Dell IoMemory VSL 3.2.15 User Guide For Linux 2017-04-11

Dell Fusion ioMemory VSL 3.2.
15
USER GUIDE FOR LINUX
Tuesday, April 11, 2017

Table of Contents
Table of Contents 2
Dell Fusion ioMemory VSL 3.2.15 User Guide for Linux 7
Legal Notices 7
CAUTION— Safety Instructions 8
Protecting Against Electrostatic Discharge 8
SAFETY— General 8
SAFETY— When Working Inside Your System 9
Introduction 10
Overview 10
About the ioMemory Platform 10
Performance 10
Endurance 10
Reliability 11
Fusion ioMemory VSL Software Installation 12
Installation Overview 12
Installing RPM Packages 13
Loading the ioMemory VSL Driver 19
Controlling Driver Loading 20
Using the init Script 20
Mounting Filesystems when Using the init Script 20
Using udev 21
On either udev or init script systems 22
Fusion ioMemory devices and Multipath Storage 22

Handling Driver Unloads 22
Upgrading the Firmware 23
Command-line Interface 23
Optional GUI - Dell ioSphere software 24
Configuration 25
Setting the ioMemory VSL Options 25
One-Time Configuration 25
Persistent Configuration 25
Enabling PCIe Power Override 25
Enabling the Override Parameter 27
Determine Serial Number(s) 27
Setting the Parameter in VSL 3.2.15 28
Virtual Controller Configuration 29
Supported Devices 30
Splitting Controllers 30
Merging Controllers 31
Using the Device as Swap 32
Setting the Amount of Preallocated Memory 33
Using the Logical Volume Manager 34
Mounting Filesystems when Using the init Script 34
Configuring RAID using the Logical Volume Manager 34
Creating a Striped Volume (RAID 0) Using LVM 35
Creating a Mirrored Volume (RAID 1) Using LVM 35
Configuring RAID Using mdadm 36
Mounting Arrays 36
RAID 0 37

Making the Array Persistent (Existing after Restart) 37
RAID 1 38
RAID 10 38
Building a RAID10 Across Multiple Devices 38
Discard (TRIM) Support 39
Discard (TRIM) on Linux 39
Performance and Tuning 41
Disable CPU Frequency Scaling 41
Limiting ACPI C-States 41
Setting ACPI C-State Options 41
C-States Under Linux 42
Setting NUMA Affinity 42
Setting the Interrupt Handler Affinity 42
Linux IRQ Balancing 42
Monitoring and Managing Devices 44
Management Tools 44
Stand-alone Tools 44
Data-source Tools 44
Example Conditions to Monitor 45
Device Status 45
Temperature 45
Health Reserves Percentage 46
Write (Health Reserves) Status 47
Device LED Indicators 47
Maintenance 49
Uninstalling the Software 49

Uninstalling the Fusion ioMemory VSL Utilities and Other Support Packages 49
Uninstalling 2.x Support Packages 49
Uninstalling 3.x Support Packages 49
Uninstalling the Fusion ioMemory VSL RPM Package 49
Unloading the Software Driver 50
Upgrading the Kernel 50
Disabling the ioMemory VSL Software 50
Disabling Auto-Attach 51
Unmanaged Shutdown Issues 51
Improving Rescan Times 51
Default Fast Rescan 51
Faster Rescans Using Module Parameters 52
RMAP Parameter 52
RSORT Parameter 53
Appendix A - Command-line Utilities Reference 56
fio-attach 57
fio-beacon 58
fio-bugreport 58
fio-detach 60
fio-firmware 60
fio-format 61
fio-pci-check 63
fio-status 63
fio-sure-erase 66
Clear Support 67
Purge Support 68

fio-update-iodrive 68
Appendix B - Monitoring the Health of Devices 71
Health Metrics 71
Health Monitoring Techniques 71
Software RAID and Health Monitoring 72
Appendix C - Using Module Parameters 73
Appendix D - NUMA Configuration 75
About NUMA Architecture 75
Using the numa_node_forced_local Parameter 75
Using the numa_node_override Parameter 75
Determining the Device ID 76
numa_node_override Parameter 76
Advanced Configuration 77
Appendix E - Upgrading Devices from VSL 2.x to 3.x 78
Upgrade Procedure 78
Dell Support 83
Contacting Dell 83

Dell Fusion ioMemory VSL 3.2.15 User Guide for Linux
Legal Notices
The information contained in this document is subject to change without notice.
® is a trademark of Western Digital

© 2017 Western Digital Corporation or its affiliates. All rights reserved. SanDisk
Corporation or its affiliates, registered in the United States and other countries. Fusion ioMemory, VSL, and others
are trademarks of Sandisk Enterprise IP LLC. Other brand names that may be mentioned herein are for
identification purposes only and may be the trademarks of their respective holder(s).
Reproduction of these materials in any manner whatsoever without the written permission of Dell Inc. is strictly
forbidden.
Trademarks used in this text: Dell™, the DELL logo, PowerEdge™ are trademarks of Dell Inc.
Other trademarks and trade names may be used in this publication to refer to either the entities claiming the
marks and names or their products. Dell Inc. disclaims any proprietary interest in trademarks and trade names
other than its own.
Part Number: 80-11-71088

Published: Tuesday, April 11, 2017
The following information boxes are used throughout the user guide to present important information:
NOTE-
Indicates a special note.
ATTENTION!
Indicates cautions to take into consideration when using the device and software.
WARNING!
Indicates warnings to take into consideration when using the device and software.

CAUTION— Safety Instructions
Use the following safety guidelines to help ensure your own personal safety and to help protect your system and
working environment from potential
NOTE-
For complete information on U.S. Terms and Conditions of Sale, Limited Warranties and Returns, Export
Regulations, Software License Agreement, Safety, Environmental and Ergonomic Instructions, Regulatory
Notices, and Recycling Information, see the Safety, Environmental and Regulatory Information (SERI), End
User License Agreement (EULA), and Warranty and Support Information (WSI) that shipped with your
system.
Protecting Against Electrostatic Discharge

Electrostatic discharge (ESD) events can harm electronic components inside your system. Under certain conditions,
ESD may build up on your body or an object, such as a peripheral, and then discharge into another object, such as
your system. To prevent ESD damage, you must discharge static electricity from your body before you interact with
any of your system's internal electronic components, such as a memory module. You can protect against ESD by
touching a metal grounded object (such as an unpainted metal surface on your system's I/O panel) before you
interact with anything electronic. When connecting a peripheral (including handheld digital assistants) to your
system, you should always ground both yourself and the peripheral before connecting it to the system.
Additionally, as you work inside the system, periodically touch an I/O connector to remove any static charge your
body may have accumulated.
You can also take the following steps to prevent damage from electrostatic discharge:
l When unpacking a static-sensitive component from its shipping carton, do not remove the component from
the antistatic packing material until you are ready to install the component. Just before unwrapping the
antistatic package, be sure to discharge static electricity from your body.
l When transporting a sensitive component, first place it in an antistatic container or packaging.
l Handle all electrostatic sensitive components in a static-safe area.
SAFETY— General
l Observe and follow service markings. Do not service any product except as explained in your user
documentation. Opening or removing covers that are marked with the triangular symbol with a lightning
bolt may expose you to electrical shock. Components inside the compartments must be serviced only by a
trained service technician.
l If any of the following conditions occur, unplug the product from the electrical outlet, and replace the part
or contact your trained service provider:
o The power cable, extension cable, or plug is damaged.
o An object has fallen in the product.
o The product has been exposed to water.
o The product has been dropped or damaged.
o The product does not operate correctly when you follow the operating instructions.
o Use the product only with approved equipment.

l Operate the product only from the type of external power source indicated on the electrical ratings label. If
you are not sure of the type of power source required, consult your service provider or local power
company.
l Handle batteries carefully. Do not disassemble, crush, puncture, short external contacts, dispose of in fire or
water, or expose batteries to temperatures higher than 60° Celsius (140° Fahrenheit). Do not attempt to
open or service batteries; replace batteries only with batteries designated for the product.
SAFETY— When Working Inside Your System

Before you remove the system covers, perform the following steps in the sequence indicated.
ATTENTION!
Except as expressly otherwise instructed in Dell documentation, only trained service technicians are
authorized to remove the system cover and access any of the components inside the system.
NOTE-
To help avoid possible damage to the system board, wait 5 seconds after turning off the system before
removing a component from the system board or disconnecting a peripheral device.
1. Turn off the system and any connected devices.

2. Disconnect your system and devices from their power sources. To reduce the potential of personal injury
or shock, disconnect any telecommunication lines from the system.
3. Ground yourself by touching an unpainted metal surface on the chassis before touching anything inside the
system.
4. While you work, periodically touch an unpainted metal surface on the chassis to dissipate any static
electricity that might harm internal components.
In addition, take note of the following safety guidelines when appropriate:
l When you disconnect a cable, pull on its connector or on its strain-relief loop, not on the cable itself. Some
cables have a connector with locking tabs. If you are disconnecting this type of cable, press in on the locking
tabs before disconnecting the cable. As you pull connectors apart, keep them evenly aligned to avoid
bending any connector pins. Also, when you connect a cable, make sure both connectors are correctly
oriented and aligned.
l Handle components and cards with care. Do not touch the components or contacts on a card. Hold a card
by its edges or by its metal mounting bracket. Hold a component such as a microprocessor chip by its edges,
not by its pins.

Introduction
Overview
Congratulations on your purchase of a Dell® solid-state storage device. This guide explains how to install,
troubleshoot, and maintain the Fusion ioMemory™ VSL® software for your Fusion ioMemory devices.
NOTE-
Throughout this manual, when you see a reference to any Fusion ioMemory device, you may substitute
your particular device(s) from the Supported Devices list located in the Dell Fusion ioMemory VSL Release
Notes.
ATTENTION!
Products with Multiple Devices
Some products, such as a Fusion ioMemory ioDrive Duo device, are actually comprised of multiple Fusion
ioMemory devices. If your product consists of multiple Fusion ioMemory devices, you will manage each
Fusion ioMemory device as an independent device.
For example, if you have a Fusion ioMemory ioDrive Duo device, you can independently attach, detach,
and/or format each of the two Fusion ioMemory devices. Each of the two devices will be presented as an
individual device to your system.
About the ioMemory Platform

The ioMemory platform combines Fusion ioMemory VSL software (VSL stands for Virtual Storage Layer) with
ioMemory hardware to take enterprise applications and databases to the next level.
Performance
The ioMemory platform provides consistent microsecond latency access for mixed workloads, multiple gigabytes
per second access and hundreds of thousands of IOPS from a single product. The sophisticated ioMemory
architecture allows for nearly symmetrical read and write performance with best-in-class low queue depth
performance, making the ioMemory platform ideal across a wide variety of real world, high-performance
enterprise environments.
The ioMemory platform integrates with host system CPUs as flash memory to give multiple (and mostly idle)
processor cores, direct and parallel access to the flash. The platform's cut-through architecture gives systems more
work per unit of processing, and continues to deliver performance increases as CPU power increases.
Endurance
The ioMemory platform offers best-in-class endurance in all capacities, which is crucial for caching and write-heavy
databases and applications.

Reliability
The ioMemory platform eliminates concerns about reliability like NAND failures and excessive wear. The all-new
intelligent, self-healing feature called Adaptive Flashback provides complete, chip-level fault tolerance. Adaptive
Flashback technology enables an ioMemory product to repair itself after a single chip or a multi-chip failure without
interrupting business continuity.

Fusion ioMemory VSL Software Installation
Before continuing with the installation of this software, please read the following:
1. Ensure that your operating system is included in the list of supported operating systems contained in the
Dell Fusion ioMemory VSL Release Notes for this release.
2. Before installing the Fusion ioMemory VSL software, make sure you have properly installed the Fusion
ioMemory device(s). Refer to the Dell Fusion ioMemory Hardware Installation Guide for full details and
hardware requirements.
ATTENTION!
Every Fusion ioMemory device in a system must be upgraded to the appropriate firmware.
For example, if you have a system running Fusion ioMemory VSL software version 2.3.1 with Fusion
ioMemory devices previously installed, and you want to install new Fusion ioMemory ioDrive2 devices
(that require the latest version of the firmware), then you will need to upgrade all of the existing devices
with firmware that supports this version of the Fusion ioMemory VSL software. Follow the upgrade path
in the Dell Fusion ioMemory VSL Release Notes to determine the upgrade sequence.
WARNING!
Upgrade Previous Devices First
If you have Legacy Fusion ioMemory devices configured for Fusion ioMemory VSL software version 2.x or
earlier, you must upgrade their firmware before installing new devices in the system. See Upgrading
Devices from VSL 2.x to 3.x on page 78 for the upgrade instructions.
NOTE-
All commands require administrator privileges. Use sudo or log in as "root" to run the install.
Installation Overview
1. Download the latest version of the software at http://dell.fusionio.com .
2. If you are installing version 3.2.15 of Fusion ioMemory VSL software on a system with Fusion ioMemory
ioDrive devices configured with firmware for Fusion ioMemory VSL software version 2.x, you must
carefully follow the instructions in the Upgrading Devices from VSL 2.x to 3.x on page 78. (Follow those
instructions instead of the normal installation instructions.)
3. If you have a previous version of the Fusion ioMemory VSL software installed, you will need to uninstall the
Fusion ioMemory VSL software and the utilities. See Uninstalling the Software on page 49 for instructions.
Once you have uninstalled the packages, return to this page.

