to it directly and storage systems on servers connected to it by LAN (local area network). The management station does not need to run the same operating system as a host whose storage systems it manages.

Using Analyzer, you can monitor the performance of storage system components — storage processors (SPs), LUNs, and component disk modules — on different hosts.

Analyzer has two parts: the GUI, which interprets and displays performance information from storage systems or log files, and the Performance Logger, which creates log files. The Analyzer GUI can communicate in real time with the Agent running on each server. The Performance Logger also communicates with each server Agent, but instead of giving data to the Analyzer, it writes the data to a log file stored on the management station.

Analyzer runs in one of two modes: real-time mode, where it takes real-time data directly from Host Agents; and playback mode, where it takes data from log files.

The sample environment that follows shows a Windows host that is a management station for storage systems connected directly to their servers. The management station shown is not a storage-system server, but it could be. You can use this host to manage the storage systems on servers running Navisphere Agent (the Windows server and three UNIX® servers shown).

1

Navisphere Environments 1-3


Figure 1-1 Sample Analyzer Environment

Windows NT, Manager,

optional CLI,optional Organizeroptional Analyzer

Windows management station

Windowsserver

Unixserver

Managed sharedstorage system

GraphicalUser

Interface

LAN



Windowsserver

Unixserver

SwitchSwitch

Windows,Agent, andoptional CLI

Unix, Agent,and optional CLI



1



Analyzer ArchitectureAnalyzer can communicate in real time with the Agent running on any servers on the network. Analyzer also lets you start and stop the Performance Logger, a companion application that creates log files of server performance data and stores them on the management station. Analyzer can interpret and display these log files as well. TCP/IP must be loaded and configured on both host and client.

In a shared storage-system management environment, the Agent on a server communicates with Application Transparent Failover (ATF) software on the server. ATF in turn communicates with the Core Software running in a storage system’s storage processors (SPs). All shared storage systems have Fibre Channel server interfaces. The Agent uses a SCSI protocol over a Fibre Channel Fabric to communicate with the Core Software.

With FC4700 storage systems, the management station uses a separate network connection to perform management functions. There are two kinds of Agents, the Host Agent running on the host and an SP Agent running in each SP.

Figure 1-2 Architectural Components of an Analyzer Environment

NavisphereManager,Analyzer

Server

TCP/IP network (LAN)

Managementstation

Storage systemwith Fiber Channel

Interface

NavisphereHost Agent

Navisphere ATF

FC connection

Networkmanagementconnection

Core softwarewith Access Logix

Navisphere SPAgent

1

Analyzer Dependencies 1-5


Analyzer DependenciesAnalyzer requires a 32-bit Microsoft Windows environment on the management station.

The management station and any remote Navisphere Agents require a LAN with TCP/IP loaded and configured on both host and client.

What Next? To install the Analyzer files and start Analyzer, continue to Chapter 2.

1



Installing and Starting Analyzer 2-1

2

This chapter describes the following topics:

• Removing Analyzer...........................................................................2-2• Installing Analyzer ............................................................................2-3• Starting a Navisphere Manager or Analyzer Session ...................2-5• Setting User Options for Navisphere ..............................................2-8• Selecting a Storage System to Manage (Manager Not Installed) .....

2-10• Selecting Storage Systems to Manage (Manager Installed) .......2-12

This manual assumes that you are familiar with the Windows environment for your management station. For the latest information on which hosts you can use and the required software revisions and service packs, refer to the Analyzer Release Notice.

If you have an existing version of Analyzer, you must remove it before installing the new version. To remove Analyzer, continue to the nexct section. To install Analyzer, go to page 2-3.

Installing and StartingAnalyzer

2


Installing and Starting Analyzer

Before you can install Analyzer, you must remove any existing revision of Analyzer. To remove an existing version, continue. If there is no existing version, skip to the next page.

Removing Analyzer1. If you are not logged in, log in to the Windows management

station as Administrator or someone who has administrative privileges.

2. If Analyzer is running, stop it by following the menu path

File → Exit

3. From the Windows taskbar, follow the path

Start → Settings → Control Panel

4. In the Control Panel dialog box, double-click Add/Remove Programs.

5. In the Add/Remove Program Properties dialog box, click Navisphere Analyzer, and then click Add/Remove.

6. In the confirmation dialog box that opens, click Yes to confirm the removal of Analyzer.

7. In any Remove Shared File dialog box that appears, click Yes To All only if you are either removing Analyzer to install a different Analyzer version or if you are removing all Navisphere software from the host; otherwise, click No to All.

8. In the Remove Programs From Your Computer dialog box, click OK.

9. In the Add/Remove Program Properties dialog box, click OK to close the dialog box.

10. Close the Control Panel dialog box.

2

Installing Analyzer 2-3


Installing AnalyzerThe host on which you install Analyzer must have the following hardware and software:

• Color graphics console with a minimum resolution of 1024 x 768 pixels

• Windows 2000 or Windows NT 4.0 operating system with Service Pack 4 or Service Pack 5

• TCP/IP Services configured with connections to the servers with storage systems that Navisphere will manage

To Install Analyzer1. Log in to the Windows management station as Administrator or

someone who has administrative privileges.

2. If a version of Analyzer is already installed, remove it as described on the previous page before installing the new version.

3. Insert the Analyzer CD-ROM in the management station’s CD-ROM drive.

The installation of Analyzer starts automatically, and the Analyzer Install splash screen and InstallShield Wizard dialog box appear. If you do not see this splash screen and dialog box, follow these steps to start the installation.

a. From the Windows taskbar, follow the path

Start → Run

b. Enter the following program name, and then click OK:

drive: \setup.exe

where drive is the letter for the CD-ROM drive.

When the InstallShield preparation is complete, the Navisphere Analyzer Setup window opens.

4. In the Navisphere Analyzer Setup window, click Next.

The License Agreement dialog box opens in the Navisphere Analyzer Setup window.

5. Read the license agreement and then click Yes to accept the agreement and continue the installation.

2



6. In the Choose Destination Location dialog box, click Next to select the default location drive:\Program Files\EMC\Navisphere version

where drive is the letter of the drive where Windows is installed and version is the Navisphere version; for example, 5.0.

7. In the Select Program folder dialog box, click Next to select the default program folder.

The setup program copies Navisphere files to the destination location.

8. In the Setup Complete dialog box, click Finish. (Do not check the Launch Analyzer box at this time.)

Any user who can access the management station can change or delete the Analyzer files you just installed. If you want to change the permissions on these files, refer to the Release Notice for Analyzer for information on the recommended permissions.

What Next? Continue to the “Starting” section.

2

Starting a Navisphere Manager or Analyzer Session 2-5


Starting a Navisphere Manager or Analyzer SessionBefore you use Analyzer, someone must have performed the following tasks:

You can start either a Navisphere Enterprise session, which lets you use all Navisphere applications installed on the management station in the same Main window; or you can start an Analyzer session, which lets you use just Analyzer in the Main window.

Any user can run an Enterprise session from any management station on which Manager or Analyzer is installed, and can use it to monitor storage systems. Only authorized users can use Navisphere to configure or reconfigure a storage system. A user is authorized if the Agent on the server is set up with configuration access for that user, as described in the Server Software manual.

CAUTION!The Agent allows more than one Navisphere Manager session to access the same storage system at the same time. As a result, two authorized users are able to configure or reconfigure the same storage system at the same time, but doing this may damage the data.

To Start a Navisphere Enterprise or Analyzer Session

Before starting a session, make sure that the Agent is running on all servers connected to storage systems whose performance you want to monitor. Do not disable automatic polling for any storage system whose performance you want to monitor. Automatic polling must be enabled (default) to provide fresh data to the Analyzer.

Task Described in

Installed the storage systems and connected them to the servers, either directly or through hubs or switches.

Storage-system installation and service manual and hub or switch documentation.

Set up the servers whose storage systems you want to manage. Setting up includes installing ATF (if purchased) and the Agent and configuring the Agent.,

Server software manual for the storage system, ATF manual, or Agent/CLI manual.

2



1. Log on the Windows management station as either Administrator or someone who has administrative privileges.

2. From the Windows taskbar, follow the path below for the session you want to start.

From the Windows taskbar, follow the path below to start just Analyzer or all installed Navisphere management applications if other Navisphere applications are installed:

AnalyzerStart → Programs → Navisphere version → Navisphere Analyzer

All management applicationsStart → Programs → Navisphere version → Navisphere Enterprise

3. Click anywhere in the Navisphere splash screen or wait for the screen to close automatically in 3 seconds.

The Main window opens.

What happens next depends on whether the application you started can access single or multiple storage systems at one time. If Manager is installed, then the application can access multiple storage systems at a time. If Manager is not installed, the application can access only one storage system at a time.

Multiple storage-systems application (Manager installed)

The application first looks for the file containing the list of hosts that were managed when you closed your last session of this application. (The default file for this list is drive:\install-directory\Profiles\username\HostAdmin.txt,where drive is the drive where Windows is installed, install-directory is the directory where the Windows operating system was installed, and username is your username.)

If this host file exists, the application tries to

• Extract hostnames from the file.

• Contact each server in the file to determine the state of the storage systems connected to it.

• For each storage system it finds, display a storage-system icon in the Equipment and Storage trees in each open Enterprise Storage dialog box in its Main window.

2

Starting a Navisphere Manager or Analyzer Session 2-7


• For each server connected to a storage system that it finds, display a host icon in the Hosts tree in each open Enterprise Storage dialog box in its Main window.

If the host file does not exist, the Enterprise Storage dialog box remains empty.

If you start the Navisphere application while the Agent is starting up on a server with managed storage systems, the application may receive a time-out error from that Agent. If such a time-out occurs, the application displays a dialog box informing you of the time-out. Once the Agent is running, you can either restart the application or add the server to the list of managed hosts using the Agent Selection dialog box.

Single storage-system application (Manager not installed)

The application first looks for the file containing the name of the server and the storage system connected to it that were managed when you closed your last session. (The default file for this list is drive:\install_directory\Profiles\username\HostAdmin.txt where drive is the drive where Windows is installed, install-directory is the directory where Windows operating system was installed, and username is your username.)

If this host file exists, the application tries to

• Extract hostnames from the file.

• Contact each server in the file to determine the state of the storage systems connected to it.

• For each storage system it finds, display a storage-system icon in the Equipment and Storage trees in each open Enterprise Storage dialog box in its Main window.

• For each server connected to a storage system that it finds, display a host icon in the Hosts tree in each open Enterprise Storage dialog box in its Main window.

If the host file does not exist, the Enterprise Storage dialog box remains empty.

What Next? Continue to the next section, “Setting User Options for Navisphere.”

2



Setting User Options for Navisphere Navisphere has the following options that you can set to determine some of its basic operations.

Network Time Out - Sets the time interval in seconds for the application to establish a connection to a managed storage system. If the time interval is exceeded, the application terminates the connection.

Host File Path - Sets the path of the file where the list of Agents with managed storage systems is saved between Navisphere sessions.

Polling Interval - Sets the time interval in seconds that determines how often Manager performs a poll operation if automatic polling for the session (background polling) is enabled.

Automatic Polling - Enables or disables background polling for the session. Background polling maintains the polling interval counter and performs poll operations. When background polling is enabled, the application automatically requests an Agent to poll a managed storage system for updated information only if automatic polling for the storage system is enabled.

If automatic polling is enabled for a storage system, the polling interval and the automatic polling priority for a storage system determine how often the Agents poll storage systems. If automatic polling is not enabled for a storage system when a poll operation occurs, the information in the application’s image of the storage system remains unchanged. When background polling is disabled, no poll operations occur.

Table 2-1 Default Settings for User Options

Option Setting

Network Timeout 120 seconds

Host File Path drive:\install_directory\Profiles\username\HostAdmin.txt where drive is the drive and install_directory is the directory where you installed Windows, and username is your user name.

Polling Interval 60 seconds

Automatic Polling Cleared (disabled)

2

Setting User Options for Navisphere 2-9


To Set the User Options for the Application1. In the Main window, follow the path View → Options.

A User Options dialog box opens, similar to the following.

2. Change any of the options that you want to change:

a. In Host File Path, type or select the path to use for the managed host file.

b. In Polling Interval, type or select the number of seconds for the polling interval.

c. Select the Automatic Polling check box to enable automatic polling for the application, or clear the check box to disable it for all managed storage systems.

To set storage-system properties, you must have Navisphere Manager installed.

When automatic polling is enabled for the session, an individual storage system is polled only if automatic polling is enabled for the storage system. You enable automatic polling for an individual storage system using the Manager Storage System Properties dialog box General tab).

3. Click OK to apply the settings and close the dialog box.

All settings are saved for your future Navisphere sessions.

What Next? What you do next depends on whether Manager is installed.

Manager not installed - Continue to the next section, “Selecting a Storage System to Manage (Manager Not Installed).”

Manager installed - Go to the section, “Selecting Storage Systems to Manage (Manager Installed).”

2



Selecting a Storage System to Manage (Manager Not Installed)You select the Agents and storage systems you want to manage using the Agent Selection dialog box. If the Agents and storage systems you want to manage were selected when you exited the previous session, you do not need to select them again.

To Open the Agent Selection Dialog BoxOn the Main window menu bar, follow the path File → Select Hosts.

An Agent Selection dialog box opens. For information on the fields in the dialog box, click Help.

If the server with the storage system you want to manage is listed in the Host field and the storage system is in the Storage System field, then click OK. If you want to learn about storage system trees or the Navisphere Main window, skip to Chapter 3. To start using Analyzer, skip to Chapter 4. If you need to select a server or storage system, continue with the procedure that follows.

2

Selecting a Storage System to Manage (Manager Not Installed) 2-11


To Manage a Storage System on a Selected Server

1. In the Select Host dialog box in Host, type the name or the IP address of the server connected to the storage system you want to manage.

2. Click Contact.

The application tries to contact the server. If the application contacts the server and the server is running the Agent, it finds the storage systems on the server and lists them in the Storage System list.

3. In the Storage System list, click the storage system you want to manage.

4. Click OK.

The application displays an icon for the managed storage system in the Equipment and Storage trees in the Enterprise Storage dialog box. A host icon for the server appears in the Hosts tree in the Enterprise Storage dialog box.

What Next? To learn about storage-system trees or the Navisphere Main window, go to Chapter 3. To start using Analyzer, skip to Chapter 4.

2



Selecting Storage Systems to Manage (Manager Installed)You select the Agents and storage systems you want to manage using the Agent Selection dialog box. If the Agents and storage systems you want to manage were selected during the previous session, you do not need to select them again.

To Open the Agent Selection Dialog Box

On the Main window menu bar, follow the path File → Select Agents.

A Host Selection dialog box opens. For information on the fields in the dialog box, click Help.

If all the servers with storage systems that you want to manage are listed under Managed Agents, then click OK. If you want to learn about storage system trees or the Navisphere Main window, skip to Chapter 3. To run Analyzer, skip to Chapter 4.

2

Selecting Storage Systems to Manage (Manager Installed) 2-13


If all servers are not listed or the list is empty, you can either manage storage systems on servers whose name you know (next section) or manage storage systems on selected servers on subnets (2-13).