ATTENTION!
Kernel Upgrades
If you ever plan to upgrade the kernel when the Fusion ioMemory VSL software is installed, you
must follow the procedure for Upgrading the Kernel on page 50.
4. Install the latest version of the Fusion ioMemory VSL software. The installation instructions will help you
determine which of these package types is supported by your kernel:
l A pre-compiled binary package
l A source-to-build package
5. Install utilities and management software (included in driver installation instructions).
6. Follow the instructions on Loading the ioMemory VSL Driver on page 19 and consider Setting the ioMemory
VSL Options on page 25.
7. Determine if you need to upgrade the firmware to the latest version, see Upgrading the Firmware on page
23.
Installing RPM Packages

To install the Linux Fusion ioMemory VSL software and utilities:
1. You will need to install a version of the Fusion ioMemory VSL software that is built for your kernel. To
determine what kernel version is running on your system, use the following command at a shell prompt:
$ uname -r
2. Compare your kernel version with the binary versions of the software available at http://dell.fusionio.com .
l If there is a binary version of the software that corresponds to your kernel version, download that.
For example:
iomemory-vsl-<kernel-version>-<VSL-version>.x86_64.rpm
l If there is no binary version of the software corresponding to your kernel, download the source
package. For example:
iomemory-vsl-<VSL-version>.src.rpm
NOTE-
Exact package names may vary, depending on software and kernel version chosen.
Use the source package that is made for your distribution. Source packages from other distributions
are not guaranteed to work.

3. Download the support RPM packages you need from http://dell.fusionio.com . These packages provide
utilities, firmware, and other files.
ATTENTION!
libvsl Packaging Change. The SDK libraries needed for 64-bit management tools (formerly
downloaded separately as libvsl-<version>.x86_64.rpm), are now packaged in the
iomemory-vsl-<kernel-version>-<VSL-version>.x86_64.rpm (or in the
iomemory-vsl-<VSL-version>.src.rpm if you are building the package). However, the
SDK libraries needed for working with 32-bit management applications must still be downloaded
separately (see table below).
For more information on available management tools, please refer to Monitoring and Managing
Devices on page 44 .
ATTENTION!
Unified Utilities. If you plan to run the 4.2.1 (or newer) Unified Utilities on a VSL 3.2.15 server,
you will need to download both the VSL 3.2.15 Release utilities package fio-util-<VSL-
version>.x86_64.rpm, and the VSL 4.2.1 (or newer) Release utilities package fio-util-
<VSL-version>.x86_64.rpm from http://dell.fusionio.com .
Be aware that only certain operating systems installed with VSL 3.2.15 can use the 4.2.1 (or
newer) utilities. For the list of operating systems compatible with Unified Utilities on VSL 3.2.15
systems, refer to the 3.2.15 Dell Fusion ioMemory VSL Release Notes.
For Unified Utilities installation, after you install the 3.2.15 Fusion ioMemory VSL and utilities
packages and upgrade the firmware, you will then install the 4.2.1 (or newer) fio-util-
<VSL-version>.x86_64.rpm package on your VSL 3.2.15 server. This installation will
replace your VSL 3.2.15 utilities with the VSL 4.2.1 (or newer) Unified Utilities. See the end of this
section for more information.

Examples:
Package What Is Installed

ioMemory VSL utilities – Strongly Recommended
NOTE-
fio-util-<VSL-
If you plan to run the 4.2.1 (or newer) Unified Utilities on a VSL
version>.x86_
3.2.15 server, you will need to download both the 3.2.15 fio-
64.rpm
util-<VSL-version>.x86_64.rpm for the VSL 3.2.15
Release, and the 4.2.1 (or newer) fio-util-<VSL-
version>.x86_64.rpm for the VSL 4.2.1 (or newer) Release
Version (preferably into separate folders or locations).
This package installs the fio-firmware utility, which reports the version
and revision of the firmware on all Fusion ioMemory devices without
fio-preinstall-
<VSL-version>.x86_ requiring you to load the ioMemory VSL software driver. – Optional
64.rpm
For fio-firmware usage information, please refer to fio-firmware on
page 60.
fio-firmware-dell_
Firmware archive package, installs the firmware archive file in a specific
<version>.<date>-
location (/usr/share/fio/firmware) – Optional
1.0.noarch.rpm
fio-firmware-dell_
Firmware archive file (not an installer package) used to upgrade the
<version>.<date>
firmware, make note of where you store this file – Optional
.fff
SDK libraries needed for working with 32-bit management applications –

Optional and only available for certain distributions.
lib32vsl- ATTENTION!
<version>.i386.rpm
You must have a full set of 32-bit system libraries installed (32-bit
compatibility layer installed on your 64-bit system) before you can
install this package.
fio-common-<VSL-
version>.x86_ Files required for the init script – Strongly Recommended
64.rpm
fio-sysvinit-<VSL-
Init script – Strongly Recommended, see Loading the ioMemory VSL Driver
version>.x86_
on page 19 for more information.
64.rpm
4. Change to the directory where you downloaded the installation packages.

5. If needed, build the Fusion ioMemory VSL software from source:
l If you downloaded a binary version of the software: skip the Building the Software from Source
instructions below and continue to the next step.
l If you downloaded the software source package: follow these Building the Software from Source
instructions:
Building the Software from Source

You only need to follow these additional instructions if you downloaded the source package.
a. Install the prerequisite files for your kernel version.
NOTE-
Some of the prerequisite packages may already be in the default OS installation.
If your system is not configured to get packages over the network, then you may
need to mount your install CD/DVD.
o On RHEL/OEL 5/6/7, you need kernel-devel, rpm-build, GCC4, make, and

rsync.
$ yum install rsync tar gcc make kernel-devel-`uname -

r` rpm-build
This command will force yum to download the exact versions for your kernel. If
the exact versions are no longer available in the repository, then you will have to
manually search for them on the Internet.
ATTENTION!
If you are using the UEK kernel on OEL, replace kernel-devel with:
kernel-uek-devel
o On SLES 10/11/12 you need kernel-syms, make, rpm, rpm-build, GCC4,

rsync, and kernel-source.
$ zypper install kernel-syms make rpm rpm-build gcc

rsync kernel-source
b. To build an RPM installation package for the current kernel, navigate to the directory with
the downloaded source RPM file and run this command:
$ rpmbuild --rebuild iomemory-vsl-<VSL-version>.src.rpm

ATTENTION!
If your kernel is an Unbreakable Enterprise Kernel (UEK), you may also need to
use the --nodeps option.
When using a .rpm source package for a non-running kernel, run this command:
$ rpmbuild --rebuild --define 'rpm_kernel_version <kernel-

version>' iomemory-vsl-<VSL-version>.src.rpm
c. The new RPM package is located in a directory that is indicated in the output from the
rpmbuild command. To find it, look for the Wrote line. In the following example, the
RPM packages are located in the /usr/src/redhat/RPMS/x86_64/ directory.
...
Processing files: iomemory-vsl-source-<version>-1.0.x86_
64.rpm
Requires(rpmlib): rpmlib(PayloadFilesHavePrefix) <= 4.0-1
rpmlib(CompressedFileNames) <= 3.0.4-1
Obsoletes: iodrive-driver-source
Checking for unpackaged file(s): /usr/lib/rpm/check-files
/var/tmp/iomemory-vsl-<version>-root
Wrote: /usr/src/redhat/RPMS/x86_64/iomemory-vsl-2.6.18-
128.el5-<version>-1.0.x86_64.rpm
Wrote: /usr/src/redhat/RPMS/x86_64/iomemory-vsl-source-
<version>-1.0.x86_64.rpm
In this example, iomemory-vsl-2.6.18-128.el5-<version>-1.0.x86_

64.rpm is the package you will use.
d. Copy your custom-built software installation RPM package into the directory where you
downloaded the installation packages and navigate to that directory.
e. Continue to the next step.
6. Enter the following command to install the custom-built software package. Use the package name that you
just copied/downloaded into that directory.
$ rpm -Uvh iomemory-vsl-<kernel-version>-<VSL-version>.x86_64.rpm
7. Enter the following commands to install the support files:
ATTENTION!
Ensure that the 3.2.15 fio-util-<VSL-version>.x86_64.rpm package (and not the
4.2.1 (or newer) fio-util-<VSL-version>.x86_64.rpm package) is located in the

directory where you will be executing the commands below.
$ rpm -Uvh lib*.rpm

rpm -Uvh fio*.rpm
NOTE-
If the installation of the 3.2.15 fio-util-<VSL-version>.x86_64.rpm package fails due
to missing dependencies, you will need to install the missing packages before you run the install
command again, for example:
$ yum install lsof pciutils
The Fusion ioMemory VSL software and utilities are installed to the following locations:
Package Type Installation Location

Fusion ioMemory VSL /lib/modules/<kernel-version>/extra/fio/iomemory-
software vsl.ko
Utilities /usr/bin
NOTE-
You may also install the Dell ioSphere software (optional GUI management software). Dell ioSphere
software and documentation are available as a separate download.
Once the packages are installed, continue to Loading the ioMemory VSL Driver on page 19.
ATTENTION!
Unified Utilities. After you install 3.2.15 Fusion ioMemory VSL and 3.2.15 utilities, and upgrade the
firmware, you can install the 4.2.1 (or newer) Unified Utilities if so desired.
To install the Unified Utilities, you will install the 4.2.1 (or newer) fio-util-<VSL-version>.x86_
64.rpm for VSL 4.2.1 (or newer) Release Version on your VSL 3.2.15 server using the following
command:
$ rpm -Uvh fio-util-4.2.x.<buildversion>.x86_64.rpm
This command will replace your VSL 3.2.15 utilities with the VSL 4.2.1 (or newer) Unified Utilities. No
reboot is required.

ATTENTION!
If the installation of the 4.2.1 (or newer) fio-util-<VSL-version>.x86_64.rpm
package fails due to missing dependencies, you will need to install the missing packages before
you run the install command again, for example:
$ yum install lsof pciutils
For information on using the Unified Utilities on VSL 3.2.15 systems, including command differences
between 3.2.15 and 4.2.1 (or newer) utilities, refer to the 3.2.15 Dell Fusion ioMemory VSL Release Notes.
Loading the ioMemory VSL Driver

To load the Fusion ioMemory VSL driver:
Run this command:
# modprobe iomemory-vsl
NOTE-
The Fusion ioMemory VSL driver automatically loads at system boot. The Fusion ioMemory device is now
available to the OS as /dev/fiox, where x is a letter (i.e., a, b, c, etc.).
To confirm the Fusion ioMemory device is attached, run the fio-status utility from the command line. The
output lists each drive and its status (attached or not attached).
ATTENTION!
If the Fusion ioMemory device is not automatically attaching, check the /etc/modprobe.d files to
ensure that the auto_attach option is turned on (set to 1).
l For the modprobe iomemory-vsl command to work on SLES 10 systems, you must edit the
/etc/init.d/iomemory-vsl file's init info and change udev to boot.udev. The file should look
like this:
### BEGIN INIT INFO

# Provides: iomemory-vsl
# Required-Start: boot.udev
l On SLES systems, you must also allow unsupported modules for the modprobe iomemory-vsl
command to work.
o SLES 12 and SLES 11 Update 2: Please refer to the /etc/modprobe.d/iomemory-vsl.conf

file for the most current instructions on enabling unsupported modules for your distribution.

o SLES 10 SP4: Modify the /etc/sysconfig/hardware/config file so the LOAD_
UNSUPPORTED_MODULES_AUTOMATICALLY sysconfig variable is set to yes, for example:
LOAD_UNSUPPORTED_MODULES_AUTOMATICALLY=yes
Controlling Driver Loading

You can control driver loading either through udev or the init script.
udev - In newer Linux distributions, you can rely on the udev device manager to automatically find and load
drivers for the installed hardware at boot time, though udev can be disabled and the init script used in
nearly all cases.
init script - On systems where udev loading of the driver doesn't work or is disabled, you can use the init
script to load the Fusion ioMemory VSL driver. We recommend using the init script to load the VSL driver if
you are managing a RAID array using LVM, mdadm, or Veritas Storage Foundation. For older Linux
distributions (such as SLES 10) without udev functionality, users must rely on a boot-time init script to load
needed drivers.
Using the init Script
The init script can be enabled to load the driver at boot. On some distros it may be enabled by default.
NOTE-
The init Script is part of the fio-sysvinit package, which must be installed before you can enable it.
You must also install the fio-common package.
l To enable driver loading in the init script, use the following command:
# chkconfig --add iomemory-vsl
l To disable driver loading in the init script, use the following command:
# chkconfig --del iomemory-vsl
The Fusion ioMemory VSL software install process places an init script in /etc/init.d/iomemory-vsl. In
turn, this script uses the setting options found in the options file in /etc/sysconfig/iomemory-vsl. The
options file must have ENABLED set (non-zero) for the init script to be used:
ENABLED=1
The options file contains documentation for the various settings—including KILL_PROCS_ON_UMOUNT, which is
discussed further in the section Handling Driver Unloads on page 22.
Mounting Filesystems when Using the init Script
There may be timing or other issues that prevent the filesystem on a Fusion ioMemory ioDrive device from coming
up automatically when using the standard Linux driver load process. This means that using the standard method for

mounting filesystems (/etc/fstab) may not work reliably for filesystems hosted on the Fusion ioMemory
device.
Instead, to correctly set up auto-mounting of a filesystem hosted on an Fusion ioMemory device, follow the
procedure below:
1. Add the filesystem mounting command to /etc/fstab as normal.