To Manage Storage Systems on Servers Whose Names You Know 1. For each server with storage systems you want to manage, type

its name in the Agent to Add box, and click →.

The server’s name appears in the Managed Agents box.

2. When the Managed Agents box contains the names of all the servers with storage systems you want to manage, click OK.

The dialog box closes, and the application does the following:

• Adds the new server names to the host file.

• Contacts each server in the file to determine the state of the storage systems connected to it.

• For each storage system it finds, displays a storage-system icon in the Equipment and Storage trees in each open Enterprise Storage dialog box.

• For each server connected to a storage system that it finds, displays a host icon in the Hosts tree in each open Enterprise Storage dialog box in its Main window.

The default host file is drive:\\install_directory\Profiles\username\HostAdmin.txt where drive is the disk and install_directory is the directory where the Windows operating system was installed and username is your username. To change the name and location of this file, follow the Main window menu path View → Options.

3. If you do not want to manage any storage systems using an icon on the Equipment and Storage trees, right-click the icon and click Unmanage.

Manager removes the storage-system icon from the Equipment and Storage trees in each open Enterprise Storage dialog box.

To Manage Storage Systems on Selected Servers on Subnets 1. For each subnet, type its IP address in the Subnet to Add box,

and click →.

The IP address moves into the Subnets to Search box.

2. When the Subnets to Search box contains the IP addresses of all the desired subnets, click Find Agents.

2



The application starts searching the subnets for servers running the Agent. When it finds a server, it displays an icon for the server in the Unmanaged Agents box. The Scanning subnets status bar displays the amount of the search completed.

3. When the search is complete, in the Unmanaged Host box, select the servers that you want to manage, and click →.

The server names move into the Managed Agents box.

The application displays host icons for each server connected to a managed storage system regardless of whether the server is managed. The icon for an unmanaged server functions differently than the icon for a managed server. As a result, we strongly recommend that you manage all servers connected to any storage system that you want to manage.

4. When the Managed Agents box contains all the desired servers, click OK.

The dialog box closes, and the application does the following:

2

Selecting Storage Systems to Manage (Manager Installed) 2-15


• Adds the new server names to the host file.

• Contacts each server in the file to determine the state of the storage systems connected to it.

• For each storage system that it finds, displays a storage-system icon in the Equipment and Storage trees in each open Enterprise Storage dialog box.

• For each server connected to a storage system that it finds, displays a host icon in the Hosts tree in each open Enterprise Storage dialog box in its Main window. (Only names of managed servers are in the host file.)

The default host file is drive:\\install_directory\Profiles\username\HostAdmin.txt where drive is the disk and install_directory is the directory where the Windows operating system was installed and username is your username. To change the name and location of this file, follow the Main window menu path View → Options.

5. If you do not want to manage any storage system with an icon on the Equipment and Storage trees, right-click the icon and click Unmanage.

The application removes the storage-system icon from the Equipment and Storage trees in each open Enterprise Storage dialog box.

What Next? For a description of the Main window, continue to Chapter 3. If you are familiar with the Main window, skip to Chapter 4.

2



Storage System Trees and the Main Window 3-1

3

Navisphere uses tree structures, similar to Windows Explorer folders, to show the physical and logical storage-system components and their relationships.

This chapter describes the following topics.

• Storage-System Trees .........................................................................3-2• Main Window ...................................................................................3-20

If you already understand these trees and the Main Window, skip to Chapter 4.

This manual assumes that you are familiar with the Windows environment for your management station.

Storage System Treesand the Main Window

3


Storage System Trees and the Main Window

Storage-System TreesTrees show the relationships between the physical and logical components of managed storage systems. Trees are analogous to the hierarchical folder structure of Microsoft Windows Explorer.

The Equipment tree shows icons for the physical components of the managed storage systems and servers and their host bus adapter (HBA) ports to which the managed storage systems are connected. The Storage and Hosts trees show icons for the logical components of the managed storage systems. The Storage tree shows the icons from a storage-system viewpoint, and the Hosts tree shows them from a host viewpoint.

A tree appears in the selected tab in the open Enterprise Storage dialog boxes in the Main window. The Equipment tree appears in the Equipment tab; the Storage tree appears in the Storage tab; and the Hosts tree appears in the Hosts tab, as shown on the following pages.

The managed storage systems are the base components in the Equipment and Storage trees. These trees display a storage-system icon for each managed storage system. The managed servers are the base icons for the Hosts tree. This tree displays a host icon for each managed server. It also displays an icon for each unmanaged server connected to a managed storage system.

You can expand and collapse the storage-system or host icons to show icons for their components (such as SP icons, disk icons, LUN icons, RAID Group icons) just as you can expand and collapse the Explorer folder structure. You use the icons to perform operations on and display the status and properties of the storage systems, their components, and their host connections.

3

Storage-System Trees 3-3


Figure 3-1 Sample Partially Expanded Equipment Tree

Storage-systemicon

3



Figure 3-2 Sample Partially Expanded Storage Tree

Storage-systemicon

3

Storage-System Trees 3-5


Figure 3-3 Sample Partially Expanded Hosts Tree

You select icons on a tree in the same way that you select items in other Microsoft Windows applications.

To Select a Single Icon

Click on the icon.

To Select Multiple Icons, Either

• Press Shift while left-clicking the first icon and last icon to select the first and last icon and all icons between them, or

• Press Ctrl while left-clicking the icons you want to select.

Host icon

3



Components of TreesThis section describes accessible, inaccessible, and unsupported storage systems and the icons that appear on the Equipment and Hosts trees.

Accessible, Inaccessible, and Unsupported Storage Systems

For Navisphere, a storage system is accessible, inaccessible, or unsupported.

Table 3-1 Accessible, Inaccessible, and Unsupported Storage Systems

Each managed storage system is represented by a storage-system icon on the Equipment, Storage, and Hosts trees. This icon consists of an image and a description.

Term Explanation

Accessible Manager can communicate with the storage system.

Inaccessible Manager has never been able to communicate with the storage system. A storage system can be inaccessible for any of these reasons:• The Agent is not running on the server. In this case, Manager displays an error message when you try

to select the server for management. Manager does not display an icon for the storage system that is inaccessible for this reason.

• The Agent running on the server was started by a user who was not logged in as root or with Administrative privileges. Manager displays an icon for a storage system that is inaccessible for this reason, and the icon indicates that the storage system is inaccessible.

• The storage system’s name is wrong in the Agent configuration file on its server. Manager displays an icon for a storage system that is inaccessible for this reason, and the icon indicates that the storage system is inaccessible.

Unsupported The storage system’s device entry in the Agent configuration file on its server is one that Manager does not support. Examples are an internal disk on the server and a seven-slot storage system with SCSI disks.

3

Components of Trees 3-7


Icons Each icon in a tree consists of an image representing the component and a description of the component. The color of the image and the letter it contains reflect the condition of the component, as follows.

Table 3-2 Icon colors

The main components of the Equipment and Storage trees are the icons for the managed storage systems, and the main components of the Hosts tree are the icons for the servers connected to managed storage system. Server icons are described below; storage-system icons are described on 3-9; and the icons for storage components are described on page 3-13.

Icons for Servers The icons for the managed servers (hosts) and their host bus adapters (HBAs) connected to managed storage systems appear in the Equipment tree. The icons for all servers (managed or unmanaged) connected to managed storage systems appear in all the trees.

Color Character Condition

Grey or green and grey

None The component and all of its components are working normally.

Faded grey or green and grey

None The component is a ghost; that is, it is an FC4700 SP that is not managed or is part of a non-FC4700 storage system that is not managed.

Orange F The component or one or more of its components has failed.

X Storage system is unsupported.

? Storage system is inaccessible.

Blue T The component or one or more of its components is in a transition state.

3



Table 3-3 Icon Images and Descriptions for Servers and Server HBAs

You display the properties of a server using the menu associated with the host icon for the server.

To Display the Menu for a Single ServerRight-click the host icon for the server whose menu you want to display.

To Display the Menu for Multiple ServersSelect the icons for the servers whose menus you want to display, and right-click.

Table 3-4 Menu Option on the Menu for Single or Multiple Servers

Image Description Tree Meaning

Hosts All trees Hosts connected to the storage system.

hostname All trees Server with name hostname connected to the storage system.

Port: UniqueID Equipment HBA port in the server connected to the storage system.UniqueID is the unique identifier of the port.

Icon description Menu option Use to

hostname Properties Display the properties of the selected servers.

3



Storage-System Icons Icons for individual storage systems appear in all the trees. In the Host tree, icons for individual storage systems connected to a host appear under a multiple-storage systems icon.

Table 3-5 Storage-System Icon Images

Image Type Meaning

FC4700 Rackmount storage system with 4 Fibre Channel host ports and Fibre Channel disks

FC 4400/4500FC5600/5700

Rackmount or deskside storage system with Fibre Channel disks.

FC5200/5300 Rackmount or deskside storage system with Fibre Channel disks.

FC5000 Rackmount or deskside storage system with Fibre Channel disks in a JBOD (just-a-bunch-of-disks) configuration. This storage system does not have SPs.

C3000 30-slot rackmount storage system with SCSI disks.

C3000 30-slot deskside storage system with SCSI disks.

C2x00 20-slot rackmount storage system with SCSI disks.

C2x00 20-slot deskside storage system with SCSI disks.

C1900 10-slot rackmount Telestor storage system with SCSI disks.

3



Table 3-6 Multiple Storage-System Icon Image

Storage-System DescriptionsA storage-system description has the following format:

storage_system_name [type]

where

C1000 10-slot rackmount storage system with SCSI disks.

C1000 10-slot deskside storage system with SCSI disks.

Image Description Tree Meaning

Storage Systems

Host Storage systems connected to the host.

storage_system_name is a name that uniquely identifies the storage system. For a storage system connected to a server running Agent revision 4.X, its format is either

A-serial# or B-serial#

where

A or B identifies either SP A or SP B as the SP used for communications with the storage system.

serial# is the unique serial number of enclosure 0 in an FC-series storage system or the chassis in a C-series storage system.

You can change this name.

type is the storage-system type for an accessible storage system, Inaccessible for an inaccessible storage system, or Unsupported for an unsupported storage system.

3



For example, A-95-2694-261 [FC4400/4500]

If automatic polling for the session (background polling) is enabled, the word “polling” appears in brackets after the description in each storage-system icon during a poll operation.

To Assign a Custom Name to a Storage System 1. In the Enterprise Storage dialog box, click the Equipment or

Storage tab.

2. Right-click the icon for the storage system whose name you want to change, and then click Set Name.

3. In the Set Storage System Name dialog box, type the new name and click OK.

Changing the storage-system name does not affect the Agent configuration file.

Storage-System MenuYou can perform operations on storage systems using the menu associated with the storage-system icon. You can display this menu for single or multiple storage systems.

To Display the Menu for a Single Storage SystemRight-click the icon for the storage system whose component you want to display.

To Display the Menus for Multiple Storage SystemsSelect the icons for the storage systems whose menus you want to display, and right-click.

Some of the options available on the menu displayed for a single storage system are unavailable (appear dimmed) when you display the menus for multiple storage systems.

3



Table 3-7 Storage-System Menu Options for a Single Storage System

Bind LUN, Create RAID Group, Create Remote Mirror, and Install Software are not available (dimmed) if the storage system is inaccessible or unsupported. If other Navisphere 4.X or 5.X applications are installed on the management station, you may see additional menu options. For information on these options, see the online help index or the manual for the application.

Option Use to

Poll Poll the storage system for status changes.

Unmanage Stop managing the storage system.

Bind LUN Bind disks into a LUN.

Create RAID Group Create a RAID Group from selected disks.

Software Installation Update existing software or install new software on the storage- system.

Set Name Assign a custom name to the storage system.

Faults Display the Fault Status Report for the storage system.

Disk Summary Display a summary of the disks in the storage system.

Create Storage Groups Create Storage Groups on the storage system.

Connect Hosts Connect servers to a Storage Group on the storage system so the servers can perform I/O to the LUNs in the Group.

Detailed View Display a graphical view of the relationships between the servers connected to the storage system and storage-system components.

Properties Display or set the storage-system properties.

3



Table 3-8 Options on the Storage-System Menus for Multiple Storage Systems

If one of the selected storage systems is unavailable or unsupported, Install Software is not available (dimmed) if the storage system is inaccessible or unsupported. If other Navisphere 4.X or 5.X applications are installed on the management station, you may see additional menu options. For information on these options, see the online help index or the manual for the application.

Storage Component IconsThe basic storage components for a storage system are

• Storage Groups (shared storage system only)

• LUNs

• RAID Groups (RAID-Group storage system only)

• Storage processors (SPs)

• Disks

The storage components for a storage system with the MirrorView options are

• Remote mirrors

• Remote mirror images

The storage components for a storage system with the SnapView option are

• Snapshot caches

• Snapshot copy LUNs

• Snapshot session

These components are represented by icons on the trees.

Option Use to

Poll Poll all selected storage systems for status changes.

Unmanage Stop managing all selected storage systems.

Software Installation Update existing software or install new software on all selected storage systems if they all are the same type of storage system.

Detailed View Display a graphical representation of the relationships between the components of all the selected storage systems.

Properties Display or set properties on all selected storage systems.

3



Table 3-9 Basic Storage Component Icons

Image Description Displayed on Meaning

Storage Groups Storage treeHost tree

Storage Groups in the storage system or accessible from the host.

StorageGroupname Storage treeHost tree

Individual Storage Group in the storage system or accessible from the host.StorageGroupname is the name of the Storage Group.

PSM LUN Storage treeHost tree

LUN in an FC4700 storage system reserved exclusively for storage-system SPs to store critical information.

LUN LUNID [RAID 5; hostnames - devicename]

LUN LUNID [RAID 5; hostnames - devicename-mirrorstatus]

Storage treeHost tree

RAID 5 LUN in RAID Group or storage system.LUNID is the ID assigned when you bound the LUN. It is a hexadecimal number. Hostnames is a list of the names of each server connected to the storage system. Devicename is the device name for the LUN on those servers. If the storage system has the MirrorView option, mirrorstatus indicates the LUN’s remote mirror status, which can be any of the following:Mirrored - LUN is the primary image LUN of a remote mirror.Mirrored/No Secondary Image - Remote mirror does not contain secondary image.Secondary Copy - LUN is a secondary image LUN for a remote mirror.

LUN LUNID [RAID 3]LUN LUNID [RAID 3; mirrorstatus]


RAID 3 LUN in RAID Group or storage system.LUNID is the ID assigned when you bound the LUN. It is a hexadecimal number. If the storage system has the MirrorView option, mirrorstatus indicates the LUN’s remote mirror status, which can be any of the following:Mirrored - LUN is the primary image LUN of a remote mirror.Mirrored/No Secondary Image - Remote mirror does not contain secondary image.Secondary Copy - LUN is a secondary image LUN for a remote mirror.

3




LUN LUNID [RAID 1/0]LUN LUNID [RAID 1/0; mirrorstatus]


RAID 1/0 LUN in RAID Group or storage system.LUNID is the ID assigned when you bound the LUN. It is a hexadecimal number.If the storage system has the MirrorView option, mirrorstatus indicates the LUN’s remote mirror status, which can be any of the following:Mirrored - LUN is the primary image LUN of a remote mirror.Mirrored/No Secondary Image - Remote mirror does not contain secondary image.Secondary Copy - LUN is a secondary image LUN for a remote mirror.

LUN LUNID [RAID 1; mirrorstatus]


RAID 1 LUN in RAID Group or storage system.LUNID is the ID assigned when you bound the LUN. It is a hexadecimal number.If the storage system has the MirrorView option, mirrorstatus indicates the LUN’s remote mirror status, which can be any of the following:Mirrored - LUN is the primary image LUN of a remote mirror.Mirrored/No Secondary Image - Remote mirror does not contain secondary image.Secondary Copy - LUN is a secondary image LUN for a remote mirror.

LUN LUNID [RAID 0; mirrorstatus]


RAID 0 LUN in RAID Group or storage system.LUNID is the ID assigned when you bound the LUN. It is a hexadecimal number.If the storage system has the MirrorView option, mirrorstatus indicates the LUN’s remote mirror status, which can be any of the following:Mirrored - LUN is the primary image LUN of a remote mirror.Mirrored/No Secondary Image - Remote mirror does not contain secondary image.Secondary Copy - LUN is a secondary image LUN for a remote mirror.

3




LUN LUNID [Disk; mirrorstatus]


Individual disk LUN in RAID Group or storage system.LUNID is the ID assigned when you bound the LUN. It is a hexadecimal number.If the storage system has the MirrorView option, mirrorstatus indicates the LUN’s remote mirror status, which can be any of the following:Mirrored - LUN is the primary image LUN of a remote mirror.Mirrored/No Secondary Image - Remote mirror does not contain secondary image.Secondary Copy - LUN is a secondary image LUN for a remote mirror.

LUN LUNID [Hot Spare] Storage treeHost tree

Hot spare in RAID Group or storage system.LUNID is the ID assigned when you bound the LUN. It is a hexadecimal number.

Unowned LUNs Storage treeHost tree

LUNs, such as hot spares, that are not owned by either SP.

SPs Equipment tree Storage tree

SPs in the storage system.

SP A Equipment tree Storage tree

In an FC-series storage system, the SP in the SP A slot in enclosure 0.In a C-series storage system, the SP in the SP A slot in the enclosure.

SP B Equipment tree Storage tree

In an FC-series storage system, the SP in the SP B slot in enclosure 0.In a C-series storage system, the SP in the SP B slot in the enclosure.

RAID Groups Storage treeHost tree

RAID Groups on the storage system.

3



Table 3-10 MirrorView and SnapView Storage Component Icons


RAID Group RAIDGroupID [RAIDtype]


Individual RAID Group identified by RAIDGroupID in the storage system.RAIDGroupID is the ID assigned when you created the RAID Group. It is a hexadecimal number between 0x00 and 0x1F.RAIDtype is Unbound if no LUNs are bound on the Group. Available RAID types are: RAID 5, RAID 3, RAID 1/0, RAID 1, RAID 0, Disk, or Hot Spare. For example, 0x03[RAID 5].

Disks Equipment treeStorage tree

Disks in the storage system.

Disk diskID All trees For an FC-series storage system, the disk in the enclosure and slot identified by diskID, which has the format m-n where m is the enclosure number and n is the slot in the enclosure containing the disk.For a C-series storage system, the disk in the slot identified by diskID, which has the format m-n where m is the letter (A, B, C, D, or E) of the SCSI bus for the slot and n is the position on the bus containing the disk.


Remote Mirrors Storage tree Remote mirrors in the storage system. This icon appears even when no remote mirror instances are defined on the storage system.

Remote Mirrormirrorname [state]

Storage tree Individual remote mirror.Mirrorname is the name of the remote mirror. It can have one of these states:Active - The remote mirror is running normallyInactive - The mirror is unavailable for host I/O. This occurs if the mirror was deactivated.Attention - Remote mirror is not running because a required condition is not met, for example, only one secondary image is available (when two are required).

3



Menus for Storage ComponentsYou can perform operations on storage components using the menu associated with the icon for the component. You can display this menu for single or multiple storage components of the same type.

If Analyzer is the only Navisphere application installed, or you started Analyzer instead of a different Navisphere application, then you can display menus for SPs, LUNs, and disks only.

To Display the Menu for a Single Storage ComponentRight-click the icon for the component whose menu you want to display.


Remote Mirror Imageimagename - imagetype [state]

Storage tree Imagename is the name of the image. Imagetype identifies whether the image is a primary or secondary image. The image can have one of these states:In-Sync (or identical or congruent) - Secondary image is identical to the primary. This state persists only until the next write to the primary image, at which time the image state becomes Consistent.Consistent - Secondary image is identical to the primary, or it was identical in the past. If the mirror is not fractured, the software will try to make the secondary image In-Sync after receiving no I/O for a given period of time (the quiesce threshold).Synchronizing - Software is applying changes to the secondary image to mirror the primary, but the current contents of the secondary are not known and are likely not usable.Out-of-Sync - None of the above; the secondary image requires synchronization with the primary image.

Snapshot Cache - SP A Storage tree SP A’s snapshot cache, which consists of any LUNs owned by SP A selected to participate in snapshot sessions.

Snapshot Cache - SP B Storage tree SP B’s snapshot cache, which consists of any LUNs owned by SP B selected to participate in snapshot sessions.

Snapshot Sessions Storage tree Snapshot sessions running in the storage system.

Snapshot Session Storage tree Individual snapshot session running in the storage system.

Snapshot Copy LUNS Storage tree LUNs participating in snapshot sessions.

SnapshotCopyLUNname Storage tree Individual LUN participating in a snapshot session.SnapshotCopyLUNname is the name of the snapshot copy LUN.

3



To Display the Menus for Multiple Storage Components of the Same TypeSelect the icons for the storage-system components whose menus you want to display, and right-click.

You cannot display the menus for multiple enclosures.

What Next? Continue to the next section, “Main Window.”

3



Main Window

The Main window is common to all Navisphere management applications. The menus and the menu options available when you right-click an icon in a dialog box in the workspace may vary with the applications installed. All other Main window components and functions are identical for all applications.

Application Icon The Application icon on the left side of the title bar provides overall status of all storage systems managed by the current session as follows.

Status bar

Equipment,Storage, and

Host tabs

Storage-systemselection filters

Application iconMenu bar

Toolbar

Workspace

3

Main Window 3-21


Table 3-11 Application Icon Image

Menu Bar From the menu bar in the Main window you can display these menus: File, View, Operations, Analyzer, Window, and Help.

If other Navisphere 4.Xor 5.X applications are installed on the management station, you may see additional menus. For information on these menus, see the on-line help Index or manual for the application.

File Menu

Icon Color Meaning

Grey Manager has detected no failures in any managed storage system.

Flashing blue

Manager has detected one or more storage systems in a transitional state.

Flashing orange

Manager has detected a failure in one or more storage systems, or one or more storage systems are inaccessible.

Option Use to

New Window Open a new Enterprise Storage dialog box.

Select Agents Change the list of servers or SPs that Navisphere uses to determine which storage systems to manage.

Save Save the application’s configuration to the most recently opened application configuration file for use by the next session.

Save As Save the application’s configuration to a new, unnamed application configuration file you specify for use by the next session.

Open Restore the application’s configuration to the one defined by the application configuration file you select.

Exit Exit the session and close the Main window.

3



View Menu

Operations Menu

The Analyzer menu bar — in addition to the File, View, Operations, Window, and Help menu options — has an Analyzer option. The Analyzer menu options follow.

Analyzer Menu

Option Use to

Toolbar Show or hide the toolbar.

Status Bar Show or hide the status bar.

Options Set the network time out, set the name and location of the host file, set the automatic polling interval for Navisphere, and enable or disable automatic polling of all managed storage systems.

Option Use to

Automatic polling Enable or disable automatic polling for all managed storage systems.

Poll All Storage Systems

Manually poll all managed storage systems; that is, survey them once for status changes.

Option Use to

View Log File Specify a log file for Analyzer to read and display. We suggest the extension .npl (performance log file) for log filenames, but you can use any extension you want.

Data Logging Display the logger startup screen that lets you specify logging parameters and start logging.

3

Main Window 3-23


Window Menu

Help Menu

Toolbar The buttons on the toolbar let you perform operations on all managed storage systems at once. To perform operations on individual storage systems, use the menu associated with the storage-system icon.

Option Use to

Close All Close all Enterprise Storage dialog boxes.

Cascade Cascade the open Enterprise Storage dialog boxes.

Tile Horizontally Tile horizontally the open Enterprise Storage dialog boxes.

Tile Vertically Tile vertically the open Enterprise Storage dialog boxes.

Enterprise Storage Activate an open Enterprise Storage dialog box.

Option Use to

Contents & Index Display the on-line Help table of contents and index.

Using Help Display information about using the on-line Help.

About Navisphere Display the version of each Navisphere application installed on the management station.

3



Table 3-12 Toolbar Buttons

If other Navisphere 4.X or 5.X applications are installed on the management station, you may see additional toolbar buttons. For information on these buttons, see the online help index or the manual for the application.

Workspace The workspace in the Main window contains the dialog boxes that you use to perform storage-system tasks. It always contains at least one Enterprise Storage dialog box, unless you have closed it. You can open additional Enterprise Storage dialog boxes in the workspace. If you have installed any additional Navisphere 4.X applications on the management station, another type of dialog box may open in the workspace when you start the application.

Button Name Click to

Poll Manually poll all managed storage systems; that is, survey them once for status changes.

Firmware Download Update Core Software and PROM code on the storage systems you select. This is a Manager button, not an Analyzer button.

Faults Display a list of any hardware faults encountered on any managed storage system. This is a Manager button, not an Analyzer button.

Help Display the on-line Help.

3

Main Window 3-25


Enterprise Storage Dialog Boxes

An Enterprise Storage dialog box displays the Equipment, Storage, or Hosts tree of the managed storage systems, depending on whether the Equipment, Storage, or Hosts tab is selected. You can specify the managed storage systems to display in the Equipment, Storage, or Hosts tree using Filter By and Filter For as listed in the table that follows.

3



Table 3-13 Filters for Displaying Managed Storage Systems

When you open a Navisphere session, it displays one Enterprise Storage dialog box with a number. During the session you can open additional Enterprise Storage dialog boxes and close them. You might want one dialog box displaying the Equipment tree, one displaying the Storage tree, and one displaying the Hosts tree. Only one dialog box is active.

To Open a New Enterprise Storage Dialog BoxFollow the path File → New window.

A new Enterprise Storage dialog box opens in the workspace and it is the active one.

To Activate a Different Enterprise Storage Dialog BoxEither click in the dialog box you want to activate, or on the Window menu click the Enterprise Storage dialog box you want to activate.

To Close All Enterprise Storage Dialog BoxesOn the Window menu, click Close All.

The Equipment tab displays the Equipment tree; the Storage tab displays the Storage tree; and the Hosts tab displays the Hosts tree. You use the Equipment tree to manage the physical components of the managed storage systems; the Storage tree to manage the logical components of the managed storage systems; and the Hosts tree to

Filter By Filter For

All N/A

Fault Condition NormalFaulted

Host Individual hostname

Storage System Type C1000 SeriesC1900 SeriesC2x00 SeriesC3000 SeriesFC50xx SeriesFC5000 SeriesFC4400/4500 SeriesFC4700 SeriesFC5200/5300 SeriesFC5600/5700 SeriesUnsupportedInaccessible

Subnet Individual subnet IP address

3

Main Window 3-27


manage the LUNs and the storage systems to which the servers connect.

You perform operations on

• All managed storage systems using the menu bar, or on selected managed storage systems using the menu associated with the storage-system icon

• Selected managed storage-system components using the menu associated with the component’s icon

• Selected servers using the menu associated with the host icon

Status Bar The status bar in the Main window contains information fields that provide the following:

• Automatic Polling indicator. If Automatic Polling is highlighted, automatic polling is enabled; if it is dimmed, automatic polling is disabled.

• Feedback about application operation.

• Brief description of a toolbar button when you position the cursor over the button.

Window Configuration

When the Main window opens, it uses the default application configuration values for the following:

• The size and position of the Main Window and any open Enterprise Storage dialog boxes.

• In the Enterprise Storage dialog boxes, any Filter By and Filter For settings and the selected tab.

If you change any of these values (for example, you filter for FC4700 storage systems and select the Storage tab), you can save them to either

• the default application configuration file so future sessions open the Main window with these values, or

• a custom application configuration file so you can restore the window to these values at any time during a session.

3



To save the current configuration to the default configuration file:

Either exit the application or on the File menu, click Save.

The current application configuration values are saved to the default application configuration file.

To change the name and location of the default configuration file:

1. In the Main window, follow the menu path

View → Options.

A User Options dialog box opens, similar to the following.

2. In Host File Path, type or select the path to use for the default application configuration file.

3. Click OK to apply your change and close the dialog box.

3

Main Window 3-29


To save the current configuration to a new custom configuration file:

We recommend that the name for a custom configuration file have the extension.nfx.

1. If the folder that you want to contain the new custom configuration file does not exist, create it.

2. On the File menu, click Save As.

3. In Save As dialog box, select the folder to hold the new custom configuration file.

4. In File name, enter the name for the new custom configuration file.

5. Click Save.

6. In the confirmation window that opens, click Yes.

The current application configuration values are saved to the new custom application configuration file.

Restoring the default or a custom configuration:

You can restore the current configuration to the values specified in either the default or a custom application configuration file.

1. On the File menu on the Main window menu bar, click Open.

2. In the Open dialog box, select the drive or folder containing the custom application configuration file you want to use.

3. Either select the desired configuration file or enter its name in File name.

4. Click Open.

The current configuration of the Main window is restored to the configuration specified in the selected file.

What Next? This chapter explained using the Navisphere trees and the Main Menu. To use Analyzer, continue to Chapter 4.

3



Analyzing Performance 4-1

4

This chapter explains using Analyzer to examine aspects of performance. It assumes you understand using the Navisphere trees as explained in the previous chapter. Topics are

• Analyzer Menu Bar and Right-Click Menu Options ....................4-2• Analyzer Modes .................................................................................4-3• Zooming and Resizing Charts..........................................................4-6• Analyzer Charts .................................................................................4-8• Performance Summary Chart...........................................................4-9• Performance Detail Chart ...............................................................4-13• IO Size Distribution Charts (LUN Only) ......................................4-20• Performance Properties...................................................................4-24• Exporting and Printing Charts.......................................................4-30• Using the Performance Logger ......................................................4-32

AnalyzingPerformance

4


Analyzing Performance

Analyzer Menu Bar and Right-Click Menu OptionsThe Analyzer menu bar — in addition to the File, View, Operations, Window, and Help menu options — has an Analyzer option. The Analyzer menu options follow.

Table 4-1 Analyzer Menu Options

Analyzer can display performance information on LUNs, SPs, and disks. The options available when you right-click on a LUN, SP, or disk icon are as follows.

Table 4-2 Analyzer Main Window Right-Click Options for SPs, LUNs, and Disks

If other Navisphere 4.X and 5.X applications are installed on the management station, you may see additional menu options. For information on these options, see the on-line help Index or the manual for the application.