2. Add the 'noauto' option and the '0 0' flag to /etc/fstab, as shown in the two following sample
entries.
/dev/fioa /mnt/fioa ext3 defaults,noauto 0 0

/dev/fiob1 /mnt/ioDrive ext3 defaults,noauto 0 0
(where the a in fioa can be a, b, c, etc., depending on how many Fusion ioMemory devices you have
installed in the system). The number of lines you add will depend on how many Fusion ioMemory devices
/partitions you have installed in the system.
WARNING!
Failure to add 'noauto 0 0' to fstab may cause a boot failure.
3. To configure the init script to automatically mount these drives after the driver is loaded, and unmount
them before the driver is unloaded, you should add a list of mount points to the options file
/etc/sysconfig/iomemory-vsl using the procedure documented there.
For the filesystem mounts shown in the earlier example, the line in the options file would look like this:
MOUNTS="/mnt/fioa /mnt/iodrive"
Using udev
On systems that rely on udev to automatically load drivers, users can modify a Fusion ioMemory VSL software
options file if they want to prevent udev from auto-loading the ioMemory VSL at boot time. To do this, locate and
edit the /etc/modprobe.d/iomemory-vsl.conf file that already has the following line:
# blacklist iomemory-vsl
To disable automatic driver loading by udev, remove the "#" from the line and save the file.
With the blacklist command in place, restart Linux. The Fusion ioMemory VSL driver will not be loaded by udev.
NOTE-
If the driver is already in the initrd, you will need to rebuild the initrd for the change to take effect.
To restore the udev-loading of the driver, replace the "#" to comment out the line.

On either udev or init script systems
For diagnostic or troubleshooting purposes on both udev and init script systems, users can disable the loading of
the driver at boot time, which in turn prevents the auto-attach process.
For example, if the Fusion ioMemory VSL software is preventing the system from booting normally, you can
prevent the Fusion ioMemory VSL driver from loading at all by appending the following parameter at the kernel
command line of your boot loader:
iomemory-vsl.disable=1
The driver will not load until the system is rebooted. However, this method is generally not preferred for diagnostic
situations because it limits the amount of troubleshooting you can perform.
For many troubleshooting scenarios, disabling auto-attach alone is sufficient. You can disable auto-attach for a single
load only, or in a persistent configuration.
l To disable the auto-attach functionality for a single load of the Fusion ioMemory VSL driver, append the
following parameter at the kernel command line of your boot loader:
iomemory-vsl.auto_attach=0
NOTE-
The above command will disable auto-attach even if the driver is loaded by the initrd.
l To disable the auto-attach functionality as a persistent configuration, follow the steps in the section Disabling
Auto-Attach on page 51.
Fusion ioMemory devices and Multipath Storage

If you are using Fusion ioMemory devices along with multipath storage, you must blacklist the Fusion ioMemory
devices to prevent device-mapper from trying to create a dm-device for each Fusion ioMemory device. This
must be done prior to activating dm-multipath and/or loading the driver. If Fusion ioMemory devices are not
blacklisted, they will appear busy and you will not be able to attach, detach, or update the firmware on the devices.
To blacklist Fusion ioMemory devices, edit the /etc/multipath.conf file and include the following:
blacklist {
devnode "^fio[a-z]"
}
Handling Driver Unloads

Special consideration must be taken during Fusion ioMemory VSL driver unload time. By default, the init script
searches for any processes holding open a mounted filesystem and kills them, thus allowing the filesystem to be
unmounted. This behavior is controlled by the option KILL_PROCS_ON_UMOUNT in the options file located at
/etc/sysconfig/iomemory-vsl. If these processes are not killed, the filesystem cannot be unmounted.

This may keep the Fusion ioMemory VSL software from unloading cleanly, causing a significant delay on the
subsequent boot.
Upgrading the Firmware

With the Fusion ioMemory VSL software loaded, you should check whether the Fusion ioMemory device's
firmware is up-to-date and then update the firmware if needed. You can do this with either the command-line
utilities or the optional Dell ioSphere software (GUI).
NOTE-
Make sure you have downloaded the firmware archive file that goes with this version of the Fusion
ioMemory VSL software.
ATTENTION!
There is a specific upgrade path that you must take when upgrading a Fusion ioMemory device. Consult
the Dell Fusion ioMemory VSL Release Notes for this Fusion ioMemory VSL software release before
upgrading Fusion ioMemory devices.
Also refer to the Dell Fusion ioMemory VSL Release Notes for this Fusion ioMemory VSL software release
for any Known Issues affecting firmware upgrades.
WARNING!
Do not attempt to downgrade the firmware on any Fusion ioMemory device; doing so may void your
warranty.
When installing a new Fusion ioMemory device along with existing devices, you must upgrade all of the currently
installed devices to the latest available versions of the firmware and Fusion ioMemory VSL software before
installing the new devices. Consult the Dell Fusion ioMemory VSL Release Notes for this Fusion ioMemory VSL
software release for any upgrade considerations.
ATTENTION!
Upgrading Guest OS
If you are using your Fusion ioMemory device within a guest OS (for example, using VMDirectPathIO), you
must power cycle the host server after you upgrade the device(s). Just restarting the virtual machine will
not apply the firmware update.
Command-line Interface
More information on these command-line utilities is available in Command-line Utilities Reference on page 56. All
command-line utilities require root permission.

1. Run the fio-status utility and examine the output. See fio-status on page 63 for usage information.
l If any device is in minimal mode and the reason is listed as outdated firmware, then update the
firmware.
l If a device is not in minimal mode, but the firmware listed for that device is a lower number than
the latest firmware version available with this version of the Fusion ioMemory VSL software, then
the firmware is old, but not outdated.
2. If the firmware is old or outdated, update it using the fio-update-iodrive utility. See fio-update-
iodrive on page 68 for complete information and warnings.
Optional GUI - Dell ioSphere software

You can use the Dell ioSphere software to check the status of your Fusion ioMemory devices. If the Dell ioSphere
software indicates that the device's firmware is outdated, you can also use the Dell ioSphere software to upgrade
the device firmware. Consult the Dell ioSphere software documentation for more information on installing and
using the software.

Configuration
Once you have your Fusion ioMemory device and Fusion ioMemory VSL software installed and loaded, and the
firmware on the device is current, you may need to configure the device and/or software. This section outlines
some of the common configurations that you may need to consider.
Setting the ioMemory VSL Options

This section explains how to set Fusion ioMemory VSL software options. For more information about setting
specific options, see Using Module Parameters on page 73.
One-Time Configuration
Fusion ioMemory VSL software options can be set at install time, on the command line of either insmod or
modprobe. For example, to set the auto_attach option to 0 (which disables auto_attach), run the command:
# modprobe iomemory-vsl auto_attach=0
This option takes effect only for this load of the Fusion ioMemory VSL software; subsequent calls to modprobe or
insmod will not have this option set.
Persistent Configuration
To maintain a persistent setting for an option, add the desired option to /etc/modprobe.d/iomemory-
vsl.conf or a similar file. To prevent the Fusion ioMemory devices from auto-attaching, add the following line to
the iomemory-vsl.conf file:
options iomemory-vsl auto_attach=0
This option then takes effect for every subsequent Fusion ioMemory VSL software load, as well as on autoload of
the Fusion ioMemory VSL software during boot time.
ATTENTION!
If the VSL driver is in the initrd, the initrd will need to be rebuilt in order for the change to take
effect.
Enabling PCIe Power Override

If you have installed any products with multiple Fusion ioMemory devices, such as the Fusion ioMemory ioDrive
Duo device, the device may require additional power to properly function (beyond the minimum 25W provided by
PCIe Gen2 slots). Even if additional power is not required for your device, all dual Fusion ioMemory devices that
receive additional power may benefit with improved performance.
Fusion ioMemory ioDrive2 Duo devices must have additional power in order to properly function. For more
information on which devices require additional power, see the section on Power Cables for Multi-device Products
in the Dell Fusion ioMemory Hardware Installation Guide.

This additional power may be provided in two ways:
l External Power Cable: See the Dell Fusion ioMemory Hardware Installation Guide for information on
installing the external cable.
NOTE-
When a power cable is used, all of the power is drawn from the cable and no power is drawn
from the PCIe slot.
l Enabling a Specified Maximum Power Draw: Some PCIe slots provide additional power (often up to 75W of
power). You may allow the device to draw a specified maximum power from the PCIe slot by setting a
Fusion ioMemory VSL software module parameter. For more information on enabling this override
parameter, see the instructions in the next section.
NOTE-
This parameter overrides the setting that prevents device(s) from drawing more than 25W from
the PCIe slot. The parameter is enabled per device (using device serial numbers with the specified
maximum power value in Watts).
Before you enable this override parameter, ensure that each PCIe slot you will use is rated to provide
enough power for all slots, devices, and server accessories. Refer to the 12G PowerEdge Server Slot
Support Matrix PDF located at http://dell.fusionio.com for information on supported 75W PCIe slots.
WARNING!
If the slot is not capable of providing the needed amount of power, enabling a specified maximum
power draw from the PCIe slot may result in malfunction or even damage server hardware. You
are responsible for any damage to equipment due to improper use of this override parameter
and Dell | SanDisk expressly disclaims any liability for any damage arising from such improper use.
Contact Customer Support if there are any questions or concerns about the override parameter
use.
Refer to the matrix to determine if your slot provides 75W. The fio-status and fio-pci-check utilities may
incorrectly display a 75W slot as a 25W slot.
The following are important considerations:
l If you are installing more than one dual Fusion ioMemory device and enabling the override parameter for
each device, make sure the motherboard is rated to provide the specified maximum power to each slot
that is used.
ATTENTION!
For example, some motherboards safely provide up to 75W to any one slot, but run into power

constraints when multiple slots are used to provide that much power. Installing multiple devices in
this situation may also result in server hardware damage. Consult with the manufacturer to
determine the total PCIe slot power available.
l The override parameter, if enabled correctly, will persist in the system, and will enable the specified
maximum power draw on an enabled device even if the device is removed and then placed in a different
slot within the same system. If the device is placed in a slot that is not rated to provide the specified
maximum power, you may damage your server hardware.
l This override parameter is a setting for the Fusion ioMemory VSL software per server, and is not stored in
the device. When moved to a new server, the device will default to the power limit set by the BIOS for that
PCIe slot unless this override parameter is enabled for that device in the new server. Consult with the
manufacturer to determine the total PCIe slot power available for the new server.
Enabling the Override Parameter
Determine Serial Number(s)
Before you enable this parameter, determine the adapter serial number for each device you will put in a
compatible slot. Use the fio-status command-line utility to determine the adapter serial number(s).
NOTE-
Serial Number Label
You may also inspect the adapter serial number label(s) on the device(s) to determine the serial number
(s). However, as a best practice, confirm that each serial number is an adapter serial number by running
fio-status. The adapter serial number label resides on the back of all Fusion ioMemory ioDrive Duo
devices and Fusion ioMemory ioDrive2 Duo devices. On Fusion ioMemory ioDrive Duo devices, it is on
the PCB component that is attached to the PCIe connector.
l Using fio-status: You determine the adapter serial number differently depending on whether the
device is an Fusion ioMemory ioDrive Duo device or an Fusion ioMemory ioDrive2 Duo device
Fusion ioMemory ioDrive Duo devices: Run the fio-status command-line utility. Sample output:
fio-status
...
Adapter: Dual Adapter
Fusion-io ioDrive Duo 1.28TB, Product Number:24X4P, SN:94126
ioDrive Duo HL, PN:00190000107
...
With Fusion ioMemory ioDrive Duo devices, the adapter serial number is simply the serial number that
appears after the adapter Product Number. In this example, that adapter serial number is 94126.
Fusion ioMemory ioDrive2 Duo devices: Run the fio-status command-line utility. Sample output:

fio-status
...
Adapter: Dual Controller Adapter
Dell ioDrive2 Duo 2410GB MLC, Product Number:7F6JV,
SN:US07F6JV7605121J0010
...
fct0 Attached as 'fioa' (block device)
SN:1149D2717-1121
...
fct1 Attached as 'fiob' (block device)
SN:1149D2717-1111
...
The adapter serial number is the number that is common to each of the fctxFusion ioMemory devices. In
this example, 1149D2717 is the adapter serial number.
ATTENTION!
With Fusion ioMemory ioDrive2 Duo devices, the adapter serial number is NOT the number that
appears after the Product Number.
l Using fio-beacon: If you have multiple devices installed, you may use the fio-beacon utility to verify
where each device is physically located. Consult the utility documentation fio-beacon on page 58 for more
information.
Setting the Parameter in VSL 3.2.15
Set the module parameter by editing the /etc/modprobe.d/iomemory-vsl.conf file and changing the
value for the external_power_override parameter. Please refer to the examples below.
Example 1:
You can use the override parameter to specify the maximum amount of power that specified devices should pull
(in Watts).
options iomemory-vsl external_power_override=<SN-value>:<W-value>
Where the <SN-value>:<W-value> for this parameter is a comma-separated list of value pairs with adapter
serial numbers and the maximum amount of power each device should pull (in Watts).
For example, 1149D0969:40,1159E0972:40,1331G0009:40 would allow three devices (with serial

numbers 1149D0969, 1159E0972, and 1331G0009) to each draw a maximum of approximately 40W.
Any cards in the server not set using this module parameter will use the power limit set by the BIOS for that PCIe
slot.
Example 2:

You can use the override parameter to specify the maximum amount of power all devices in the system should pull
(in Watts). Use the new serial number wildcard (*) to specify all devices in the system.
options iomemory-vsl external_power_override=*:<W-value>
Where the *:<W-value> for this parameter is the serial number wildcard (*) that includes all cards in the
system, and the maximum amount of power all devices should pull (in Watts).
For example, *:50 would allow all cards in the system to draw a maximum of approximately 50W.
ATTENTION!
Do not use serial numbers with the (*) wildcard (do not mix commands in Examples 1 and 2).
ATTENTION!
You must reboot or unload/load the driver to enforce any parameter changes.
Virtual Controller Configuration

Depending on your use case and application, you may benefit from configuring supported devices to use Virtual
Controller technology.
When configured, each physical Fusion ioMemory device is split into two (virtual) logical devices. Splitting the
Fusion ioMemory device into two virtual devices has the following implications:
l Latency: There is no affect on latency.

l Throughput: The total peak I/O bandwidth of the device is approximately the same.
l IOPS: Depending on the use of the virtual devices (especially the average I/O size), the peak IOPS for each
virtual device is about the same for a non-split device. In other words, the combined peak IOPS of the two
virtual devices can be nearly double that of a non-split device. For details, see in the Maintenance section.
l Capacity: Due to virtualization overhead, the combined capacity of the two virtual devices is slightly less
than that of a single-controller device. See the Dell Fusion ioMemory VSL Release Notes for a list of
compatible devices and their Virtual Controller capacities.
Converting your Fusion ioMemory device to a Virtual Controller configuration will split the Fusion ioMemory device
into two logical devices.
For 512B I/Os, the combined IOPS performance of the two virtual devices is approximately double that of a single-
controller device. For 4KiB I/Os, there is more than an 80% improvement in IOPS performance with virtual devices.
For 16kB and larger I/Os, there is no improvement of total IOPS performance over a non-Virtual Controller
configuration.
Latency in the virtual devices is unaffected, and the combined bandwidth of the two virtual devices is the same as it
would be without the split. Due to the overhead of an additional device, the combined capacity of the two virtual
devices is slightly less than that of a single-controller device.

Splitting a single physical device into multiple virtualized devices, or merging multiple virtualized devices back to a
single physical device, requires a low-level format, which will erase all of the data on the device. Be sure to back up
all of your data.
Supported Devices
Only relatively new devices (with few writes performed) may be split or merged. Devices with too much wear are
unsuitable for converting to or from a Virtual Controller configuration. Merging virtual devices may also result in
additional wear (depending on the wear differences of the two virtual devices).
To be suitable for splitting or merging, devices (including Virtual Controller devices) must have 90% or more of
their remaining rated endurance of Petabytes Written (PBW). This rating as well as the current percentage
remaining is visible in fio-status with the -a option. For example:
fio-status /dev/fct1 -a
...
Rated PBW: 17.00 PB, 99.95% remaining
In the above example, the device is suitable for conversion because it has more than 90% of the rated PBW
remaining.
If you attempt to merge or split a device that does not support Virtual Controller technology or a device that has
too much wear, the update utility will not allow the conversion and the firmware upgrade will not take place. See
the Release Notes for a list of devices that support Virtual Controller technology and their capacities after the
conversion.
Multi-device Products
For products with more than one Fusion ioMemory device, such as a Fusion ioMemory ioDrive2 Duo device, you
must configure all of the Fusion ioMemory devices to Virtual Controller technology at the same time. All of the
devices must also be merged at the same time. For example, the two Fusion ioMemory devices in a Fusion
ioMemory ioDrive2 Duo device will be converted into four virtual devices. The utility will not allow a conversion if
you attempt to split or merge only one physical device in a multi-device product.
Splitting Controllers
Be sure to use firmware that supports Virtual Controller technology. Consult the Release Notes to determine if the
firmware for that release supports Virtual Controller technology.
1. Back up all of your data. Because a low-level format is needed to complete the conversion, all of the user
data on your device will be erased.
2. Use the fio-update-iodrive command-line utility to configure a Fusion ioMemory device to use
Virtual Controller technology:
l Use the --split option to split the controller.
l Use the -d option to specify a device, otherwise all installed devices that can be split will be split.
l Specify the firmware path, and check the Dell Fusion ioMemory VSL Release Notes to make sure the

firmware supports Virtual Controller technology.
Example:
fio-update-iodrive --split -d /dev/fct0 <firmware-path>
After rebooting, each physical device will be split into two virtual devices. Each Fusion ioMemory
device will therefore split into two logical devices, each with a unique device path. For example,
/dev/fct0 may become /dev/fct0 and /dev/fct1. You will manage each device as a
unique device.
3. Reboot.
4. Load the Fusion ioMemory VSL driver.
5. Run fio-status to determine which devices need to be formatted.
6. Low-level format the device(s). For example:
fio-format /dev/fct0 /dev/fct1
Formatting will erase all user data, be sure to back up your data. You can reverse the split by merging the
controllers (without losing data) up until you format the virtual devices.
Merging Controllers
If your Fusion ioMemory device (including the two virtual devices) is suitable for merging, then you will be able to
use the fio-update-iodrive utility to merge the virtual devices back into one physical device.
1. Back up all of your data. Because a low-level format is needed to complete the merge, all of the user data
on your device will be erased.
2. Use the fio-update-iodrive command-line utility to configure the device for merging:
a. Use the --merge option to merge the virtual devices.
b. Use the -d option to specify a device.
ATTENTION!
The fio-update-iodrive utility only successfully works against one of the two virtual
devices for each physical Fusion ioMemory device. Out of the two virtual devices, only the
first virtual device (in terms of device numbering) is linked to the physical device (and the
firmware). The second virtual device is not linked, and any firmware operation against that
second virtual device will fail with this message:
Error: Device '/dev/fctx' had an error while updating. This
device does not support firmware update.
This is expected, and the error will not affect the update/merge of the first (linked) virtual

device. The update operation will complete on all devices that can merge and otherwise
accept firmware changes.
c. Specify the firmware path, and check the Dell Fusion ioMemory VSL Release Notes to make sure the
firmware supports Virtual Controller technology.
Example:
fio-update-iodrive --merge -d /dev/fct0 <firmware-path>
3. Reboot.
4. Load the Fusion ioMemory VSL driver.
5. Run fio-status to determine which devices need to be formatted.
6. Low-level format the device(s). For example:
fio-format /dev/fct0
WARNING!
Formatting will erase all user data, be sure to back up your data. You can reverse the merge by
splitting the controllers (without losing data) up until you format the merged device.
Using the Device as Swap

To safely use the Fusion ioMemory device as swap space requires passing the preallocate_memory kernel
module parameter. To set this parameter, use either the optional Dell ioSphere software (see Dell ioSphere
software documentation), or add the following line to the /etc/modprobe.d/iomemory-vsl.conf file (see
Using Module Parameters on page 73 for more information on using parameters):
options iomemory-vsl preallocate_memory=1149D2717-1121,1149D2717-1111,10345
l Where 1149D2717-1111,1149D2717-1111,10345 are serial numbers obtained from fio-status,

see fio-format on page 61.
NOTE-
Be sure to provide the serial numbers for the Fusion ioMemory device, not an adapter, when applicable.
ATTENTION!
You must have enough RAM available to enable the Fusion ioMemory device with pre-allocation enabled
for use as swap. Attaching a Fusion ioMemory device, with pre-allocation enabled, without sufficient RAM
may result in the loss of user processes and system instability.

Consult the Dell Fusion ioMemory VSL Release Notes for RAM requirements with this version of the Fusion
ioMemory VSL software.
NOTE-
The preallocate_memory parameter is recognized by the Fusion ioMemory VSL software at load
time, but the requested memory is not actually allocated until the specified device is attached.
Setting the Amount of Preallocated Memory

If you enable devices for preallocation (as described above), the Fusion ioMemory VSL software will automatically
preallocate the amount of memory based on your formatted sector size. If you have a sector size that is less than
4KiB (for example, 512B sectors), then the Fusion ioMemory VSL software will preallocate a very large amount of
memory (calculated for worst-case scenarios based on 512B sectors).
See the Dell Fusion ioMemory VSL Release Notes for worst-case RAM requirements based on formatted sizes. The
Fusion ioMemory VSL software preallocates enough memory for worst-case scenarios in order to avoid an out-of-
memory situation where the Fusion ioMemory device would no longer function due to insufficient memory
(resulting in a system crash).
Operating systems normally use at least 4KiB blocks of data for virtual memory, so it would be safe in most cases
for the Fusion ioMemory VSL software to preallocate enough memory for a worst-case scenario based on 4KiB
sectors. There are two ways to force the Fusion ioMemory VSL software to preallocate based on 4KiB block sizes:
l Format the device to 4KiB sectors. With preallocation enabled, the Fusion ioMemory VSL software will
automatically preallocate memory based on 4KiB blocks of data. Use fio-format to format the Fusion
ioMemory device with 4KiB sector sizes.
l Use the preallocate_mb parameter. This parameter sets the amount of memory that the Fusion
ioMemory VSL software will preallocate for every Fusion ioMemory device in the system that is enabled for
preallocation (as described in the previous section).
WARNING!
Ensure that you are preallocating enough system memory. If you do not allocate enough memory
and you attempt to use the device for swap space (due to a miscalculation or entry error), the
system may crash as memory resources are depleted.
1. Determine the amount of system memory needed for every device in a worst-case scenario.
a. See the Dell Fusion ioMemory VSL Release Notes for worst-case RAM requirements based on
sector sizes.
b. You can base this calculation on the sector size that aligns with the data block size that your
operating system uses for virtual memory.
2. Set the parameter by adding the following line to the /etc/modprobe.d/iomemory-

vsl.conf file (see Using Module Parameters on page 73 for more information on using

parameters):
options iomemory-vsl preallocate_mb=<value>
Where <value> is the amount of system memory in MB that the Fusion ioMemory VSL software
should preallocate for every Fusion ioMemory device that is enabled for preallocation. For example,
if you entered a value of 3500, then the Fusion ioMemory VSL software will preallocate about
3.5GB of RAM for every Fusion ioMemory device that is enabled for preallocation in that system.
ATTENTION!
In order for the preallocation of memory to be effective, this value should be larger than
the default memory usage of the device (as reported by runing fio-status -a).
Using the Logical Volume Manager

The Logical Volume Manager (LVM) volume group management application handles mass storage devices like
Fusion ioMemory devices if you add the Fusion ioMemory device as a supported type:
1. Locate and edit the /etc/lvm/lvm.conf configuration file.
2. Add an entry similar to the following to that file:
types = [ "fio", 16 ]
The parameter "16" represents the maximum number of partitions supported by the device.
NOTE-
If using LVM or MD, do not use udev to load the Fusion ioMemory VSL driver. The init script will
ensure that the LVM volumes and MD devices are detached before attempting to detach the
Fusion ioMemory device.
Mounting Filesystems when Using the init Script

If you are using the init script to load the ioMemory VSL software, once you have configured your devices you will
need to follow the instructions in the Mounting Filesystems when Using the init Script subsection of the section
Controlling Driver Loading on page 20.
Configuring RAID using the Logical Volume Manager

The simplest way to using your Fusion ioMemory devices with LVM is to use the entire block device with it (for
example: /dev/fioa rather than /dev/fioa1, /dev/fioa2, etc.). This way the block device does not need
to partitioned ahead of time, though partitioning is also supported. The examples that follow assumes the entire
block device is used.
Whether you plan to stripe (RAID 0) or mirror (RAID 1) the devices, the first two steps are the same:

1. First, create physical volumes, for example:
$ pvcreate /dev/fioa /dev/fiob
2. Next add these physical volumes to a volume group, for example:
$ vgcreate iomemory_vg /dev/fioa /dev/fiob
Creating a Striped Volume (RAID 0) Using LVM
With the volume group created, you can create logical volumes within this volume group. In this instance, only one
will be created and we name it iomemory_lv:
1. Create the striped volume using the -i2 option for two stripes, for example:
$ lvcreate -l 100%VG -n iomemory_lv -i2 iomemory_vg
2. Create a file system on the newly created volume, for example:
$ mkfs.ext3 /dev/iomemory_vg/iomemory_lv
3. In order to make this persistently mount at boot time, edit the /etc/sysconfig/iomemory-vsl file
by adding the volume group path under the example line:
Example: LVM_VGS="/dev/vg0 /dev/vg1"

LVM_VGS="/dev/iomemory_vg"
Be sure to just add the volume group path and not the logical volume as well. For more information on
using the init script and the /etc/sysconfig/iomemory-vsl file, see Loading the ioMemory VSL
Driver on page 19.
Creating a Mirrored Volume (RAID 1) Using LVM
With the volume group created, you can create logical volumes within this volume group. In this instance, only one
will be created and we name it iomemory_lv:
1. Create the mirrored volume using the -m1 option for an original linear volume plus one copy, for example:
$ lvcreate -l 50%VG -n iomemory_lv -m1 --corelog iomemory_vg
You can monitor the progress of the initial mirror synchronization using the lvs command. For example:
lvs -a -o +devices
ATTENTION!
With the \--corelog option, LVM uses an in-memory region or log to track the state of the

mirror legs. Since it is in-memory and lost at reboot, it must be regenerated at boot by scanning
the mirror legs. For alternative configurations, consult the LVM documentation.
2. Create a file system on the newly created volume, for example:
$ mkfs.ext3 /dev/iomemory_vg/iomemory_lv
3. In order to make this persistently mount at boot time, edit the /etc/sysconfig/iomemory-vsl file
by adding the volume group path under the example line:
# Example: LVM_VGS="/dev/vg0 /dev/vg1"

LVM_VGS="/dev/iomemory_vg"
Be sure to just add the volume group path and not the logical volume as well. For more information on
using the init script and the /etc/sysconfig/iomemory-vsl file, see Loading the ioMemory VSL
Driver on page 19.
Configuring RAID Using mdadm

You can configure two or more Fusion ioMemory devices into a RAID array using software-based RAID solutions.
NOTE-
If you are using RAID1/Mirrored and one device fails, be sure to run fio-format on the replacement
device (not the existing, good device) before rebuilding the RAID. Following are some examples of some
common RAID configurations using the mdadm utility.
ATTENTION!
The Linux kernel RAID 5 implementation performs poorly at high data rates. This is an issue in the Linux
kernel. Alternatives include using RAID 10, or possibly a third-party RAID stack.
Mounting Arrays
Once you are done making your array (by following one of the configuration samples below), you must edit the
/etc/sysconfig/iomemory-vsl file and add the array path under the example line. This will make the
array mount at boot time.
In all of the examples below, we create arrays named md0. Here is is how you would edit the
/etc/sysconfig/iomemory-vsl file to include the array using md0 as an example (you add the array just
below the example line in that file):
# Example: MD_ARRAYS="/dev/md0 /dev/md1"

MD_ARRAYS="/dev/md0"

If you are using the init script to load the ioMemory VSL software, once you have configured your devices you will
need to follow the instructions in the Mounting Filesystems when Using the init Script subsection of the section
Controlling Driver Loading on page 20.
RAID 0
To create a striped set, where fioa and fiob are the two Fusion ioMemory devices you want to stripe, run this
command:
$ mdadm --create /dev/md0 --chunk=256 --level=0 --raid-devices=2 /dev/fioa

/dev/fiob
Making the Array Persistent (Existing after Restart)
On some versions of Linux, the configuration file is in /etc/mdadm/mdadm.conf, not /etc/mdadm.conf.
Inspect /etc/mdadm.conf. If there are one or more lines declaring the devices to inspect, make sure one of
those lines specifies "partitions" as an option. If it does not, add a new DEVICE line to the file specifying "partitions"
like this:
DEVICE partitions
Also add a device specifier for the fio Fusion ioMemory devices:
DEVICE /dev/fio*
To see if any updates are needed to /etc/mdadm.conf, issue the following command:
$ mdadm --examine --scan
Compare the output of this command to what currently exists in mdadm.conf and add any needed sections to
/etc/mdadm.conf.
NOTE-
For example, if the array consists of two devices, there will be three lines in the output of the command
that are not present in the mdadm.conf file: one line for the array, and two device lines (one line for
each device). Be sure to add those lines to the mdadm.conf so it matches the output of the command.
For further details please see the mdadm and mdadm.conf man pages for your distribution.
With these changes, on most systems the RAID 0 array will be created automatically upon restart. However, if you
have problems accessing /dev/md0 after restart, run the following command:
$ mdadm --assemble --scan

You may also want to disable udev loading of the Fusion ioMemory VSL driver, if needed, and use the init script
provided for driver loading. Please see the Using the init Script subsection of the section Controlling Driver Loading
on page 20 for further details on how to use the init script.
NOTE-
In SLES 11, you may need to run the following commands to make sure these services are run on boot:
chkconfig boot.md on
chkconfig mdadmd on
RAID 1
To create a mirrored set using the two Fusion ioMemory devices fioa and fiob, run this command:
$ mdadm --create /dev/md0 --level=1 --raid-devices=2 /dev/fioa /dev/fiob
RAID 10
To create a striped, mirrored array using four Fusion ioMemory devices (fioa, fiob, fioc, and fiod), run this
command:
$ mdadm --create /dev/md0 -v --chunk=256 --level=raid10 --raid-devices=4

/dev/fioa /dev/fiob /dev/fioc /dev/fiod
Building a RAID10 Across Multiple Devices

In a RAID10 configuration, sets of two disks are mirrored, and then those mirrors are striped. When setting up a
RAID10 across multiple Fusion ioMemory devices, it is best to make sure that no mirror resides solely on the two
Fusion ioMemory devices that comprise an a single product (such as a Fusion ioMemory ioDrive Duo device).
In order to get the data to lay out properly,
l Use the --layout=n2 option when creating the RAID10 (though it should be the default)
l Ensure that no two Fusion ioMemory devices from the same device are listed side by side.
The following sample code shows some recommended configurations.
NOTE-
The following commands assume that all Fusion ioMemory devices have been freshly formatted with the
fio-format utility.
ATTENTION!
The ordering of the fiox devices is critical.
# 2 Dual Device Products RAID10

$ mdadm --create --assume-clean --level=raid10 --layout=n2 -n 4 /dev/md0 \

/dev/fioa /dev/fioc \
/dev/fiob /dev/fiod
# Mirror groups are: fioa,fioc and fiob,fiod

/dev/fioa /dev/fiod \
/dev/fioc /dev/fiof \
/dev/fioe /dev/fiob

/dev/fioe /dev/fioh \
/dev/fiog /dev/fiob

/dev/fioe /dev/fioh \
/dev/fiog /dev/fioj \
/dev/fioi /dev/fiol \
/dev/fiok /dev/fion \
/dev/fiom /dev/fiop \
/dev/fioo /dev/fiob
Discard (TRIM) Support

With this version of the Fusion ioMemory VSL software, Discard (also known as TRIM) is enabled by default.
Discard addresses an issue unique to solid-state storage. When a user deletes a file, the device does not recognize
that it can reclaim the space. Instead the device assumes the data is valid.
Discard is a feature on newer filesystem releases. It informs the device of logical sectors that no longer contain
valid user data. This allows the wear-leveling software to reclaim that space (as reserve) to handle future write
operations.
Discard (TRIM) on Linux

Discard is enabled by default in the Fusion ioMemory VSL software release. However, for discard to be
implemented, the Linux distribution must support this feature, and discard must be turned on.
In other words, if your Linux distribution supports discard, and discard is enabled on the system, then discard will
be implemented on your Fusion ioMemory device.
Under Linux, discards are not limited to being created by the filesystem; discard requests can also be generated
directly from userspace applications using the kernel's discard ioctl.

ATTENTION!
There is a known issue that ext4 in Kernel.org 2.6.33 or earlier may silently corrupt data when discard is
enabled. This has been fixed in many kernels provided by distribution vendors. Please check with your
kernel provider to be sure your kernel properly supports discard. For more information, see the Errata in
the Dell Fusion ioMemory VSL Release Notes for this version of the software
NOTE-
On Linux, MD and LVM support discards on newer kernels. LVM supports this on kernels 2.6.29 and
newer. MD supports discard on kernels 3.7 and newer.

Performance and Tuning
Fusion ioMemory devices provide high bandwidth, high Input/Output per Second (IOPS), and are specifically
designed to achieve low latency.
As Fusion ioMemory devices improve IOPS and low latency, the device performance may be limited by operating
system settings and BIOS configuration. These settings may need to be tuned to take advantage of the revolutionary
performance of Fusion ioMemory devices.
While Fusion ioMemory devices generally perform well out of the box, this section describes some of the common
areas where tuning may help achieve optimal performance.
Disable CPU Frequency Scaling

Dynamic Voltage and Frequency Scaling (DVFS) are power management techniques that adjust the CPU voltage
and/or frequency to reduce power consumption by the CPU. These techniques help conserve power and reduce
the heat generated by the CPU, but they adversely affect performance while the CPU transitions between low-
power and high-performance states.
These power-savings techniques are known to have a negative impact on I/O latency and IOPS. When tuning for
performance, you may benefit from reducing or disabling DVFS completely, even though this may increase power
consumption.
DVFS, if available, is often configurable as part of your operating systems power management features as well as
within your system's BIOS interface. Within the operating system and BIOS, DVFS features are often found under
the Advanced Configuration and Power Interface (ACPI) sections; consult your computer documentation for
details.
Limiting ACPI C-States

Newer processors have the ability to go into lower power modes when they are not fully utilized. These idle states
are known as ACPI C-states. The C0 state is the normal, full power, operating state. Higher C-states (C1, C2, C3,
etc.) are lower power states.
While ACPI C-states save on power, they can have a negative impact on I/O latency and maximum IOPS. With each
higher C-state, typically more processor functions are limited to save power, and it takes time to restore the
processor to the C0 state.
When tuning for maximum performance you may benefit from limiting the C-states or turning them off
completely, even though this may increase power consumption.
Setting ACPI C-State Options
If your processor has ACPI C-states available, you can typically limit or disable them in the BIOS interface
(sometimes referred to as a Setup Utility). APCI C-states may be part of of the Advanced Configuration and Power
Interface (ACPI) menu. Consult your computer documentation for details.

C-States Under Linux
Newer Linux kernels have drivers that may attempt to enable APCI C-states even if they are disabled in the BIOS.
You can limit the C-state in Linux (with or without the BIOS setting) by adding the following to the kernel boot
options:
intel_idle.max_cstate=0 processor.max_cstate=0
In this example, the maximum C-state allowed will be C0 (disabled).
Setting NUMA Affinity

Servers with a NUMA (Non-Uniform Memory Access) architecture may require special installation instructions in
order to maximize Fusion ioMemory device performance. This includes most multi-socket servers.
On some servers with NUMA architecture, during system boot, the BIOS will not associate PCIe slots with the
correct NUMA node. Incorrect mappings result in inefficient I/O handling that can significantly degrade
performance. To prevent this, you must manually assign Fusion ioMemory devices optimally among the available
NUMA nodes.
See NUMA Configuration on page 75 for more information on setting this affinity.
Setting the Interrupt Handler Affinity

Device latency can be affected by placement of interrupts on NUMA systems. We recommend placing interrupts
for a given device on the same NUMA node that the application is issuing I/O from. If the CPUs on this node are
overwhelmed with user application tasks, in some cases it may benefit performance to move the the interrupts to
a remote node to help load-balance the system.
Many operating systems will attempt to dynamically place interrupts across the nodes, and generally make good
decisions.
Linux IRQ Balancing

In Linux this dynamic placement is called IRQ Balancing. You can check to see if the IRQ balancer is effective by
checking /proc/interrupts. If the interrupts are unbalanced (too many device interrupts on one node) or on
an overwhelmed node, you may need to stop the IRQ balancer and manually distribute the interrupts in order to
balance the load and improve performance.
NOTE-
Restarting the IRQ Balancer after the Fusion ioMemory VSL software loads (and the Fusion ioMemory
devices are attached) may resolve interrupt affinity issues. For example, run one of the following
commands (depending on your distribution):
/etc/init.d/irqbalance restart
/etc/init.d/irq_balancer restart
If that does not resolve the affinity issues, then we recommend manual pinning the device interrupts to
specific nodes.

Hand-tuning interrupt placement in Linux is an advanced option that requires profiling of application performance
on any given hardware. Please see your operating system documentation for information on how to pin specific
device interrupts to specific nodes.