Option Use to

View Log File Specify a log file for Analyzer to read and display. We suggest the extension .npl (performance log file) for log filenames, but you can use any extension you want.

Data Logging Display the logger startup screen that lets you specify logging parameters and start logging.

Icon Description Menu Option Displays

SP A or SP B Performance Summary Performance Summary window (performance range).

Performance Detail Performance Detail window (line chart).

LUN lunID [RAIDtype] Performance Summary Performance Summary window (performance range).


IO Size Distribution Summary

Line chart showing numbers of I/Os of various sizes.

IO Size Distribution Detail

Line chart showing numbers of I/Os of various sizes over time.

Disk diskID Performance Summary Performance Summary window (performance range).


4

Analyzer Modes 4-3


Analyzer ModesAnalyzer has two modes: real-time mode, where it takes data directly from agents; and playback mode, where it takes data from performance log files recorded previously. For Analyzer to gather data in real-time mode, automatic polling must be enabled or you must poll manually.

Analyzer starts in real-time mode. It will close any real-time charts and switch to playback mode when you choose Analyzer → View Log File from the Main window bar. Log file playback requires at least one Analyzer log file on the management station. (A sample log file ships with the Analyzer in directory Program files\EMC\Navisphererevision\Analyzer\Log Files.) Starting the logger service to create log files on page 4-32.

Analyzer closes any log file display charts and switches to real-time mode any time you request a chart from the Storage tree window. However, Analyzer remains in playback mode if you request a chart by right-clicking from the Analyzer tree window.

To Run Analyzer in Real-time Mode1. Select one or more LUNs, SPs, or disks in the Storage tree view.

Generally, LUNs and SPs provide broader workload metrics than disks.

2. Right-click the selected item.

3. Select Performance Summary, Performance Detail, IO Size Distribution Summary (LUN only), or IO Size Distribution Detail (LUN only) on the drop-down menu.

4. If you are currently in playback mode, click OK on the confirmation dialog box. (If you click Cancel, the operation aborts.)

After pausing for a few moments to gather data, Analyzer displays a chart of the selected type for the selected components.

To Run Analyzer in Playback Mode1. On the Main menu bar, follow the path Analyzer → View Log

File.

2. Select OK on the confirmation dialog box that appears.

4



3. From the standard Windows file selection dialog box, select a log file to play back. A sample log file ships with the Analyzer in the Program files\EMC\Navisphererevision\Analyzer\Log Files directory. Creating new analyzer log files is explained in “Using the Performance Logger,” starting on 4-32.

After you select a log file, a display similar to the following appears.

4. In the Log File Display Start Time dialog box, specify a date/time for the first sample you wish to display. You can use the left- and right-arrow keys, and you can scroll through the options using the up- and down-arrow keys.

Select the year, month, date, hours, minutes, and/or seconds you want. The range of start times is limited to the period included in the log file. Analyzer will play back the file starting at the specified interval and update the display based on the polling interval used when the log file was recorded.

5. Click OK.

Analyzer displays a Performance Summary chart using the log file.

4

Analyzer Modes 4-5


In the Analyzer tree, you can check the boxes of one or more LUNs, SPs, or disks. Generally, LUNs and SPs provide broader workload metrics than disks.

Analyzer will leave playback mode and return to real-time mode any time you select a chart by right-clicking from the Storage tree window. However, it will stay in playback mode if you select a chart by right-clicking from the Analyzer tree window.

The Analyzer can display up to 255 data samples for each component, beginning with the first sample at the start time entered.

4



Zooming and Resizing ChartsAnalyzer lets you zoom into (expand) any type of chart that has a time sequence as its X-axis.

To zoom into a chart:

• Click the X Axis or below and drag to select a new date range.

• Release the mouse button.

The chart expands to display the date range you selected.

You can use the horizontal scroll bar to display different parts of the chart.

To reset after zooming (that is, restore the chart display to its original state), right-click in the chart area and select the Reset Zoom option.

The following three figures show a graph before zooming, the same graph with a zoom area selected, and the graph after zooming.

Figure 4-1 Display Before Zoom

4

Zooming and Resizing Charts 4-7


Figure 4-2 Chart With Zoom Area Selected (Left Mouse Button)

Figure 4-3 Chart After Zoom (Release of Left Mouse Button)

You can resize a chart as you can many windows by positioning the mouse pointer over a corner of a chart to change the cursor to a

4



double-ended arrow and then left-clicking and pulling the chart to the desired size. You can resize the tree and chart panes the same way.

Analyzer ChartsAnalyzer has four types of chart: Performance Summary, Performance Detail, IO Size Distribution Summary (for LUNs only), and IO Size Distribution Detail (LUNs only). To choose among these, you select the objects you want, and then right-click to display a menu.

Each Analyzer chart overlays the Main window. The Performance Summary and Performance Detail charts include their own tree.The Analyzer tree differs from the Navisphere tree in that each component on the Analyzer tree has a checkbox that you can select or clear. Each component also has an icon that shows the device type and the device name.

Analyzer displays information for only those components whose boxes are checked. The display depends on data from the Agent or a log file. Data from the Agent (not a log file) becomes available only after two polling cycles. You may need to wait several minutes for a meaningful display.

In the Analyzer tree, each storage component you selected displays as the primary component, which you can expand to show related components. That is, if you select a LUN in the Main window and specify Performance Summary, then LUNs will appear as primary in the Analyzer tree. You can expand the LUN’s icon to show individual LUNs; and you can expand an individual LUN component to show the owning SP and the LUN’s constituent disks.

Analyzer takes the data directly via Host Agents from the SP Agents (in real-time mode) or from a log file generated by the Performance Logger (in playback mode) and displays the data graphically in the type of chart you specify.

4

Performance Summary Chart 4-9


Performance Summary ChartThe Performance Summary chart shows the activity, based on the performance property selected, for each device whose checkbox is selected. Each device display is cumulative and shows the range of all values seen since the Analyzer started receiving data (in real time mode) or at the time the log file recorded the data (in playback mode).

The Summary chart shows a value range over an entire period, while the Detail chart (described next) tracks the progression of values at intervals during the period.

The chart has three parts: Analyzer tree, to select devices; performance properties list, to select type of information (such as Utilization or Response Time); and the chart display space, which graphs the performance properties of the devices.

The chart display uses the following symbols:

• a red line for the range of values, minimum to maximum

• a green circle for the average value, and

• a pink diamond for the latest value

To Display a Performance Summary Chart

1. Select the SP Agent(s) whose storage-system performance you want to analyze (Chapter 2). You can analyze performance in real time, or by the log file you select using the sequence Analyzer → View Log File.

2. Click the tab for the type of tree (Equipment, Storage, Agents) from which you want to work. The Storage tab provides the best selection of storage components.

3. Expand a storage-system enclosure and, for an SP, right-click the SP. For LUNs or disks, expand the icon RAID groups or LUNs until an individual LUN is displayed.

4. From the pull-down menu, select Performance Summary.

5. In the Analyzer display, select the check boxes for the components whose performance you want to graph. You can de-select check boxes to remove the display of a component.

4



Analyzer displays information for only those components whose boxes are checked. The display depends on data acquired from an Agent (or a log file). Data acquired from an Agent (not a log file), becomes available only after a polling cycle. You may need to wait several minutes for a meaningful display.

6. From the list on the lower left, select the performance property you want for the chart. The default performance property is Utilization. For a Performance Summary chart, you can select one performance property (described later in this chapter, on page 4-24) at a time.

For example, a Performance Summary chart for a LUN, an SP, and the LUN’s constituent disks looks like the following.

Figure 4-4 Sample Performance Summary Chart

You can select a device by clicking it. To remove a device's values from the chart, clear the checkbox.

4

Performance Summary Chart 4-11


The performance properties list lets you select any listed property; the one shown is Utilization. All the performance properties are explained starting on page 4-24.

To display plots for all disks in a LUN, double-click the plot for the LUN. Double-click again to remove the disk plots.

Performance Summary Device Menu

To display the menu for the Performance Summary chart, right-click a device in the tree. The Performance Summary device menu options follow.

Table 4-3 Performance Summary Device Menu Options

Performance Summary Chart MenuTo display the menu for the Performance Summary chart, right-click anywhere in the chart area. The menu for the Performance Summary chart has the following items.

Menu Option Displays


IO Size Distribution Summary (LUN only)

Line chart showing numbers of I/Os of various sizes that occurred with the LUN.

IO Size Distribution Detail (LUN only)

Line chart showing numbers of I/Os of various sizes that occurred with the LUN over time.

Remove Object After you confirm, deletes the object (and any components) from the tree and chart.

Properties Appears only if Manager is installed. Displays device properties.

Help An explanation of the chart.

4



Table 4-4 Performance Summary Chart Options

Performance Summary Chart Scroll Bar

Immediately below the Performance Summary chart area is the scroll bar: use it to move the display to the left or right.

Menu Item Use to

Collapse/Expand Remove each LUN’s SP and disks from the chart and hide their tree entries; or add each LUN’s SP and disks to the chart and display and select their tree entries. Collapse is available only if the LUN is currently expanded; expand is available only if it is collapsed.

Hide Remove the component nearest the menu and its plot from the chart.

Performance Detail Display a Performance Detail chart for the component.


Line chart showing numbers of I/Os of various sizes.


Line chart showing numbers of I/Os of various sizes over time.

Export Chart Display a submenu that lets you save chart• Values as comma-separated text to the file you specify, with the

extension .csv, or to the Windows clipboard• Graph as a jpeg (.jpg) image to the file you specify or to the

Windows clipboard• For details on exporting chart data, see “Exporting and Printing

Charts” on page 4-30.

Print Chart Display a standard Windows Print dialog box so you can send an image of the chart to the printer selected by the dialog box.

Sort Sort the items along the X (horizontal) axis by axis by different criteria. When you click the Sort button, a chart-sorting dialog box appears that lets you sort by• Minimum values• Maximum values, • Average values,• Latest values, or• Default order (restores the original order)

Help Display an explanation of the Summary chart.

4

Performance Detail Chart 4-13


Performance Detail ChartThe Performance Detail chart is a line chart that shows the performance property (or properties) of the components selected when you selected the chart. Unlike the Summary chart, you can select more than one performance property at a time. A sample line chart showing Utilization of the SP, the LUN, and the disks in the LUN follows.

The Detail chart shows the progression values at intervals during a period, unlike the Summary chart, which shows the range of all values seen during a period.

Samples are taken at the specified sampling interval (seconds) and displayed on the X-axis until 21 samples are on display. After the 21st sample, the total number displayed remains 21. However, the intervals between the samples are scaled so that each represents

(total-logging-time) / 20

All sample points are still displayed on the chart at the times at which they were taken.

Analyzer can plot a maximum of 255 samples at a time. Each sample after the 255th replaces the oldest sample.

To Display a Performance Detail Chart

1. Select the SP Agent(s) whose storage-system performance you want to analyze (Chapter 2). You can analyze performance in real time, or by the log file you select using the path Analyzer → View Log File.

2. Click the tab for the type of tree (Equipment, Storage, Agents) you want to work from. The Storage tab provides the best selection of storage components.

3. Expand a storage-system enclosure and, for an SP, right-click the SP. For LUNs or disks, expand the icon RAID groups or LUNs until an individual LUN is displayed.

4. From the pull-down menu, select Performance Detail.

5. In the Analyzer display, select the check boxes for the components whose performance you want to graph. You can de-select check boxes to remove the display of a component.

4



Analyzer displays information for only those components whose boxes are checked. Also, the display depends on data acquired from the Agent (or a log file). Data acquired from the Agent (not a log file), becomes available only after a polling cycle. You may need to wait several minutes for a meaningful display.

6. In the list box on the lower left, select the performance properties you want for the chart. The default performance property is Utilization. All the performance properties are explained starting on page 4-24.

With a Detail chart, you can select multiple performance properties; for example, you can select both Utilization and Total IOs/second. And you can select a time span to expand (zoom) with the left mouse button as explained on page 4-6.

A sample performance detail chart follows.

4



Figure 4-5 Sample Performance Detail Chart Showing Utilization of SP

You can zoom (expand) this chart by clicking the time axis, dragging to include the area to zoom, and then releasing the mouse button. To reset, right-click in the chart area and select Reset Zoom.

Performance Detail Device Menu

To display the menu for the Performance Detail chart, right-click a device in the tree. The Performance Detail device menu options follow.

4



Table 4-5 Performance Detail Device Menu Options

Performance Detail Chart Menu

To display the menu for the Performance Detail chart, right-click anywhere in the chart area. The menu for the Detail chart has the following items.

Table 4-6 Performance Detail Chart Options

Menu Option Displays

Performance Summary

Performance Summary window (performance range).


Line chart showing the number of I/Os of various sizes.


Line chart showing the number of I/Os of various sizes over time, displayed at each interval.

Remove Object After you confirm, deletes the object (and any components) from the tree and chart.

Properties Appears only if Manager is installed. Displays device properties.

Help Display an explanation of the chart.

Menu Item Use to

Reset Zoom Restores the display of a zoomed chart to its original size. Appears only if a chart has been zoomed.

Hide/Show Legends Hide or display legends (map of line and marker colors and shapes).

Configure Detail Chart Display the Configure Detail dialog box, which lets you customize the color and line style of each line that represents an item-characteristic combination. See below for details.

Export Chart Lets you choose to export Chart Data to File, Chart Data to Clipboard, Graph to Clipboard, or Graph to File. For details on exporting chart data, see “Exporting and Printing Charts” on page 4-30.

Print Chart Display a standard Windows Print dialog box, and after you choose a filename and click OK, send an image of the chart to the printer selected and configured via the dialog box.


4



Performance Detail Chart Data PopupYou can obtain data on any device whose behavior is displayed in the Detail chart by moving the mouse cursor to the display and pressing the left mouse button. Analyzer will display device data while you depress the button. To retain the data display, move the mouse out of the chart area. This feature lets you learn which device performed as shown on the chart. A sample device identifier display follows.

Figure 4-6 Sample Performance Detail Chart Data Popup

4



To Change Details on a Performance Detail Chart1. Right-click anywhere in the Performance Detail chart.

The menu shows the button Configure Detail Chart.

2. Click the button Configure Detail Chart

The Configure Detail Chart dialog box appears as follows:

To Change the Chart Configuration

1. Select a line style.

Line Style lets you specify dashed or solid lines. You can select Solid (default), None, Long Dash, Dotted, Short Dash, Long Short Long, and Dash Dot.

2. Select a line color.

Line Color shows a palette from which you can choose line color. To display the palette, click the color displayed beside Click to pick.

3. Select a point style.

4



Point Style lets you specify a graphic to mark the end of each line. You can select None (default), Dot, Box, Triangle, Diamond, Star, Vertical Line, Horizontal Line, Cross, Circle, Square, Inverse Triangle, Diagonal Triangle, Open Triangle, Open Diamond, Inverse Triangle, or Open Inverse Triangle.

4. Select a point color.

Point Color shows a palette from which you can choose point color. To display the palette, click the color displayed beside Click to pick.

5. Select a line width.

Line sets the relative line width: 1 (default), 2, 3, 4, or 5.

6. Select a point size.

Point Size sets the size of the point graphic (Dot, Box, Triangle, and so on, as selected above). You can choose values from 0 (invisible point display graphic) through 11 points.

7. Select a Y-axis scale.

The Y-axis scale sets the order of magnitude for the Y-axis (vertical axis) numbers. For example, if the Y-axis numbers range from 0 to 8000, and you change the scale to 1, the range will change from 0 to 8.

To see the chart display in the display area, click OK. The chart settings you specify become standard for detail chart display after you save them.

To display the menu for the Detail chart, right-click anywhere in the chart area.

4



IO Size Distribution Charts (LUN Only)The IO Size Distribution charts show how many I/Os within given sizes occurred with a LUN. (A chart that shows the relative sizes of writes or reads to a LUN is also known as histogram.) There are two charts: a summary and a detail chart.

The IO Size Distribution Summary chart shows the total number of reads and writes of specific sizes since the chart opened.

The IO Size Distribution Detail chart shows read and write totals over a period of time at each interval since the chart opened. The Detail chart can include as many 255 data sampling intervals.

To Display an IO Size Distribution Chart

1. Select the SP Agent(s) whose storage system performance you want to analyze (Chapter 2). You can analyze performance in real time, or by the log file you select using the sequence Analyzer → View Log File.

2. Click the tab for the type of tree (Equipment, Storage, Agents) you want to work from. The Storage tab provides the best selection of storage components.

3. Expand a storage-system enclosure and, for LUNs, expand the icon RAID groups or LUNs until an individual LUN is displayed.

4. From the pull-down menu, select IO Size Distribution Summary or IO Size Distribution Detail.

The Analyzer display depends on data acquired from the Agent (or a log file). Data acquired from the Agent (not a log file), becomes available only after a polling cycle. Also, Distribution chart display is based on data acquired since this chart opened, not on data acquired since data logging was enabled. Therefore, you may need to wait several minutes or longer for a meaningful display.

With a Detail chart, you can select a time span to expand (zoom) with the left mouse button as explained on page 4-6.

4

IO Size Distribution Charts (LUN Only) 4-21


The following figure shows and IO Size Distribution Summary chart. A detail chart appears on page 4-22.

Figure 4-7 Sample IO Size Distribution Summary Chart

This display shows I/O size distribution concentrated at 1K, 8K, and 64K for reads and writes.

IO Size Distribution Summary MenuTo display the menu for an IO Size Distribution summary chart, right-click a device in the tree. The menu options follow.

4



Table 4-7 IO Size Distribution Size Summary Chart Menu Options

A Sample IO Size Distribution Detail chart follows.

Figure 4-8 Sample IO Size Distribution Detail Chart

Menu Item Use to

IO Size Distribution Detail Display the Distribution Detail chart.

Export Chart Export Chart Data to File, Chart Data to Clipboard, Graph to Clipboard, or Graph to FIle. For details on exporting and printing charts, see page 4-30.


Help Display an explanation of the IO Size Distribution chart.

4

IO Size Distribution Charts (LUN Only) 4-23


The previous display shows a consistent rate for read, write, and total I/O sizes for a 22-hour period.

You can zoom (expand) this chart by clicking the time axis, dragging to include the area to zoom, and then releasing the mouse button. To reset, right-click in the chart area and select Reset Zoom.

IO Size Distribution Detail Chart Menu

To display the menu for the IO Size Distribution chart, right-click anywhere in the chart area. The menu for the IO Size Distribution Detail chart shows the following choices.

Table 4-8 IO Size Distribution Detail Chart Options

Menu Item Use to

Hide/Show Legends Hide or show the legend key display that identifies the various IO sizes.

Configure Detail Chart Display the Configure Detail dialog box, which lets you customize the color and line style of each line that represents an item-characteristic combination. See page 4-18 for details.

Export Chart Lets you choose to export Chart Data to File, Chart Data to Clipboard, Graph to Clipboard, or Graph to FIle. For details on exporting chart data, see “Exporting and Printing Charts” on page 4-30.



4



Performance PropertiesThe Analyzer Performance Properties apply to disks, SPs, LUNs, and/or snapshot sessions, shown as D, S, L, and SS in the selection list. For example, Utilization D/S/L means that the Utilization property applies to disks, SPs, and LUNs.

The snapshot properties apply to a Snapshot Session node only. In the Snapshot Session, these properties are shown as SS.

The performance properties fall into six general categories:

• Basic (includes the properties utilization, queue length, response time, total bandwidth, and total throughput)

• Workload (includes max outstanding requests, max request count, max requests in queue, and read/write throughput, size, and bandwidth)

• Read cache (includes percentage of used prefetches, hit ratio, miss rate, and hit rate)

• Write cache (includes miss rate, hit ratio, hit rate, flush ratio, dirty page percentage, block flush rate, forced flush rate, high watermark flush on rate, idle flush on rate, low watermark flush off rate, and flush rate)

• Miscellaneous (includes average seek distance, disk crossing percentage, disk crossing rate, average busy queue length, and service time)

• SnapView (includes read and write information about the SnapView source LUN and Snapview).

The performance properties are as follows.

4

Performance Properties 4-25


Table 4-9 Performance Properties

Property Description

Utilization (D/S/L) For Disk, SP, LUN (D/S/L). Describes the fraction of a certain observation period that the system component is busy serving incoming requests. An SP or disk that shows 100% (or close to 100%) utilization is a system bottleneck since an increase in the overall workload will not affect the component throughput; the component has reached its saturation point. Since a LUN is considered busy if any of its disks is busy, LUN utilization usually presents a pessimistic view. That is, a high LUN utilization value does not necessarily indicate that the LUN is approaching its maximum capacity.

Queue Length (D/S/L)

The average number of requests within a certain time interval waiting to be served by the component, including the one in service. An (average) queue length of zero indicates an idle system. If three requests arrive at an empty service center at the same time, only one of them can be served immediately; the other two must wait in the queue, resulting in a queue length of three.

Response Time (ms) (D/S/L)

The average time, in milliseconds, required for one request to pass through a system component, including its waiting time. The higher the queue length for a component, the more requests are waiting in its queue, thus increasing the average response time of a single request. For a given workload, queue length and response time are directly proportional.

Total Bandwidth (MB/s) (D/S/L)

The average amount of data in Mbytes that is passed through a system component per second. Larger requests result usually in a higher total bandwidth than smaller requests. Total bandwidth includes both read and write requests.

Total Throughput (IO/s) (D/S/L)

The average number of requests that pass through a system component per second. Since smaller requests need a shorter time for this, they usually result in a higher total throughput than larger requests do. Total throughput includes both read and write requests.

Max Outstanding Requests (S)

The largest number of commands on the SP at one time since statistics logging was enabled. The value measures the biggest burst of requests sent to this SP at a time.

Max Request Count (L)

The largest number of requests queued to this LUN at one time since statistics logging was enabled. This value also indicates the worst instantaneous response time due to the maximum number of waiting requests.

Max Requests in Queue (D)

The maximum number of requests waiting to be serviced by this specific disk since statistics logging was enabled.

4



Read Bandwidth (MB/s) (D/S/L)

The average number of Mbytes read that were passed through a component per second. Larger requests usually result in a higher bandwidth than smaller ones.

Read Size (KB) (D/S/L)

The average read request size in KBytes. This number indicates whether the read workload is oriented more toward throughput (I/Os per second) or bandwidth (MB/second). For a finer distinction of I/O sizes, use an IO Size Distribution chart.

Read Throughput (IO/s) (D/S/L)

The average number of read requests passed through a component per second. Since smaller requests need less processing time, they usually result in a higher read throughput than larger requests.

Write Bandwidth (MB/s) (D/S/L)

The average number of Mbytes written that were passed through a component per second. Larger requests usually result in a higher bandwidth than smaller ones.

Write Size (KB) (D/S/L)

The average write request size in Kbytes. This number indicates whether the read workload is oriented more toward throughput (I/Os per second) or bandwidth (MB/second). For a finer distinction of I/O sizes, use an IO Size Distribution chart.

Write Throughput (IO/s) (D/S/L)

The average number of write requests passed through a component per second. Since smaller requests need less processing time, they usually result in a higher write throughput than larger requests.

Used Prefetches (%) (L)

The indication of prefetching efficiency. To improve read bandwidth, two consecutive requests trigger prefetching, thereby filling the read cache with data before it is requested. Thus, sequential requests will receive the data from the read cache instead of from the disks, which results in a lower response time and higher throughput. As the percentage of sequential requests rises, so does the percentage of used prefetches.

Read cache Hit Ratio (L)

The fraction of read requests served from both read and write caches vs. the number of read requests to this LUN. The higher the ratio, the better the read performance.

Read Cache Hits/s (L)

The number of read requests that were satisfied by either write or read cache within a second. A read cache hit occurs when recently accessed data is referenced while it is still in the cache.

Read Cache Misses/s (L)

The rate of read requests that could not be satisfied by the SP cache and therefore required a disk access.

Dirty Pages (%) (S) Percentage of cache pages owned by this SP that were modified since they were last read from or written to disk. In an optimal environment, the dirty-pages percentage will not exceed the high watermark for a long period.


4



Flush Ratio (S) The fraction of the number of flush operations performed vs. the number of write requests. A flush operation is a write of a portion of the cache to make room for incoming write data. Since the ratio is a measure for the back-end activity vs. front-end activity, a lower number indicates better performance.

Write cache Hit Ratio (L)

The fraction of write requests served from the write cache vs. the total number of write requests to this LUN. The higher the ratio, the better the write performance.

Write Cache Hits/s (L)

The number of write requests per second that were satisfied by the write cache since they have been referenced before and not yet flushed to the disks. Write cache hits occur when recently accessed data is referenced again while it is still in the write cache.

Write Cache Misses/s (L)

The number of write requests per second that could not be satisfied by the write cache since their data was not currently in the cache from a previously disk access.

MBs Flushed/ s (MB/s) (S)

The number of megabytes per second written from the write cache to the disks. The value is a measure of back-end activity.

Forced Flushes/ s (L) Number of times per second the cache had to flush pages to disk to free space for incoming write requests. Forced flushes indicate that the incoming workload is higher than the back end workload. A relatively high number over a long period of time suggests that you spread the load over more disks.

High Water Flush On (S)

Number of times, since the last sample, that the number of modified pages in the write cache reached the high watermark. The higher the number, the greater the write workload coming from the host.

Idle Flush On (S) Number of times, since the last sample, that the write cache started flushing dirty pages to disk due to a given idle period. Idle flushes indicate a low workload.

Low Water Flush Off (S)

Number of times, since the last sample, that the number of modified pages in the write cache reached the low watermark, at which point the SP stops flushing the cache. The higher the number, the greater the write workload coming from the host. This number should be close to the High Watermark Flush On number.

Write Cache Flushes/s (S)

Number of times per second that the write cache performed a flush operation. A flush operation is a write of a portion of a cache for any reason; it includes forced flushes, flushes resulting from high watermark, and flushes from an idle state. This value indicates back-end workload.


4



Disk Crossing (%) (L) Percentage of requests that require I/O to at least two disks vs. the total number of server requests. A disk crossing may involve more than two disks; that is, more than two stripe element crossings. Disk crossings relate to the LUN stripe element size. Generally, a low value is needed for good performance.

Average Busy Queue Length (D/S)

Average number of requests waiting at a busy system component to be serviced, including the request that is currently in service. Since the queue length is counted only when the component is busy, the value indicates the frequency variation (burst frequency) of incoming requests. The higher the value, the bigger the burst, and the longer the average response time at this component.

Average Seek Distance (GB) (D)

Average seek distance in gigabytes. Longer seek distances result in longer seek times and therefore higher response times. Defragmentation might help to reduce seek distances.

Disk Crossings/s (L) Indicates how many back-end requests per second used an average of at least two disks. Disk crossings are counted for read and write requests. Generally, a low value is needed for good performance.

Service Time (ms)(D/S/L)

Time, in milliseconds, a request spent being serviced by a component. It does not include time waiting in a queue. Service time is mainly a property of the system component. However, larger I/Os take longer and therefore usually result in lower throughput (IO/s) but better bandwidth (MB/s).

Reads from Snapshot Cache (SS)

With the optional SnapView software only: The number of reads during this session that have resulted in a read from the Snapshot cache (instead of a read from the source LUN).


4



Reads from Snapshot LUN (SS)

With the optional SnapView software only: The number of read requests on SnapViews during this Snapshot Session.

Reads from Snapshot Source LUN (SS)

With the optional SnapView software only: The number of reads during this Snapshot session from the source LUN. It is calculated by the difference between Total Reads in Session and Reads From Cache.

Writes to Snapshot Source LUN (SS)

With the optional SnapView software only: The number of writes during this Snapshot session to the source LUN (on the pertinent SP).

Writes to Snapshot Cache (SS)

With the optional SnapView software only: The number of writes to the source LUN this session that triggered a copy-on-write operation (the first write to each snapshot cache chunk region).

Writes Larger Than Cache Chunk Size (SS)

With the optional SnapView software only: The number of writes to the source LUN during this session that were larger than the chunk size (they have resulted in multiple writes to the cache).

Cache Chunks Used in Snapshot Session (SS)

The number of chunks that this session has used.


4



Exporting and Printing ChartsYou can export chart data, export graph data, or print a chart as follows.

To Export Chart Data

1. Move the mouse cursor to the chart area, and right-click to display the pull-down menu.

2. Select Export Chart → Export Chart Data to File.

The Save As dialog box appears.

3. Specify the filename and select the location.

4. The Analyzer saves the file as a comma delimited file under the name you specify, with the extension .csv. The default filename is xxxData.csv, where xxx is Summary, IO Summary, and IO Detail, depending on the kind of chart you are working with.

The output format looks like the following (for Performance Summary data):

LUN 0x01 [RAID5;toolsdev1 - \\.\PhysicalDrive1],SP A,Enclosure 0 Disk 6,Enclosure 1 Disk 0,Enclosure 1 Disk 1,Enclosure 1 Disk 2,Current Values:,99.9823,31.9149,44.3425,48.1707,41.4634,43.5976,Average Values,99.981,28.5642,45.4579,44.3035,44.5652,43.5922,Max Values:,99.9859,32.2188,49.3902,50.9146,48.6322,47.2561,Min Values:,99.9734,20.4545,41.2844,38.9058,38.7195,40.2439

4

Exporting and Printing Charts 4-31


5. Double click the saved file to have Microsoft Excel open it. In Excel, the file will display like this (for a Performance Summary data):

The Analyzer uses the same format for Export Data to Clipboard.

To Export Graph Data in jpeg (.jpg) Format1. Move the mouse cursor to the chart area, and right-click to

display the pull-down menu.

2. Select Export Chart → Export Chart Graph to File.

The Save As dialog box appears.

3. Specify the filename and select the location.

4. The Analyzer saves the file in jpeg (.jpg) format under the name you specify, with the extension .jpg. The default filename is xxxData.jpg, where xxx is Summary, Detail, or IO Summary, or IO Detail, depending on the kind of chart you are working with.

The Analyzer also uses jpeg format for Export Graph to Clipboard.