Monitoring and Managing Devices
Dell | SanDisk provides many tools for managing your Fusion ioMemory devices. These tools will allow you to
monitor the devices for errors, warnings, and potential problems. They will also allow you to manage the devices
including performing the following functions:
l Firmware upgrades
l Low-level formatting
l Attach and detach actions
l Device status and performance information
l Configuring Swap and Paging
l Generating bug reports
Management Tools
Dell | SanDisk has provided several tools for monitoring and managing Fusion ioMemory devices. These include
stand-alone tools that require no additional software and data-source tools that can be integrated with other
applications.
Consider the descriptions of each tool to decide which tool (or combination of tools) best fits your needs.
ATTENTION!
The Fusion ioMemory VSL software does print some error messages to the system logs, and while these
messages are very useful for troubleshooting purposes, the Fusion ioMemory VSL software log messages
are not designed for continual monitoring purposes (as each is based on a variety of factors that could
produce different log messages depending on environment and use case). For best results, use the tools
described in this section to regularly monitor your devices.
Stand-alone Tools
These stand-alone tools do not require any additional software.
l Command-line Utilities: These utilities are installed with the ioMemory VSL software and are run manually
in a terminal. The fio-status utility provides status for all devices within a host. The other utilities allow
you to perform other management functions. See Command-line Utilities Reference on page 56 for full
details.
l Dell ioSphere software: The GUI browser-based Dell ioSphere software allows you to monitor and manage
every Fusion ioMemory device installed in multiple hosts across your network. It collects all of the alerts for
all Fusion ioMemory devices and displays them in the Alert Tab. You may also set up the Dell ioSphere
software to send email or SMS messages for specific types of alerts or all alerts. The Dell ioSphere software
packages and documentation are available as separate downloads.
Data-source Tools
These data-source tools provide comprehensive data, just like the stand-alone tools, but they do require integration
with additional software. At a minimum, some tools can interface with a browser. However, the benefit of these
tools is that they can be integrated into existing management software that is customized for your
organization.

These tool packages and documentation are also available as separate downloads (separate from the Fusion
ioMemory VSL software packages).
l SNMP Subagent: The Dell | SanDisk SNMP AgentX subagent allows you to monitor and manage your Fusion
ioMemory devices using the Simple Network Management Protocol. You can use a normal SNMP browser,
or customize your existing application to interface with the subagent.
l SMI-S CIM Provider: The CIM provider allows you to monitor and manage your devices using the Common
Information Model. You can use a normal CIM browser, or customize your existing application to interface
with the CIM provider.
Example Conditions to Monitor

This section gives examples of conditions you can monitor. It is intended as an introduction and not as a
comprehensive reference. These conditions will have slightly different names, states, and values, depending on the
tool you choose. For example, an SNMP MIB may have a different name than a SMI-S object or an API function.
In order to properly monitor these conditions, you should become familiar with the tool you choose to implement
and read the documentation for that tool. You may also discover additional conditions that you wish to frequently
monitor.
For quick reference, the possible states/values of these conditions are described as Normal (GREEN), Caution/Alert
(YELLOW), or Error/Warning (RED). You may implement your own ranges of acceptable states/values, especially if
you use a data-source tool.
Device Status
All of the monitoring tools return information on the status of the Fusion ioMemory devices, including the following
states:
GREEN Attached
YELLOW Detached, Busy (including: Detaching, Attaching, Scanning, Formatting, and Updating)
RED Minimal Mode, Powerloss Protect Disabled
If the device is in Minimal Mode, the monitoring tool can display the reason for the Minimal Mode status.
Required Actions
If the device is in Minimal Mode, the action will depend on the reason. For example, if the reason is outdated
firmware, then you will need to update the firmware.
Temperature
Fusion ioMemory devices require adequate cooling. In order to prevent thermal damage, the Fusion ioMemory
VSL software will start throttling write performance once the on-board controller reaches a specified temperature.
If the controller temperature continues to rise, the software will shut down the device once the controller
temperature reaches the maximum operating temperature.

These temperatures depend on the device. Newer Fusion ioMemory devices have higher thermal tolerances.
Consult the Dell Fusion ioMemory Hardware Installation Guide to determine the thermal tolerances of all devices
you will monitor. This table uses the follwing controller thermal tolerances: 93°C throttling, 100°C shutdown.
GREEN <93°C
YELLOW 93-99°C
RED 100°C
You may wish to shift the conditions by a few degrees so the YELLOW condition exists before throttling occurs. For
example:
GREEN <90°C
YELLOW 90-96°C
RED 97°C
ATTENTION!
NAND Board Temperature
Newer Fusion ioMemory devices also report the temperature of the NAND Boards. This is also a critical
temperature to monitor. Consult the Dell Fusion ioMemory Hardware Installation Guide to see if your
device reports this temperature and to see the temperature thresholds.
Required Actions
If the temperature is at or approaching the YELLOW condition, you must increase the cooling for your system. This
may include increasing the fan speed, bringing down the ambient temperature, reducing write load, or moving the
device to a different slot.
Health Reserves Percentage

Fusion ioMemory devices are highly fault-tolerant storage subsystem with many levels of protection against
component failure and the loss nature of solid-state storage. As in all storage subsystems, component failures may
occur.
By pro-actively monitoring device age and health, you can ensure reliable performance over the intended product
life. The following table describes the Health Reserve conditions.
GREEN >10%
YELLOW 4% -10%
RED 0% - 3%

At the 10% healthy threshold, a one-time warning is issued. At 3%, the device is considered unhealthy and enters
write-reduced mode. At the 1% threshold, the device will enter read-only mode.
For complete information on Health Reserve conditions and their impact on performance, see Monitoring the
Health of Devices on page 71.
Required Actions
The device needs close monitoring as it approaches 0% reserves and goes into write-reduced mode, which will
result in reduced write performance. Prepare to replace the device soon.
Write (Health Reserves) Status

In correlation with the Health Reserves Percentage, the management tools will return write states similar to these:
GREEN Device is healthy
YELLOW Device is getting close to entering reduced write mode.
RED Device has entered reduced-write or read-only mode to preserve the flash from further wearout.
Required Actions
The device needs close monitoring as it approaches 0% reserves and goes into write-reduced mode, which will
result in reduced write performance. Prepare to replace the device soon.
Device LED Indicators

If you have physical access to the devices and depending on your device configuration, you can use the LED
indicator(s) on the bracket to monitor their status.
ATTENTION!
Fusion ioMemory devices may have an additional LEDs (that are not on the bracket, as shown below. You
can ignore those other LEDs, as they are not meant for monitoring device and software functionality.
The LEDs on your device should be similar to one of these configurations:

This table explains the information that these LEDs convey:
Green Yellow Amber Indication Notes

OFF OFF OFF Power is off.
Power is on. Problem with device, or Use fio-status to view

OFF OFF LIT driver not loaded (and device problem, or load driver (and attach
unattached). device).
Power is on. Driver loaded (device may

LIT OFF OFF You may need to attach the device.
or may not be attached).
Writing (Rate indicates volume of Can appear in combination with the

LIT FLASHING OFF
writes). Read LED indication.
Can appear in combination with the

FLASHING OFF OFF Read (rate indicated volume of reads).
Write LED indication.
Use the fio-beacon utility to

LIT LIT LIT Location Beacon.
initiate this behavior.

Maintenance
This section explains additional software maintenance functions not covered in the sections Configuration on page
25 and Monitoring and Managing Devices on page 44.
NOTE-
All commands require administrator privileges. Log in as "root" or use sudo to run the commands.
Uninstalling the Software
NOTE-
If you came to this section from the Installation Overview section, return to Installation Overview on page
12 after you uninstall previous versions of the Fusion ioMemory VSL software and utilities.
Uninstalling the Fusion ioMemory VSL Utilities and Other Support Packages
Uninstalling 2.x Support Packages
To uninstall the support RPM packages, run this command (adding or removing package names as needed):
# rpm -e fio-util fio-snmp-agentx fio-common fio-firmware iomanager-gui

iomanager-jre libfio libfio-doc libfusionjni fio-sysvinit fio-smis fio-snmp-
mib libfio-dev
Uninstalling 3.x Support Packages
To uninstall the support RPM packages, run this command (adding or removing package names as needed):
# rpm -e fio-util fio-snmp-agentx fio-common fio-firmware libvsl libvsl-doc

fio-sysvinit fio-smis fio-snmp-mib libvsl-dev
Uninstalling the Fusion ioMemory VSL RPM Package

With versions 2.x and later (including 3.x releases) of the Fusion ioMemory VSL software, you must specify the
kernel version of the package you are uninstalling. Run this command to find the installed driver packages:
# rpm -qa | grep -i iomemory
Sample output:
iomemory-vsl-2.6.18-194.el5-2.2.2.82-1.0
Uninstall the Fusion ioMemory VSL software by running a command similar to this example (specify the kernel
version of the driver you wish to uninstall):
# rpm -e iomemory-vsl-2.6.18-194.el5-2.2.0.82-1.0

1. Block version command:
SCSI version command:
esxcli --server <servername> software vib remove -n scsi-iomemory-vsl
Unloading the Software Driver

To unload the driver, run this command:
# modprobe -r iomemory-vsl
Upgrading the Kernel

If you ever plan to upgrade the kernel when the Fusion ioMemory VSL software is installed, you must:
1. Unload the Fusion ioMemory VSL driver.

2. Uninstall the Fusion ioMemory VSL software.
3. Upgrade the kernel.
4. Install a Fusion ioMemory VSL software package that is compiled for the new kernel.
Failure to follow this procedure may result in driver load issues.
Disabling the ioMemory VSL Software

The Fusion ioMemory VSL software automatically loads by default when the operating system starts. You can
disable the Fusion ioMemory VSL software for diagnostic or troubleshooting purposes.
To disable auto-load, you can use one of the following methods:
l Uninstall the Fusion ioMemory VSL software to keep it from loading.

l Move the Fusion ioMemory VSL software out of the /lib/modules/<kernel_version> directory to
keep it from loading.
l On systems that rely on udev to load drivers, locate and edit the /etc/modprobe.d/iomemory-vsl.conf file
that already has the following line:
# blacklist iomemory-vsl
To disable loading, remove the "#" from the line and save the file. With the blacklist command in place,
restart Linux. The Fusion ioMemory VSL software will not be loaded by udev.
ATTENTION!
If the driver is in the initrd, the initrd will need to be rebuilt in order for the modprobe
configuration changes to take effect.
To re-enable udev-loading of the driver, replace the "#" to commment out the line.

Disabling Auto-Attach
When the Fusion ioMemory VSL software is installed, it is configured to automatically attach any devices when the
Fusion ioMemory VSL software is loaded. Sometimes you may want to disable the auto-attach feature (to assist in
troubleshooting or diagnostics). To do so:
NOTE-
You can also use the Dell ioSphere software to enable or disable auto-attach. See the Dell ioSphere
software documentation for more information.
1. Edit the following file:
/etc/modprobe.d/iomemory-vsl.conf
2. Add the following line to that file:
3. Save the file. To re-enable auto-attach, simply edit the file and either remove that line or change it to the
following:
Unmanaged Shutdown Issues

Unmanaged shutdowns due to power loss or other circumstances can force the Fusion ioMemory device to
perform a consistency check during the restart. This may take several minutes or more to complete.
Although data written to the Fusion ioMemory device is not lost due to unmanaged shutdowns, important data
structures may not have been properly committed to the device. This consistency check (also called a rescan)
repairs these data structures.
Improving Rescan Times

The rescan of the device (also called a consistency check) the VSL performs after an unmanaged shutdown may
take an extended period of time depending on the total capacity of the device(s) that the Fusion ioMemory VSL
software needs to scan.
Default Fast Rescan
By default, all Fusion ioMemory devices formatted with the fio-format utility or ioSphere are formatted to have
improved rescan times. You can disable this default fast rescan by reformatting the device and using the -R option.
Disabling this feature will reclaim some reserve capacity that is normally set aside to help improve rescan times.
If you leave the default fast rescan feature in place you can also take further steps to improve rescan times by
implementing one of the following module parameters.