To Print a Graph1. Move the mouse cursor to the chart area, and right-click to

display the pull-down menu.

2. Select Print.

Analyzer sends the chart to the default printer.

LUN 0x00 [RAID1/0; toolsdev1]

SP B Enclosure 0 Disk 0 Enclosure 0 Disk 1

Current Values: 0.952381 11.538462 0 0.952381

Average Values: 0.469667 9.01991 0.152207 0.31746

Max Values: 0.952381 11.538462 0.456621 0.952381

Min Values: 0 7.33945 0 0

4



Using the Performance LoggerGenerally, the views Analyzer provides in real time mode are most useful as snapshots. For detailed study, you may want to record log files during specific times on specific SPs and play these files back.

Logging is controlled by the Performance Logger. The logger has two parts:

• A Windows system service, accessible from Windows via the Control Panel/Services as Array Performance Logger

• Logger control, accessible from the Analyzer menu button or the Windows Start icon, configures and controls the logger executable.

The logger runs independently of the Analyzer GUI. To acquire log information, the logger requires an agent running on the pertinent server.

You can change logger parameters while the logger is running, but your changes will not take effect until you restart the logger.

Logger error messages are recorded in the Event Log.

Setting up the Performance Logger

Setting up the Performance Logger consists of identifying the servers whose performance data you want to log.

The logger can record information for any server on the network that is running the Navisphere Agent. When you specify multiple servers, the logger polls them all and adds their performance data to the performance log file. The Main window displays data from all agents at the same time. You can use this to compare the load and performance of all your storage systems.

To Set Up the Performance Logger

1. From the Main window, use the path Analyzer → Data Logging. You can also click on the Start button and follow the path Programs → Navisphererevision → Logger Control.

The logger control process starts and displays the Logger Control dialog box.

2. On the Logger Control dialog box, click the Agents tab.

4

Using the Performance Logger 4-33


The logger control process displays a dialog box that lets you add or remove an Agent. By default, the local SP Agent (localhost) is listed as an Agent to log. The logger can record data for localhost only if the host running Analyzer is connected to a storage system.

Figure 4-9 Performance Logger Agents Dialog Box

The Agents to Log box lets you select a different server or multiple servers on which to monitor storage systems.

With FC4700-based storage systems, each SP is a server with its own IP address and agent name.

The logger treats each server whose name you add on its own line as a separate server, and, if you check the Separate Log File (Set) for Each Agent box, it will generate a separate log file for that server. If an FC4700 storage system you want to analyze has two SPs, you will probably not want the data for those SPs written to different log files. You can tell the logger to write both SPs’ data to one log file by entering the SP Agent names on the same line, separated by spaces, in the entry box.

For example, if you specify different log files for each agent, then

• entry sys1_spa sys1_spb in the same Agents to Log box will send performance data on both agents to the same log file; and

4



• entries

sys1_spasys1_spb in different Agents to Log boxes will send performance data to different log files.

3. To add one or more servers, type the SP Agent name(s) or IP address(es) in the Agents to Log field and click Add. To remove a server, select the agent name from the list and click Remove. (The logger does not check the validity of the names and addresses you specify until you start logging by clicking Start.)

4. If you want separate log files for each agent, select the Separate Log File (Set) for each Agent box.

By default, one log file records information for all selected agents. However, even if you select different log files for each agent, you can have data for two agents written to a single log by entering both SP Agent names on the same line, as explained above.

5. If you want these agent specifications saved for the next time the logger runs, click the Save button.

Continue to set log file properties as follows.

Setting Log File PropertiesThe log file properties include the log filename(s), maximum size, and start and stop times.

To Set Log File Properties:1. On the menu bar, select Analyzer → Data Logging

The logger process displays the Logger Control dialog box.

2. On the Logger Control dialog box, click the Files tab.

The logger control program displays the Files dialog box, as follows.

4



3. For the Log Files Directory, specify a directory.

The Log File Directory stores the log files the logger creates. You can use the Navisphere Analyzer directory for these files or create and use a separate directory for them. The default directory, named Log Files in the Analyzer directory, is created at installation. For Analyzer 5.0, the pathname is drive:\Program Files\EMC\Navipshere 5.0\Analyzer\Log Files.

If you want to use a different directory (which must already exist), use the Browse button to find and specify the log file directory, and then click OK.

4. For the Log File Basename, specify a basename.

The Log File Basename forms the base of the log filename. The basename is the filename of the first log file created, and is expanded only if you specify more than one log file (described below). When logging is on, the logger writes performance data to this file. The sample log file basename is arraylogfile.npl. We

4



suggest the extension .npl for log files. If the basename you specify will produce the filename of an existing file, then you must also check the box Overwrite Existing Files.

If you check Create New Log File and specify more than one log file in the Max Number of Files box below, then the logger creates a new log file after the old one fills, using the form basename1.npl, basename2.npl, basename3.npl, and so on.

If the logger Agents tab checkbox Separate Log File (Set) for Each Agent is selected, the SP Agent name is added as a prefix, with an underscore, to the basename. So, with three agents (abcsp, xyz, sp3), the filenames would be abcsp_log.npl, xyz_log.npl, and sp3_log.npl. If the Max Number of Files is set to 5 for this example, there would potentially be 15 log files – named abcsp_log1.npl through abcsp3_log5.npl. If you specify no extension for the basename, then the log filenames will have no extension.

5. Select or deselect the Overwrite Existing File box.

If you check the Overwrite Existing File box, and the specified log file exists when the logger starts, then the logger will overwrite the existing log file.

If you do not check the Overwrite Existing File box, and the specified log file exists when the logger starts, then the logger stops and posts an error message to the Windows Event Log and in the status bar at the bottom of the logger dialog.

6. Specify the action you want after the last log file fills.

In the After Log File Fills section, when the log file fills, the logger takes the action you specify:

• If you select the Stop Logging button, the logger will stop logging.

• If you select the Overwrite Oldest Sample button, the logger will start overwriting the oldest sample in the current log file.

• If you select the Create New Log File button, the logger will close the current log file and either create a new log file or stop, depending on settings you specify.

7. Specify the maximum number of log files.

4



In the Max Number of Files box, specify the maximum number of log files you want to allow per server. As each log file fills, the logger will close it and create a new file, using a filename with the next integer appended to the basename. For example, assume the maximum number is three and the basename is log. The logger will create and fill three log files, log1, log2, and log3; then it will either delete log1 and create log4 or it will stop logging, depending on what you specify. This mechanism makes it easy to identify the most recent log file: the most recent file has the highest file name suffix.

8. Specify the action you want after the logger has filled the last file it is allowed to create.

In the After Last File section, you tell the logger what to do after it fills the last log file it creates.

• If you select the Delete Oldest File button, the logger will delete the oldest file and create and use a new one in its place. The new filename will have the same basename plus the next available integer.

• If you select the Stop Logging button, the logger will stop logging

9. If you want these file specifications saved for the next time the logger runs, click the Save button.

10. Continue to specify the maximum log files sizes.

Setting Log File Sizes The log file sizes dialog box includes settings for the sampling interval, times covered in each log file, and a size estimator.

4



1. In the General section, specify a sampling interval.

The Sampling Interval is the interval, in seconds, at which the logger retrieves data from the Agents on the selected servers. This value is independent of the poll interval set by Analyzer or Navisphere. You can enter a value or select one from the Sampling Interval pull-down list. A long interval has less impact on system performance and produces smaller log files; a smaller interval shows short-term spike effects more precisely. One approach to performance analysis is to time your applications over both a long period (such as a work week, with an interval of 10 minutes or more) and a short peak period (such as 20 minutes, with an interval of 60 seconds). You can specify a value from 5 seconds to 86,400 seconds (24 hours).

The logger gets status data from the storage-system Agent running on each server. The logger cannot get this data faster than the Agent. Therefore, it is useless to set the logger sampling interval to less than the Agent polling interval. The default Agent polling interval is 60 seconds. For details on Agent polling, see the Navisphere Agent manual.

Polls may take a long time, depending on the number of storage systems and the amount of I/O.

4



2. For Time Represented by Each Log File (h:m:s), specify the time period you want each log file to include. A longer time period combined with a shorter interval produces larger log files that have more information; a shorter period with a longer interval produces shorter files with less information.

You can use the Log File Size Estimator button to estimate the log file sizes that of different Sampling Interval-Time Represented combinations will produce.

For the Snapshot Copy section, specify the maximum number of Snapshot sessions for which you want performance data logged. The logger will record information on an unlimited number of sessions. If you specify a number of sessions n, use the First or Last buttons to specify whether the Logger should record the first or last n sessions.

You can also use the Log File Size Estimator button to estimate the log file sizes needed for Snapshot sessions.

If you want to schedule times for logging to occur automatically, without human intervention, continue to specify the start and end times for logging as follows.

Scheduling Logger Operation (Advanced Tab)The Advanced properties specify when the logger should start and stop.

To Schedule Logging Operation

1. On the menu bar, select Analyzer → Data Logging

The logger control program starts (if not already running).

2. Click the Advanced tab.

The program displays the Advanced dialog box as follows.

4



By default, logging starts when you click the logger control Start button, and continues until you click the Stop button. If you want to specify another start time and/or end time, you can do so with the Schedule Logging checkbox.

The Schedule Logging checkbox directs control to start and/or stop logging data at a specific date or time. While running with a scheduled request, the logger polls and logs data only between the specified start and end date/time. The logger control need not be running when the start date/time occurs. The logger stops logging at the end date/time.

1. Select or deselect the Schedule Logging checkbox. If you do not select this, skip to step 3.

The Start Date, End Date, Start Time, and End Time list boxes let you set the time range for data logging. You can change the sample start and end times using the Start and End list boxes; or you can use the arrow buttons. The two arrow buttons next to the date boxes display a calendar from which you can select the date.

2. Specify a start date and time and/or an end date and time. To specify a date and time, select part of the displayed date: month, date, year, hours, minutes, or seconds; and set as you wish. Set each part until you reach the start date and time you want.

4



3. To save these file specifications for the next time the logger runs, click the Save button.

Continue to start logging as follows.

Starting and Stopping Logging

Buttons on the right panel of the logger dialog boxes let you start, stop, and restart logging as follows.

To Start Logging

Click the logger control Start button. The logger starts logging immediately or at your specified start time. Logging stops when you click the Stop button or at your specified end time. Closing the Logger Control dialog box does not stop logging.

The logger runs independently of Analyzer. Exiting Analyzer or logging off the management station does not stop logging.

To Stop Logging

Click the logger control Stop button. Logging stops while the control program continues.

To Restart Logging

Click the logger control Restart button. The control program writes the values currently in the dialog pages to the Windows registry. Then, the control program sends the logger a message telling it to restart with the newly written values. The Restart button is available only when the logger is running.

To Save Logger Settings

Click the Save button. The Save button writes the values set currently in the dialog pages to the registry. It neither starts nor stops the logger. Changing the registry values has no effect on any currently running logger. However, the next time you start the logger, the new values (unless changed) control it. Closing logger control without saving or restarting discards any changes.

To Close the Logger Control Program

Click the Close button. The Close button terminates the logger control program. Close neither starts nor stops the logger – it merely stops the control program. Any logging operation, start and/or stop,

4



scheduled for the future will occur whether or not the control program is running.

Using the Log File Size EstimatorThe log file size estimator provides an approximate size, in Mbytes, of each log file. There may be multiple log files, based on the settings made in the Logger Files dialog box.

To Run the Log File Size Estimator

1. On the menu bar, select Analyzer → Data Logging

The logger control program starts (if not already running) and displays the Sizes dialog box.

2. Click the Log File Size Estimator button. The Estimator dialog box appears as follows.

3. Specify a trial sampling interval, as detailed on page 4-38.

4



4. Specify a trial time interval, also as detailed on page 4-39.

5. Specify the number of LUNs, SPs, and Disks attached to all storage systems on the agent(s) whose data this log file will record. If you don’t want the log to include Snapshot copy performance data, skip to step 8.

6. For Snapshot copy data, specify the number of Snapshot sessions whose data you will want to include in the log file.

7. For each session, specify the estimated time each snapshot session will run (Snapshot session start to Snapshot session stop).

8. Click Estimate.

The estimated log file size appears in the Estimated field. Based on the estimated size, you may want to change the log file File and/or Sizes settings, starting on page 4-37.

What Next? To use Analyzer information to diagnose the cause of performance problems, continue to the next chapter.

4



Understanding Performance Data 5-1

5

This chapter describes how to solve performance problems using the Navisphere Analyzer. Major topics are

• Understanding Workload .................................................................5-2• Identifying a Performance Problem ................................................5-5• Performance Analysis Flowchart.....................................................5-6• Working With the Storage-System Cache.......................................5-9• Addressing Server Problems ..........................................................5-14

UnderstandingPerformance Data

5


Understanding Performance Data

Understanding WorkloadBefore trying to solve any performance problem, you should understand the workload that your application is applying to the storage system. This will help you correlate your knowledge of your application to Analyzer statistics.

Application workload has two important factors: the characteristics of the workload and its intensity. The main characteristics are

• I/O size

• Read/Write ratio

• Locality of reference (seek area, hot spots, and random/sequential access patterns)

These workload characteristics are not constant; they vary with time depending on the work the application performs, and they vary from LUN to LUN.

To use Analyzer to examine the workload characteristics, select the LUNs in the Storage tree and create Performance Summary charts to show their characteristics. Read and Write, throughput and bandwidth will tell you about the read/write traffic both in terms of data and I/O requests. Clicking an individual LUN will let you look at a time series of the quantity of interest.

Use this information along with your knowledge of your application to make sure the storage system is performing as expected. For example, suppose you are running a file server that is mostly used for data retrieval. If you find that the LUN workload is write intensive, this indicates that something is wrong with the storage system and further investigation is needed.

I/O size data is also very important, particularly the quantity of traffic at each I/O size. For example, if you set the buffer size of a file system to 16 Kbytes, then you would expect a large amount of 16-Kbyte traffic. If the amount of data traffic is small, then a volume manager or I/O driver might be combining or splitting requests. Average Read and Average Write size data are useful but can be misleading. An average I/O size of 20 Kbytes may be mostly 16-Kbyte I/O with some 256-Kbyte I/Os; this is very different from half 8 Kbytes and half 32 Kbytes that would give the same result. To distinguish sizes, Analyzer lets you look at the I/O size distribution (histogram), which records the number of I/O of each size.

5

Understanding Workload 5-3


You cannot directly derive application program locality data using Analyzer. Such detail requires more complex analysis that cannot be achieved in real time. You can draw some locality conclusions from Analyzer cache data and disk seek data.

The quantity of work that is applied to a LUN is also important. A single-threaded application presents requests one at a time to the LUN. If the thread supplying requests waits between I/O operations (“think time”), then the LUN will have periods of inactivity. During these lightly loaded conditions, LUN utilization is a good measure of load. As load increases, the utilization rises until it approaches 100%. At this point, a single thread cannot request any more work from the LUN and the throughput is at its maximum. If the application is multi-threaded, the load and throughput can continue to increase; at this point, Queue length is a better measure of applied load.

LUN response times measure the way that the storage system responds to an applied load. As the load increases, response times do not change until one of the components that implements the LUN reaches saturation (100% utilization). When this happens, response time will rise linearly with applied load. Unfortunately, response time depends on both load level and load characteristics because writes take longer than reads and large I/Os take longer than small ones. LUN response time is an excellent measure of how the storage system is affected by changing loads, but you must identify the factor of the workload that is changing.

If the system is to perform at its theoretical maximum, the workload intensity must be evenly distributed across all of the LUNs. In practice, this is difficult to achieve. Workload intensity varies with time as does the distribution of work between LUNs. The variations can occur over widely different time scales. Business cycles such as end of month, end of quarter, and even end of year processing may be very different from normal day-to-day operation. Operational schedules such as backups and data- mining operations can cause problems. The user load associated with logging in and at break period can also vary intensity and change distributions.

Analyzer lets you look at the LUN workloads and determine

• the load at critical times of day

• which LUNs have most traffic

• which LUNs are idle

5



This might help you decide how to move data from one LUN to another to better distribute the load. With some thought, you might be able to eliminate some bottlenecks by shifting the schedules for some system operations. After implementing any easy or obvious schedule improvement, you can use Analyzer to understand the impact of your changes. But remember, distributing data to solve one problem may cause another at a different time.

Storage Systems Under Light LoadsUnder light load situations, indicated by low disk and SP utilizations, the response times experienced by the requests are mostly due to the processing times at the various system components. The queuing that can occur at higher loads does not influence the response times. For requests serviced from the cache, the response times are typically a few milliseconds. For requests serviced by the disks, the response times can be as high as 25 milliseconds. Under the following situations with light loads, you may be able to reduce response time by reconfiguring the read or write caches.

• Single-threaded sequential reads: If the cache is configured properly, prefetching will allow almost all the requests to be serviced from the cache. This will be indicated by a read cache hit ratio of 100%.

• Locality of reference by reads: If the size of the most frequently used data set is smaller than the read cache size, most of the requests will be serviced from the cache.

For write operations, data is written to the cache. Under light loads, a small write cache might suffice to absorb the data. Beyond that, tuning the cache configuration properties is unlikely to influence performance.

Configuring the cache involves setting the read and write cache sizes, and the cache page sizes. Some of the general guidelines for this are provided later in this chapter.

Storage Systems With BottlenecksThere are several areas of computer system operations that pertain to bottlenecks:

• Operating cycles (for resource-hungry procedures such as backups),

5

Identifying a Performance Problem 5-5


• Business cycles (for periodic demanding applications such as data reduction for reports), and

• Time-of-day cycles (such as a 9:00 am surge when users log in).

To use Analyzer effectively with bottlenecks, first you need to determine when the bottleneck(s) occur, so that you can record a meaningful Analyzer log file that includes them.

Then note the components that have high utilizations (SPs and disks), click them to expand and show the time course. In general, if the utilization of a component is higher than 70%, the device is operating near its capacity.

Do not use LUN utilization values to detect bottleneck conditions. Since a LUN is considered busy if at least one of its disks is busy, LUN utilization usually presents a pessimistic view. Thus, a high LUN utilization value does not necessarily indicate that the LUN is operating at its maximum capacity.

If the utilization of components such as the disks and SP is low, the problem is likely at the server. See the section "Addressing Server Problems" on page 5-14.

Identifying a Performance ProblemAfter appraising your workload and deciding on obvious solutions, you can try a procedural approach to identify performance problems. The following flowchart guides you toward this goal.

5



Figure 5-1 Performance Analysis Flowchart

Run Analyzer

Any SP or disk showgreater than 70%

utilization?

Yes

Component is a bottleneck

Is a hardware failure (failed disk or loose connection)

degrading performance?

Yes

Fix hardware No

Improve efficiency of component - See section"Eliminating bottlenecks."

No

Yes

Fix server problem (more processing power or fewer demands). See section"Addressing server problems."

No

Problem may be single-threaded application

Use Analyzer to examineaverage busy queues ofSPs and disks

Any average busy queue length close to 1 (1 to 1.5)?

Work with storage system cacheto improve I/O efficiency for application. See section "Working with the storage-system cache."

Server processing problem(network I/O, or too many users

overtaxing processor)?

Start

No

Yes

Storage system is running efficiently; tuning probably cannot improve performance.

Application that uses componentis probably single-threaded and contributing to problem

5

Eliminating Bottlenecks 5-7


Eliminating BottlenecksGenerally, if the disks have higher utilization than the SP, then the disks are the bottleneck. You can work with the cache or change the RAID type to push the SP harder, as described later in this section.

CAUTION!Changing the RAID type of a LUN requires you to unbind the LUN, which eliminates all data stored on it. Before unbinding, back up all data that you want on the LUN; then after rebinding with the new RAID type, load the backed-up data.

If the SP utilization is higher than the disks, then the SP is the bottleneck. You should consider moving some load off the SP (perhaps by moving some LUNs to another SP), or changing the schedules of some operations.

When you identify the component with the highest utilization, you identify the performance bottleneck and limiting factor. If this component is running at maximum capacity, system performance is limited by the component and any improvement in system performance requires eliminating the bottleneck. This can be achieved in several ways.

• Reduce the demand on the component. This might be as simple as scheduling some operations to another time of day. You can reduce disk workload by using caching so that data can be accessed in the cache rather than from disk. You may also be able to reduce workload by choosing a more appropriate RAID type (but see the caution above).

• Move some workload to another component. You can do this by moving files and databases (or index tables of databases) to different LUNs. If the disk module utilizations within a striped LUN are not equal (within 15% of each other), then the I/O distribution of the striping needs to be improved. Reducing the stripe element size increases the chances of localizing I/O to one disk, but it also can have the detrimental effect of increasing the number of stripe-crossing I/Os.

• Add components. If Analyzer shows disks as the bottleneck, then you can improve performance by adding more disks or LUNs. But if Analyzer shows an SP as a bottleneck, adding more disks

5



will not help performance. For example, adding disks will not help if disk utilization is 40% and the SP utilization is 65%, since the SP is the limiting factor.

Disk-Based Factors - Raid Type and Stripe Element SizeYou may be able to reduce the effect of a disk bottleneck by using a different RAID type for one or more LUNs.

• For single-threaded applications that perform large, sequential I/Os, use RAID 3. For RAID 3, stripe element size cannot be modified. However, stripe size (number of disks) does influence performance. For instance, a nine-disk RAID 3 LUN has a higher bandwidth capacity than a five-disk RAID 3 LUN. However, partial stripe writes should be avoided since they result in poor performance. Since the chance of a partial write with the larger stripe size is higher than with the smaller size, the better stripe size depends on the distribution of I/O sizes in a workload.

• For general-purpose throughput and response time with high availability, use RAID 5. The RAID 5 stripe-element size should be larger than most I/O sizes, but smaller than the disk buffer size. Ideally, each read request and small write request should be less than the stripe-element size. The choice for an optimal stripe size (number of disks in a RAID 5 LUN) also depends on the distribution of I/O sizes in a workload. The tradeoffs are the same as those mentioned in the case of RAID 3. Small, partial stripe writes, where the I/O sizes are smaller than half the stripe size, are expensive.

• For throughput and response time without high availability, use RAID 0. RAID 0 yields better write performance than RAID 5 by avoiding the overhead of computing and writing parity data. RAID 0 read performance is identical to RAID 5. The RAID 0 stripe element size should be larger than most I/O sizes, but smaller than the disk buffer size (if the disk buffer size is known). The number of disks in a single LUN does not influence performance significantly.

• For highest availability and better read response time, use a mirrored RAID type: RAID 1 or RAID 1/0. Advantages include higher fault tolerance and better read response times than RAID 5. Disadvantages include significantly slower writes and higher storage cost.

5

Working With the Storage-System Cache 5-9


Working With the Storage-System CacheFor best performance with most applications, each SP should have its maximum amount of cache memory and you should use the default settings for the cache properties. Analyzer shows how the cache affects the storage system, and lets you tune the cache properties to best suit your application.

A storage-system cache has two parts: a read cache and a write cache. The read cache uses a read-ahead mechanism that lets the storage system prefetch data from the disk. Therefore the data will be ready in the cache when the application needs it. The write cache buffers and optimizes writes by absorbing peak loads, combining small writes, and eliminating rewrites.

You can change read cache size, write cache size, and cache page size to achieve optimal performance. The best sizes of the read and write caches depend on the read/write ratio. A general norm for the ratio of reads to writes is two reads per write; that is, reads represent 66% of all I/Os.

Since the contents of write cache are available for read operations as well, you should allocate most of the available SP memory to the write cache. However, since the write cache is flushed after a certain timeout period, a read cache is also required to hold active data for longer periods of time.

Read cache size: The read cache holds data that is expected to be accessed in the near future. If a request for data that is in the cache arrives, the request can be serviced from the cache faster than from the disks. And each request satisfied from cache eliminates the need for a disk access, reducing disk load. If the workload exhibits a “locality of reference” behavior, where a relatively small set of data is accessed frequently and repeatedly, the read cache can improve performance. In read-intensive environments, where more than 70% of all requests are reads, the read cache should be large enough to accommodate the data set that is most frequently accessed. For sequential reads from a LUN, data that is expected to be accessed by subsequent read requests is read (prefetched) into the cache prior to being requested. Therefore, for optimal performance, the read cache should be large enough to accommodate prefetched data for sequential reads from each LUN.

Write cache size: Write cache serves as a temporary buffer where data is stored temporarily before it is written to the disks. Cache writes are

5



far faster than disk writes. Also, write-cached data is consolidated into larger I/Os when possible, and written to the disks more efficiently. (This reduces the expensive small writes in case of RAID 5 LUNs). Also, in cases where data is modified frequently, the data is overwritten in the cache and written to the disks only once for several updates in the cache. This reduces disk load. Thus, the write cache absorbs write data during heavy load periods and writes them to the disks, in an optimal fashion, during light load periods. However, if the amount of write data during an I/O burst exceeds the write cache size, the cache fills. Subsequent requests must wait for cached data to be flushed and for cache pages to become available for writing new data.

The write cache provides sustained write speed by combining sequential RAID 5 write operations and writing them in RAID 3 mode. This eliminates the need to read old data and parity before writing the new data. To take advantage of this feature, the cache must have enough space for one entire stripe of sequential data (typically 64 KBytes x [number-of-disks -1], or, for a five-disk group, 256 KBytes) before starting to flush. Note that the sequential stream can be contained in other streams of sequential or random data.

Cache page size: A page size of two Kbytes is a good general-purpose choice. Larger page sizes will reduce the cache page management overhead (due to the page allocation table and cache page header), but result in more memory fragmentation. Ideally, the cache page size should be slightly larger than most I/O sizes. For the write cache, if the I/O sizes are 16 Kbytes and larger, a page size of 16 Kbytes is optimal. If the I/O sizes are significantly smaller than 16 Kbytes, a 16-Kbyte page size will result in poor performance. In most cases, a two-Kbyte page size yields reasonably good performance.

Using Analyzer to Monitor Cache PerformanceAnalyzer provides counters to monitor cache performance, as follows.

Write Cache Counters The write cache counters show forced flush rates, dirty cache pages, and cache hit ratio. They are as follows.

Forced flush rate: Forced flushing occurs when a write request arrives and there are no empty pages in the cache. Forced flushes are undesirable since they incur the overhead of cache management with

5



little chance of optimizing writes. A high forced flush rate indicates that the write cache is not large enough to accommodate the modified data. If the workload imposes heavy loads for long periods of time, the cache will eventually run out of empty pages and cause forced flushes. However, if the workload consists of short bursts, the cache size should be large enough to accommodate the write data during a burst. The cache content will be flushed during the reduced load periods. You might reduce the forced flush rate either by increasing the cache size or by reducing the cache page size to reduce memory fragmentation. As a general rule, if the number of forced flushes is more than half the number of writes, performance will be better without the write cache.

% Dirty pages: Dirty pages are pages that have been modified in the write cache but have not yet been flushed to disk. The Dirty pages % count should be lower than the high watermark under most situations. A consistently high percentage of dirty pages might indicate a balance between the incoming load and the rate at which cache pages are flushed. A larger cache might be required to absorb any I/O burst. Alternatively, you might resolve the problem by decreasing the cache page sizes to reduce memory fragmentation in cache, if you suspect fragmentation. A high percentage of dirty pages does not necessarily indicate a problem. However, when the forced flush rate is high, the dirty pages percentage is likely to be high as well.

Write cache hit ratio: A high cache hit ratio indicates good performance. A low value might be caused by large I/O sizes, since large writes bypass the cache. In this case, the forced flush rate will be low and cache tuning will not help. A low hit ratio value might also be caused if the cache is too small for the incoming workload. In this case, the forced flush rate will be high. You might resolve this problem by increasing the write cache size, if possible, or by reducing the cache page size to reduce cache memory fragmentation.

High and low watermark flush rates: The high watermark specifies the maximum value for dirty pages beyond which the cache page flushing starts. This is done so that empty cache pages will be available for subsequent write requests. The low watermark specifies the minimum value of dirty pages below which the cache page flushing stops, triggered by the high watermark. (Note that there are other conditions that also trigger cache page flushing, but the low watermark does not stop cache flushing under other conditions). A high value for a high watermark flush rate indicates a heavy

5



incoming workload. However, a high value does not necessarily indicate a performance problem.

The default values for the watermarks yield reasonably good performance under most situations. However, the high and low watermarks can be useful tuning parameters. A high value for the dirty pages percentage as the high watermark results in less frequent cache page flushing, but leaves fewer empty pages in the cache to accommodate any subsequent I/O burst. A lower value can accommodate I/O bursts better, but causes more frequent cache page flushing. Thus, the loads imposed on the disks are different under the two situations. Likewise, a lower value for the low watermark is better suited to accommodate I/O bursts, but imposes a steady load on the disks. This may influence performance differently, depending on the load imposed on the disks by read requests. A higher value for the low watermark allows the more frequently modified pages to stay in the cache, allowing more data rehits, which in turn reduce the load on the disks. An optimal value for the high and low watermark depends on the nature of the incoming workload.

Read Cache Counters The read cache counters show prefetch effectiveness and cache hit ratio. They are as follows.

% Used prefetches: A high value indicates good performance for sequential reads. In general, it is difficult to detect a sequential stream of references to a LUN if there are other active reference streams to that LUN. However, if there is a single sequential stream addressing a LUN and the value of Used prefetches % is low, this indicates a performance problem. You might improve performance either by increasing the read cache size or decreasing the cache page size.

Read cache hit ratio: The hit ratio is a measure of the efficiency of your tuning and should exceed 20%. The read cache hit ratio should be high, particularly for sequential reads from a LUN. The prefetch occurs at the speed of the disks. If the arrival rate of requests is high and the I/O sizes are large, the prefetch process might not be able to keep up with the load, causing a low cache hit ratio.

For random reads from a LUN, the cache hit ratio will be high only if

• the requests exhibit good locality of reference (accessing a limited set of data during a phase) and

• the read cache is large enough to hold the active data set (working set)

5



For random reads, or if the working set is too large, the read cache will not provide any advantage. Therefore, a low read cache hit ratio indicates a performance problem only if

• the access stream is sequential and the load is low, or

• the access stream is random but has a good locality of reference behavior.