Faster Rescans Using Module Parameters
These two module parameters require the default fast rescan formatting structure, and they also use system
memory (RAM) to help improve rescan times. The extra memory enables the rescan process to complete faster,
which reduces downtime after a hard shutdown. This memory allocation is only temporary and is freed up after
the rescan process is complete.
If you decide to use one of these parameters, you will need to set the upper limit of RAM used by that parameter.
To do this, you will need to determine how much RAM each parameter may use in your scenario, how much
system RAM is available, and (therefore) which parameter is more suited for your use case.
For more information on setting module parameters, see Using Module Parameters on page 73.
Here is a quick comparison of the two parameters:
l RMAP Parameter
o Fastest: This improvement results in the fastest rescan times.
o Less Scalable: (All or nothing.) This parameter requires enough RAM to function. If the RAM limit is
set too low, then the Fusion ioMemory VSL software will not use RMAP at all, and it will revert back
to the default fast rescan process.
o Target Scenario: This parameter will improve any use case if there is enough RAM available for the
parameter. It is more suited for smaller capacity Fusion ioMemory devices and/or systems with
fewer Fusion ioMemory devices installed. We also recommend it for devices that have been used
for many small random writes.
l RSORT Parameter
o Faster: This improves rescan times over the default fast rescan process.
o Scalable: With this parameter, the Fusion ioMemory VSL software works with the system RAM to
improve rescan times until it reaches the RAM limit set in the parameter. At that point, the software
reverts back to the default fast rescan process.
o Target Scenario: This parameter will improve rescan times in any use scenario. It is especially useful
in systems with multiple Fusion ioMemory devices and/or larger-capacity Fusion ioMemory devices.
We also recommend it when Fusion ioMemory devices are used to store databases.
RMAP Parameter
The rmap_memory_limit_MiB parameter sets the upper memory (RAM) limit (in mebibytes) used by the
Fusion ioMemory VSL software to perform the RMAP rescan process. You should only use this option if you have
enough memory for all of your Fusion ioMemory devices in the system. If you do not have enough memory to use
this option, use the RSORT parameter instead.
Because this parameter requires a set amount of memory, it often works best with fewer Fusion ioMemory
devices and/or smaller-capacity Fusion ioMemory devices in a system, but the determining factor is how much
memory is in the system and whether there is enough to set the appropriate memory limit.

This parameter requires 4.008 bytes of RAM per block of Fusion ioMemory device capacity.
1. First determine the number of blocks that are formatted for each device.
a. This information is visible when you format the device using the fio-format utility.
b. Or you can estimate the number of block using the device capacity and the formatted sector size.
This example shows a quick estimation of the number of blocks on a 400GB device with 512B size
sectors (2 sectors per KB):
400GB * 1000MB/GB * 1000KB/MB * 2 Blocks/kB = 800,000,000 Blocks
2. Multiply the number of blocks by 4.008 bytes of RAM per block (and translate that into MiB) to determine
the memory limit that is required for this parameter to function.
a. In the example above there were 800 million blocks:
800,000,000 Blocks * 4.008B/Block * 1KiB/1024B * 1MiB/1024KiB =

~3058MiB of RAM
b. In this example, you would need about 3100 MiB of RAM available in your system for a 400GB
Fusion ioMemory device formatted for 512B sectors, and you would need to set the RMAP
parameter to 3100.
NOTE-
Default Value
The RMAP parameter is, by default, set to 3100. It is set to this low default value so the
rescan process does not use all of the RAM in systems that have less available memory.
l If the RMAP value is too low for the number of Fusion ioMemory device blocks in the system,
then the Fusion ioMemory VSL software will not use the RMAP process to improve rescan
times, it will just use the default fast rescan process. (RMAP is an all-or-nothing setting.)
l If you don't have enough system memory to use the RMAP parameter, consider using the
RSORT parameter. The RSORT parameter will use its RAM limit to improve the rescan
process, and then the Fusion ioMemory VSL software will revert to the default fast rescan
process to finish the consistency check.
3. Set the module parameter to the value you have determined. See Using Module Parameters on page 73 for
more information on setting parameters.
RSORT Parameter
The rsort_memory_limit_MiB parameter sets the memory (RAM) limit used by the Fusion ioMemory VSL
software to perform the RSORT rescan process. The RSORT rescan process is faster than the default rescan process
and we recommend using it to rescan devices that are used as datastores for databases.

If this parameter is given any memory limit, the Fusion ioMemory VSL software will use the RSORT process until
either the rescan is done or it consumes the memory limit. If the process runs out of memory, it will revert to the
default fast rescan process. However, in order to optimize the use of this process, you can calculate the target
RAM usage and set the limit based on that target. There is no penalty for setting a high limit, the RSORT process will
only use the RAM it needs (up to the limit that is set).
This target is based on 32 bytes per write extent. For example, if your database writes 16kB at a time, there is one
write extent per 16kB of Fusion ioMemory device capacity.
NOTE-
Blocks per Write Extent
One measure of the the benefits of the RSORT process is to see how many blocks are written per write
extent. The RSORT process improves rescan times over the default fast rescan process on when a device
has 8 or more blocks written per extent. For example, if your Fusion ioMemory device is formatted to
512B sector sizes (2 sectors per KB), and your database writes in 8KB chunks, then your database writes
16 blocks per write extent and RSORT would improve the rescan times.
1. First determine the number of blocks that are formatted for each device.
a. This information is visible when you format the device using the fio-format utility.
b. Or you can estimate the number of block using the total device capacities and their formatted sector
sizes.
This example shows a quick estimation of the number of blocks on 1200GB of Fusion ioMemory
device capacity with 512B size sectors (2 sectors per KB):
1200GB * 1000MB/GB * 1000KB/MB * 2 Blocks/kB = 2,400,000,000

Blocks
2. Divide the number of blocks by the write extents per block to determine the total possible number of write
extents on the device(s).
a. In the example above there were 2.4 billion blocks. We will assume 16KB write extents (32
blocks per write on 512B sectors):
2,400,000,000 Blocks * 1 Write Extent/32 Blocks = 150,000,000

Writes
3. Multiply the number of writes by 32 bytes of RAM per write (and translate that into MiB) to determine the
memory target for this parameter.
a. In the example above there were 150 million write extents:
150,000,000 Writes * 32B/Write * 1KiB/1024B * 1MiB/1024KiB =

~4578MiB of RAM

b. In this example, you would want to set the RSORT limit to about 4600 MiB of RAM available in your
system for 1200GB of Fusion ioMemory device capacity formatted for 512B sectors.
NOTE-
Default Value
The RMAP parameter is, by default, set to 0m and it has a maximim of 100000 (100GB).
4. Set the module parameter to the value you have determined. See Using Module Parameters on page 73 for
more information on setting parameters.

Appendix A - Command-line Utilities Reference
The Fusion ioMemory VSL software installation packages include various command-line utilities, installed by default
in /usr/sbin. These provide a number of useful ways to access, test, and manipulate your device.
ATTENTION!
Unified Utilities. If you plan to use the 4.2.1 (or later) utilities (Unified Utilities) on a VSL 3.2.15 server,
please refer to the 3.2.15 Dell Fusion ioMemory VSL Release Notes for information on which 3.2.15
operating systems can use Unified Utilities, and for differences between the 3.2.15 and 4.2.1 (or later)
utilities.
ATTENTION!
There are some additional utilities installed in the /usr/sbin directory that are not listed below. Those
additional utilities are dependencies (used by the main Fusion ioMemory VSL utilities), and you should not
use them directly unless Customer Support advises you to do so.
Utility Purpose
fio-
Makes a Fusion ioMemory device available to the OS.
attach
fio-
Lights the Fusion ioMemory device's external LEDs.
beacon
fio-
Prepares a detailed report for use in troubleshooting problems.
bugreport
fio-
Temporarily removes a Fusion ioMemory device from OS access.
detach
Reports the version and revision of the firmware on Fusion ioMemory devices without requiring
fio- you to load the Fusion ioMemory VSL software driver. By reporting the firmware version(s) on
firmware installed devices, this utility can help you determine which version of the Fusion ioMemory VSL
software to load.
fio-
Used to perform a low-level format of a Fusion ioMemory device.
format
fio-pci-
Checks for errors on the PCI bus tree, specifically for Fusion ioMemory devices.
check
fio-
Displays information about the device.
status

Utility Purpose
fio-sure-
Clears or purges data from the device.
erase
fio-
update- Updates the Fusion ioMemory device's firmware.
iodrive
NOTE-
There are -h (Help) and -v (Version) options for all of the utilities. Also, -h and -v cause the utility to
exit after displaying the information.
fio-attach
Description
Attaches the Fusion ioMemory device and makes it available to the operating system. This creates a block device in
/dev named fiox (where x is a, b, c, etc.). You can then partition or format the Fusion ioMemory device, or set
it up as part of a RAID array. The command displays a progress bar and percentage as it operates.
NOTE-
In most cases, the Fusion ioMemory VSL software automatically attaches the device on load and does a
scan. You only need to run fio-attach if you ran fio-detach or if you set the Fusion ioMemory VSL
software's auto_attach parameter to 0.
NOTE-
If the Fusion ioMemory device is in minimal mode, then auto-attach is disabled until the cause of the
device being in minimal mode is fixed.
Syntax
fio-attach <device> [options]
where <device> is the name of the device node (/dev/fctx), where x indicates the device number: 0, 1, 2,
etc. For example, /dev/fct0 indicates the first Fusion ioMemory device detected on the system.
You can specify multiple Fusion ioMemory devices. For example, /dev/fct1 /dev/fct2 indicates the second
and third Fusion ioMemory devices installed on the system.

Option Description
Force a metadata rescan. This may take an extended period of time, and is not normally required.
-r ATTENTION!
Only use this option when directed by Customer Support.
-c Attach only if clean.
-q Quiet: disables the display of the progress bar and percentage.
fio-beacon
Description
Lights the Fusion ioMemory device's LED(s) to locate the device.
Syntax
fio-beacon <device> [options]
where <device> is the name of the device node (/dev/fctx), where x indicates the card number: 0, 1, 2, etc.
For example, /dev/fct0 indicates the first Fusion ioMemory device detected on the system. The device
numbers are visible using fio-status.
Option Description
-0 Off: (Zero) Turns off the LED beacon.
-1 On: Lights the LED beacon.
Prints the PCI bus ID of the device at <device> to standard output. Usage and error information may be
-p
written to standard output rather than to standard error.
fio-bugreport
Description
Prepares a detailed report of the device for use in troubleshooting problems. The results are saved in the /tmp
directory in the file that indicates the date and time the utility was run.
Example:
/tmp/fio-bugreport-20100121.173256-sdv9ko.tar.bz2
Syntax
fio-bugreport

Notes
This utility captures the current state of the device. When a performance or stability problem occurs with the
device, run the fio-bugreport utility and contact Customer Support at http://support.dell.com for assistance
in troubleshooting.
Sample Output
-bash-3.2# fio-bugreport
Collecting fio-status -a
Collecting fio-status
Collecting fio-pci-check
Collecting fio-pci-check -v
Collecting fio-read-lebmap /dev/fct0
Collecting fio-read-lebmap -x /dev/stdout/dev/fct0
Collecting fio-read-lebmap -t /dev/fct0
Collecting fio-get-erase-count/dev/fct0
Collecting fio-get-erase-count -b /dev/fct0
Collecting lspci
Collecting lspci -vvvvv
Collecting lspci -tv
Collecting messages file(s)
Collecting procfusion file(s)
Collecting lsmod
Collecting uname -a
Collecting hostname
Collecting sar -r
Collecting sar
Collecting sar -A
Collecting syslog file(s)
Collecting proc file(s)
Collecting procirq file(s)
Collecting dmidecode
Collecting rpm -qa iodrive*
Collecting find /lib/modules
Please attach the bugreport tar file
/tmp/fio-bugreport-20090921.173256-sdv9ko.tar.bz2
to your support case, including steps to reproduce the problem.
If you do not have an open support case for this issue, please open a
support
case with a problem description and then attach this file to your new case.
For example, the filename for a bug report file named fiobugreport-20090921.173256-
sdv9ko.tar.bz2 indicates the following:
l Date (20090921)
l Time (173256, or 17:32:56)
l Misc. information (sdv9ko.tar.bz2)

fio-detach
Description
Detaches the Fusion ioMemory device and removes the corresponding fctx Fusion ioMemory device block
device from the OS. The fio-detach utility waits until the device completes all read/write activity before
executing the detach operation. By default, the command also displays a progress bar and percentage as it
completes the detach.
ATTENTION!
Before using this utility, ensure that the device you want to detach is NOT currently mounted and in use.
Syntax
fio-detach <device> [options]
For example, /dev/fct0 indicates the first Fusion ioMemory device detected on the system.
You can specify multiple Fusion ioMemory devices. For example, /dev/fct1 /dev/fct2 indicates the second
and third Fusion ioMemory devices installed on the system. You can also use a wildcard to indicate all Fusion
ioMemory devices on the system. For example, /dev/fct*
Option Description
-q Quiet: Disables the display of the progress bar and percentage.
Notes
Attempting to detach a Fusion ioMemory device may fail with an error indicating that the device is busy. This
typically may occur if the Fusion ioMemory device is part of a software RAID (0,1,5) volume, is mounted, or some
process has the device open.
The tools fuser, mount, and lsof can be helpful in determining what is holding the device open.
fio-firmware
Description
This utility is useful for diagnosing system issues. The utility reports the version and revision of the firmware on all
Fusion ioMemory devices without requiring you to load the Fusion ioMemory VSL software driver. By indicating the
firmware version(s) on installed devices, this utility can help you determine which version of the Fusion ioMemory
VSL software to load.
ATTENTION!
The 3.2.15 fio-firmware utility will not report firmware version information on Gen3 devices
including Fusion ioMemory SX300 devices, Fusion ioMemory SX350 devices, and Fusion ioMemory PX600

devices.
If you need to run fio-firmware across product generations, please use the latest VSL 4 fio-
firmware version.
Output is in CSV format, with each Fusion ioMemory device on a separate line:
<PCIe address>, <FW Version>, <FW Revision>
Syntax
fio-firmware [options]
Option Description
-f Print only x.y.z firmware version.
-n Disable printing header information on columns.
-p Print pending version of firmware to be activated on next reboot
-r Print only firmware revision.
fio-format
Description
NOTE-
Fusion ioMemory devices ship pre-formatted, so fio-format is generally not required except to
change the logical size or block size of a device, or to erase user data on a device. To ensure the user data
is truly erased, use fio-sure-erase, see fio-sure-erase on page 66 for more information.
Performs a low-level format of the Fusion ioMemory device. By default, fio-format displays a progress-
percentage indicator as it runs.
ATTENTION!
Use this utility with care, as it deletes all user information on the device. You will be prompted as to
whether you want to proceed with the format.
NOTE-
Using a larger block (sector) size, such as 4096 bytes, can significantly reduce worst-case Fusion
ioMemory VSL host memory consumption. However, some applications are not compatible with non-
512-byte sector sizes.