Read cache hit rate: The hit rate shows the impact of caching on your application. It should be high if the read cache hit ratio is high and the workload is relatively heavy. A low cache hit rate with a high cache hit ratio indicates that the workload is not driving the cache to its maximum capacity. If the cache hit ratio is low, the cache hit rate will also be low.

5



Addressing Server ProblemsWorkload analysis is often the key to solving problems that appear to be caused by the storage system but that you can solve by changing the way the server accesses the storage system.

Server Throttles Often software designers face problems of resource allocation. There are not enough resources to maintain as many parallel activities (paths, threads, tasks, queued requests, and so on) as desired and the designer resorts to an execution throttle. The ultimate effect of these throttles is to limit the number of concurrent requests submitted to the storage system. A storage system can process multiple requests in parallel. By reducing the number of concurrent requests, the storage system reduces the throughput. On average, 16 concurrent threads are required to reach 90% maximum throughput of a five-disk RAID 5 LUN. Reducing concurrency below this level will considerably reduce the system throughput, resulting in longer response times seen by the application.

Analysis of LUN queue lengths can provide the clues needed to diagnose server throttles. You might find that throttles are imposed on each LUN, some LUNs, SPs, buses and even the entire storage system.

Data Alignment The alignment of large I/Os (larger than 256 Kbytes) and small I/Os (smaller than eight Kbytes) does not normally impact performance. I/O sizes between these can unduly affect performance. With random I/O, a stripe crossing can require additional seek time. Even though this additional seek time occurs in parallel and does not directly affect response time, it causes extra busy time for a second disk. The second disk is not available for other work; this reduces the maximum throughput of the LUN, which can consequently lead to increased response times.

Misalignment can be caused by file systems, table spaces, and other storage objects not being aligned to stripe boundaries. When this happens, the I/O is misaligned, and in sequential operations, every nth I/O will cause a stripe crossing. To resolve this kind of problem, you must find some kind of padding mechanism to realign the objects in question, or change a block or buffering factor to improve alignment.

5

Addressing Server Problems 5-15


Sometimes the application has no concept of alignment, and I/Os are generated with random starting addresses. This is called unaligned I/O and unless you can create some kind of alignment, there is nothing to be done.

Comparing LUN I/O sizes and disk I/O sizes can give you clues about alignment, as can the DiskCrossing performance properties in the properties list. Precise analysis requires a fibre channel analyzer or trace facility and consultation with a qualified storage system engineer.

5



EMC Navisphere Analyzer Version 5.X Administrator’s Guide s-1

Sales and ServiceLocations

CorporateHeadquarters

EMC CorporationHopkinton, Massachusetts01748-9103 U.S.A.508-435-10001-800-424-EMC2

Asia PacificHeadquarters

EMC Japan K.K.P.O. Box 323Shinjuku-Mitsui Bldg., 551-1, Nishi-Shinjuku2-chome Shinjuku-kuTokyo 163-0466 JAPANTel: 81-3-3345-3211Fax: 81-3-3345-3221

North American Salesand Service Locations

Atlanta, GA 404-705-4750Baltimore, MD 410-850-4324Bellevue, WA 425-201-6400Bethesda, MD 301-530-0091Boston, MA 617-449-8100Charlotte, NC 704-521-9773Chicago, IL 708-390-8800Cincinnati, OH 513-745-0300Cleveland, OH 216-573-1162Columbus, OH 614-436-3900Dallas, TX 972-233-5676Denver, CO 303-294-9966Farmington Hills, Ml 313-553-4810Ft. Lauderdale, FL 305-776-3622Greensboro, NC 910-665-1077

s-2 EMC Navisphere Analyzer Version 5.X Administrator’s Guide

Sales and Service Locations

Houston, TX 713-621-9800Indianapolis, IN 317-577-9766Inglewood, CA 310-364-1222Irvine, CA 714-833-1442Lenexa, KS 913-469-9292Long Island, NY 516-393-5045Minneapolis, MN 612-835-1994Montreal, Quebec 514-856-6166Nashville, TN 615-781-4394New York City, NY 212-564-6866Ottawa, Ontario 613-233-0111Orlando, FL 407-855-4087Phoenix, AZ 602-955-0702Philadelphia, PA 610-834-7740Pittsburgh, PA 412-922-5222Portland, OR 503-293-8450Raleigh, NC 919-420-0405Rochester, NY 716-387-0970Salt Lake City, UT 801-532-1454San Diego, CA 619-576-1880San Francisco, CA 415-871-1970Southington, CT 860-620-6730Springfield, NJ 201-467-7979St. Louis, MO 314-469-9005Tampa, FL 813-282-0274Toronto, Ontario 905-206-1580Tulsa, OK 918-663-2255Vancouver, BC 604-270-1657Vienna, VA 703-893-8400

International SalesLocations

Berlin, Germany (49) 30 25 49 31 86Bologna, Italy (39) 51 522579Brescia, Italy (39) 30 2421791Brussels, Belgium (32) 2 725 74 25Cape Town, South Africa (27) 21 686 1430Dublin, Ireland (353) 1 475 4172Durban, South Africa (27) 31 83 6611Dusseldorf, Germany (49) 2131 9191 0Frankfurt, Germany (49) 6196 4 72 80Hamburg, Germany (49) 4069 69 610Hong Kong, Taikooshing (852) 2839 9600Johannesburg, South Africa (27) 11 807 5300Leatherhead, Surrey, UK (44) 1372 36 0000

EMC Navisphere Analyzer Version 5.X Administrator’s Guide s-3


Lyon, France (33) 7 860 1330Marseille, France (33) 4292 2549Melbourne, Australia (61) 3 653 9519Milan, Italy (39) 02 409081Munich, Germany (49) 89 14 31 320Nagoya, Japan (81) 52 223 1900Newton Le Willows, Merseyside, UK (44) 1942 275 511Nieuwegein, Netherlands (31) 03 6055777Nuernberg, Germany (49) 911 2379260Osaka, Japan (81) 6 373 8300Padua, Italy (39) 49 8235853Paris, France (33) 1 3082 5100Pretoria, South Africa (27) 12 663 6635Rome, Italy (39) 06 8552116Schlieren, Switzerland (411) 743 4949Seoul, Korea (82) 2 539 4455Singapore (65) 338 9265Stuttgart, Germany (49) 7152 979340Sydney, Australia (61) 2 922 7888Toulouse, France (33) 6131 6262Turin, Italy (39) 11 746527Vienna, Austria (43) 1 599 99 627

s-4 EMC Navisphere Analyzer Version 5.X Administrator’s Guide


EMC Navisphere Analyzer Version 5.X Administrator’s Guide i-1

Aaccessible storage system, defined 3-6accessible storage systems 3-6adapter port, <italic>see HBA portAgent

about 1-5poll interval 4-38

alignment, data 5-14Analyzer

architecture 1-4charts, introduced 4-8dependencies 1-4environments 1-2hardware and software requirements for 2-3menu option 3-22, 4-2starting 2-5

application configurationdefault configuration file, changing name

and location of 3-28defined 3-27saving to

custom configuration file 3-29default configuration file 3-28

Application icon 3-20architecture, Analyzer 1-4audience for manual iii-ixAutomatic Polling option

default value 2-8defined 2-8required for analyzing performance 2-5setting 2-9

Bbasic storage component icons 3-14bottlenecks 5-4

eliminating 5-6buttons

Faults 3-24Firmware Download 3-24Help 3-24Poll 3-24

Ccache

page size 5-10performance 5-9

monitoring 5-10snapshot 3-18

changing name and location of 3-28chart

Analyzer, introduced 4-8IO Size Distribution 4-20Performance Detail 4-13Performance Summary 4-13

chart data, exporting 4-30chart menu

Performance summary 4-11charts

resizing, zooming 4-6comma separated values (csv) 4-30component objects, selecting 4-10components, Analyzer 1-4controlling logger operation 4-39csv (comma-separated values) 4-30

Index

EMC Navisphere Analyzer Version 5.X Administrator’s Guidei-2

Index

Ddata

alignment on stripe and performance 5-14date, to start/stop logging 4-40dependencies, Analyzer 1-4Detailed View window

icons 3-6workspace 3-6

device menuPerformance summary 4-11

dialog boxesEnterprise Storage 3-25Host Selection 2-10, 2-12Main 3-20User Options 2-9, 3-28

directorylog file 4-35

dirty pagesmeaning 5-11

diskfactors in performance 5-8menu 4-2

disksicons for 3-17

documentation conventions iii-x

EEnterprise management 2-5Enterprise Storage dialog boxes 3-25

activating 3-26closing 3-26Equipment tab 3-2Hosts tab 3-2opening 3-26Storage tab 3-2

Equipment tree 3-2exporting chart data 4-30

FFaults button 3-24FC4700

icon for 3-9File menu 3-21Filter By filter, Enterprise Storage dialog box 3-26

Filter For filter, Enterprise Storage dialog box 3-26

filters for displaying managed storage systems 3-26

Firmware Download button 3-24forced flush rate 5-10format conventions iii-x

HHBA port, icon for 3-8Help

button 3-24menu 3-23

high watermark flush rate 5-11hit ratio

read cache 5-12write cache 5-11

hostperformance issues 5-14selecting for logging 4-33selecting to manage 2-10, 2-12

host, <italic>see serverHost File Path option

default value 2-8defined 2-8setting 2-9

Host Selection dialog box 2-10, 2-12opening 2-10, 2-12using to select storage systems for

management 2-13hosts

tree 3-2Hosts tree 3-2hot spare

icon for 3-16

Iicons 3-7

Application 3-20basic storage components 3-14color of 3-7disks 3-17HBA port 3-8letter in 3-7LUNs 3-14MirrorView components 3-17

i-3EMC Navisphere Analyzer Version 5.X Administrator’s Guide

Index

multiple storage systems 3-13RAID Groups 3-16remote mirror 3-17server 3-8snapshot copy LUNs 3-18SnapView components 3-17SPs 3-16Storage Group 3-14, 3-18storage system 3-9storage-system components 3-13

inaccessible storage system, defined 3-6individual disk

icon for 3-16IO Size Distribution chart 4-20

introduced 4-20

Jjpeg (.jpg) images 4-22jpg (jpg) images 4-30

Llight loads 5-4locality of reference 5-4log file

about 4-23directory 4-35playing back 4-3properties 4-34viewing 4-3

loggercontrolling operation 4-39running 4-41setting up 4-32starting 4-23stopping (close) 4-41

loggingscheduling 4-40

low watermark flush rate 5-11LUNs (logical units)

icon for snapshot copy 3-18icons for 3-14menu 4-2unowned

defined 3-16

MMain menu

View menu 3-22Main window 3-20

application configurationcustom 3-29default, restoring current configuration

to 3-29default configuration file 3-28defined 3-27saving to

custom configuration file 3-29default configuration file 3-28

Application icon 3-20Enterprise Storage dialog boxes 3-25menu bar 3-21

File menu 3-21Help menu 3-23Operations menu 3-22

Navispheremenu bar 3-23

status bar 3-27toolbar 3-23

buttons 3-24management station

defined 1-2Manager

Automatic Polling optionsetting 2-9

Polling Interval optionsetting 2-9

managing storage systems onselected servers on subnets 2-13servers whose names you know 2-13

menuPerformance Summary 4-11Performance Summary chart 4-11

menu bar 3-21File menu 3-21Help menu 3-23Navisphere 3-23Operations menu 3-22View menu 3-22

menu options 3-23menus

for storage system 3-11


Index

for storage-system components 3-18disk 4-2LUN 4-2SPs 4-2

on menu barFile 3-21Help 3-23Operation 3-22View 3-22Window 3-23

MirrorViewcomponent icons 3-17storage components for 3-13

NNavisphere

Automatic Polling optiondefault value 2-8

Host File Path optiondefault value 2-8defined 2-8setting 2-9

Main window 3-20Polling Interval option

default value 2-8Network Time Out option

defined 2-8Network Timeout option


OOperations menu 3-22option

default value 2-8

Ppage

dirty 5-11size, cache 5-10

performancedata, understanding, see also Chapter 4 5-1logger, running 4-23problem, identifying 5-5

properties 4-24Performance Detail chart

about 4-13introduced 4-13

Performance Loggersetting up 4-32

Performance Logger, See also loggerPerformance Summary chart

about 4-13introduced 4-9

pictures from charts (Export Chart Graph) 4-22pictures from charts (exporting) 4-30poll

time needed for 4-38Poll button 3-24poll, automatic, defined 2-8Polling Interval option


printing graphic displays 4-30properties

log file 4-34performance 4-24

Qqueue length property

using 5-3

RRAID 0

LUN, icon for 3-15RAID 1

LUN, icon for 3-15RAID 1/0

LUN, icon for 3-15RAID 3

LUNicon for 3-14

RAID 5LUN, icon for 3-14

RAID factors in performance 5-8RAID Groups

icons for 3-16read cache

performance 5-9

i-5EMC Navisphere Analyzer Version 5.X Administrator’s Guide

Index

remote mirrorsicons for 3-17

resizing charts 4-6restoring current configuration to 3-29running the logger 4-41

Ssaving logger settings 4-41scheduling logging 4-40server

icon for 3-7, 3-8performance issues 5-14selecting for logging 4-33selecting to manage 2-10, 2-12throttle 5-14

snapshot cache 3-18snapshot copy LUNs

icon for 3-18SnapView

storage component icons 3-17storage components for 3-13

Sorting display of devices 4-12sorting values 4-12SP (storage processor)

icons for 3-16menu 4-2snapshot cache 3-18

starting Analyzer 2-5status bar 3-27

information fields 3-27stopping logger (close) 4-41Storage Groups

icons for 3-14, 3-18storage processor, <italic>see SP (storage

processor)storage system

accessible 3-6automatic polling

interval, defined 2-8interval, setting 2-9

component menus, displaying 3-18filters for displaying 3-26hosts, selecting 2-10, 2-12icons 3-9

for components 3-13for multiple storage systems 3-13

inaccessible 3-6managing on

selected servers on subnets 2-13servers whose name you know 2-13

menu 3-11for components 3-18

name, custom, assigning 3-11trees for 3-2unsupported 3-6

storage systemsaccessible 3-6icons 3-9inaccessible 3-6unsupported 3-6

Storage tree 3-2stripe size factors and performance 5-8

Tthrottle, server 5-14time, to start/stop logging 4-40toolbar 3-23toolbar buttons 3-23tools 3-23trees 3-13

defined 3-2icons 3-7<italic>see also Equipment tree or Hosts tree

or Storage tree

Uunowned LUNs

defined 3-16icon for 3-16

unsupported storage system, defined 3-6unsupported storage systems 3-6User Options dialog box 2-9, 3-28user options for Navisphere

default values 2-8defined 2-8setting 2-9

utilization propertyusing 5-3, 5-5, 5-7

VView menu options 3-22


Index

Wwatermark

flush rates 5-11Window menu 3-23workload

light 5-4write cache

counters, using 5-10performance 5-9size 5-9

Zzooming charts 4-6

emc enterprise storage · emc enterprise storage emc corporation 171 south street, hopkinton, ma...

Documents