NOTE-
If you do not include the -s or -o options, the device size defaults to the advertised capacity. If used, the
-s and -o options must include the size or percentage indicators.
WARNING!
Do not interrupt the formatting! We recommend adding power backup to your system to prevent
power failures during formatting. If formatting is interrupted, please contact Customer Support.
Syntax
fio-format [options] <device>
where <device> is the name of the device node (/dev/fctx), where x indicates the device number: 0, 1, 2,
etc. For example, /dev/fct0 indicates the first Fusion ioMemory device detected on the system. Use fio-
status to view this number.
Options Description
-b <size Set the block (sector) size, in bytes or kibibytes (base 2). The default is 512 bytes. For
B|K> example: -b 512B or -b 4K (B in 512B is optional).
Force the format size, bypassing normal checks and warnings. This option may be needed in
-f rare situations when fio-format does not proceed properly. (The "Are you sure?"
prompt still appears unless you use the -y option.)
-q Quiet mode: Disable the display of the progress-percentage indicator.
Set the device capacity as a specific size (in TB, GB, or MB) or as a percentage of the
advertised capacity:
-s <size
l T Number of terabytes (TB) to format
M|G|T|%>
l G Number of gigabytes (GB) to format
l M Number of megabytes (MB) to format
l % Percentage, such as 70% (the percent sign must be included)
Over-format the device size (to greater than the advertised capacity), where the maximum
size equals the maximum physical capacity. If a percentage is used, it corresponds to the
maximum physical capacity of the device. (Size is required for the -o option; see the -s
-o <size option above for size indicator descriptions.)
B|K|M|G|T|%>
ATTENTION!
Before you use this option, please discuss your use case with Customer Support.
-R Disable fast rescan on unclean shutdown to reclaim some reserve capacity.
-y Auto-answer "yes" to all queries from the application (bypass prompts).

You must re-attach the device in order to use the Fusion ioMemory device. See fio-attach on page 57 for details.
fio-pci-check
Description
Checks for errors on the PCI bus tree, specifically for Fusion ioMemory devices. This utility displays the current
status of each Fusion ioMemory device. It also prints the standard PCI Express error information and resets the
state.
NOTE-
It is perfectly normal to see a few correctable errors when fio-pci-check is initially run. Subsequent
runs should reveal only one or two errors during several hours of operation.
Syntax
fio-pci-check [options]
Option Description
-d
1 = Disable the link; 0 = bring the link up (Not recommended).
<value>
-e Enable PCI-e error reporting.
-f Scan every device in the system.
-n Do not perform any writes to config space. Will prevent errors from being cleared.
Optimize the Fusion ioMemory device PCIe link settings by increasing the maximum read request
-o
size if it is too low.
-r Force the link to retrain.
-v Verbose: Print extra data about the hardware.
fio-status
Description
Provides detailed information about the installed devices. This utility operates on either fctx or fiox devices.
The utility depends on running as root and having the Fusion ioMemory VSL driver loaded. If no driver is loaded, a
smaller set of status information is returned.
fio-status provides alerts for certain error modes, such as a minimal-mode, read-only mode, and write-
reduced mode, describing what is causing the condition.
Syntax
fio-status [<device>] [<options>]

For example, /dev/fct0 indicates the first Fusion ioMemory device detected on the system.
If <device> is not specified, fio-status displays information for all cards in the system. If the Fusion
ioMemory VSL driver is not loaded, this parameter is ignored.
Option Description
-a Report all available information for each device.
Show all errors and warnings for each device. This option is for diagnosing issues, and it hides
-e
other information such as format sizes.
-c Count: Report only the number of Fusion ioMemory devices installed.
Show basic information set plus the total amount of data read and written (lifetime data
-d
volumes). This option is not necessary when the -a option is used.
Format JSON: creates the output in JSON format, which can be used for scripted management
-fj
applications.
-fx Format XML: creates the output in XML format.
-u Show unavailable fields. Only valid with -fj or -fx.
Show unavailable fields and details why. Only valid with -fj or -fx.
NOTE-
-U
Some fio-status fields are unavailable depending on the operating system or
device. For example, some legacy fields are unavailable on newer Fusion ioMemory
devices.
Print the value for a single field (see the next option for field names). Requires that a device be
-F<field>
specified. Multiple -F options may be specified.
-l List the fields that can be individually accessed with -F.
ATTENTION!
Output Change
The standard formatting of fio-status ouput has changed compared to the output from Fusion
ioMemory VSL software version 2.x. This will affect any custom management tools that used the output of
this utility.
Basic Information: If no options are used, fio-status reports the following basic information:
l Number and type of devices installed in the system

l Fusion ioMemory VSL software version
Adapter information:

l Adapter type
l Product number
l External power status
l PCIe power limit threshold (if available)
l Connected Fusion ioMemory devices
Block device information:
l Attach status
l Product name
l Product number
l Serial number
l PCIe address and slot
l Firmware version
l Size of the device, out of total capacity
l Internal temperature (average and maximum, since Fusion ioMemory VSL software load) in degrees
Centigrade
l Health status: healthy, nearing wearout, write-reduced or read-only
l Reserve capacity (percentage)
l Warning capacity threshold (percentage)
Data Volume Information: If the -d option is used, the following data volume information is reported in addition
to the basic information:
l Physical bytes written

l Physical bytes read
All Information: If the -a option is used, all information is printed, which includes the following information in
addition to basic and data volume information:
Adapter information:
l Manufacturer number
l Part number
l Date of manufacture
l Power loss protection status
l PCIe bus voltage (avg, min, max)
l PCIe bus current (avg, max)
l PCIe bus power (avg, max)
l PCIe power limit threshold (watts)
l PCIe slot available power (watts)
l PCIe negotiated link information (lanes and throughput)
Block device information:
l Manufacturer's code
l Manufacturing date
l Vendor and sub-vendor information
l Format status and sector information (if device is attached)
l FPGA ID and Low-level format GUID

l PCIe slot available power
l PCIe negotiated link information
l Card temperature, in degrees Centigrade
l Internal voltage (avg and max)
l Auxiliary voltage (avg and max)
l Percentage of good blocks, data and metadata
l Lifetime data volume statistics
l RAM usage
Error Mode Information: If the Fusion ioMemory VSL software is in minimal mode, read-only mode, or write-
reduced mode when fio-status is run, the following differences occur in the output:
l Attach status is "Status unknown: Driver is in MINIMAL MODE:"

l The reason for the minimal mode state is displayed (such as "Firmware is out of date. Update firmware.")
l "Geometry and capacity information not available." is displayed.
l No media health information is displayed.
fio-sure-erase
ATTENTION!
As a best practice, do not use this utility if there are any Fusion ioMemory devices installed in the system
that you do not want to clear or purge. First remove any devices that you do not want to accidentally
erase. Once the data is removed with this utility it is gone forever. It is not recoverable.
WARNING!
Before you use this utility, be sure to back up any data that you wish to preserve.
WARNING!
Do not interrupt fio-sure-erase (by issuing a forced kill command, or powering down or resetting
the system)! Interrupting fio-sure-erase could cause the Fusion ioMemory device to become
unusable.
ATTENTION!
If the device is in Read-only mode, perform a format using fio-format before running fio-sure-
erase. If the device is in Minimal mode, then fio-sure-erase cannot erase the device. Updating
the firmware may take the device out of Minimal Mode. If the device remains in Minimal mode, contact
Customer Support at http://support.dell.com for further assistance.
In order to run fio-sure-erase, the block device must be detached. See fio-detach on page 60 section for
more information.
Description
The fio-sure-erase is a command-line utility that securely removes data from Fusion ioMemory devices.

Syntax
fio-sure-erase [options] <device>
Where <device> is the name of the device node (/dev/fctx), where x indicates the card number: 0, 1, 2, etc.
For example, /dev/fct0 indicates the first Fusion ioMemory device detected on the system. Use fio-status
to view this device node, see fio-status on page 63.
NOTE-
Products with Multiple Devices
fio-sure-erase works on individual Fusion ioMemory devices. For example, if you are planning to
purge a Fusion ioMemory ioDrive Duo device, you will need to perform this operation on each of the
product's two Fusion ioMemory devices.
Option Description
Purge instead of Clear: performs a write followed by an erase. For more information on Purge, see
below.
-p ATTENTION!
Purging the device may take hours to accomplish, depending on the size of the device that
needs to be purged.
-y No confirmation: do not require a yes/no response to execute the utility.
-t Do not preserve current format parameters, including device and sector size (reset to default).
-q Quiet: do not display the status bar.
NOTE-
If you run fio-sure-erase with no options, a Clear is performed. For more information, see below.
When the utility completes, each block of memory consists of uniform 1 bits or 0 bits.
Clear Support
A "Clear" is the default state of running fio-sure-erase (with no options), and refers to the act of performing a
full low-level erase (every cell pushed to "1") of the entire NAND media, including retired erase blocks.
Metadata that is required for operation will not be destroyed (media event log, erase counts, physical bytes
read/written, performance and thermal history), but any user-specific metadata will be destroyed.
The following describes the steps taken in the Clear operation:

1. Creates a unity map of every addressable block (this allows fio-sure-erase to address every block,
including previously unmapped bad blocks).
2. For each block, performs an erase cycle (every cell is pushed to "1").
3. Restores the bad block map.
4. Formats the device (the purpose of this is to make the device usable again, the utility erases all of the
headers during the clear).
Purge Support
A "Purge" is invoked by using the -p option with fio-sure-erase. Purge refers to the act of first overwriting
the entire NAND media (including retired erase blocks) with a single character (every cell written to logical "0"),
and then performing a full chip erase (every cell pushed to "1") across all media (including retired erase blocks).
Metadata that is required for operation will not be destroyed (media event log, erase counts, physical bytes
read/written, performance and thermal history), but any user-specific metadata will be destroyed.
The following describes the steps taken in the Purge operation:
1. Creates a unity map of every addressable block (this allows fio-sure-erase to address every block,
including previously unmapped bad blocks).
2. For each block, performs a write cycle (every cell written to "0").
3. For each block, performs an erase cycle (every cell pushed to "1").
4. Restores the bad block map.
5. Formats the drive (the purpose of this is to make the drive usable again, the utility erases all of the headers
during the clear).
fio-update-iodrive
ATTENTION!
You should back up the data on the Fusion ioMemory device prior to any upgrade as a precaution.
Description
Updates the Fusion ioMemory device's firmware. This utility scans the PCIe bus for all Fusion ioMemory devices
and updates them. A progress bar and percentage are shown for each device as the update completes.
WARNING!
It is extremely important that the power not be turned off during a firmware upgrade, as this could
cause device failure. If a UPS is not already in place, consider adding one to the system prior to
performing a firmware upgrade.
WARNING!
Note that when running multiple firmware upgrades in sequence, it is critical to reboot and then both
load the Fusion ioMemory VSL driver and attach the Fusion ioMemory device(s) after each firmware
upgrade step. Otherwise the on-device format will not be changed, and there will be data loss.

WARNING!
Do not use this utility to downgrade the Fusion ioMemory device to an earlier version of the firmware.
Doing so may result in data loss and void your warranty. Contact Customer Support at
http://support.dell.com if you have issues with your upgrade.
NOTE-
The default action (without using the -d or -s option) is to upgrade all Fusion ioMemory devices with the
firmware contained in the fio-firmware-dell_<version>.<date>.fff firmware archive file.
Confirm that all devices need the upgrade prior to running the update. If in doubt, use the -p (Pretend)
option to view the possible results of the update.
ATTENTION!
You must detach all Fusion ioMemory devices before updating the firmware.
ATTENTION!
Upgrade Path
There is a specific upgrade path that you must take when upgrading Fusion ioMemory device. Consult the
Dell Fusion ioMemory VSL Release Notes for this Fusion ioMemory VSL software release before upgrading
Fusion ioMemory devices.
NOTE-
If you receive an error message when updating the firmware that instructs you to update the midprom
information, contact Customer Support.
To update one or more specific devices:
l If the Fusion ioMemory VSL driver is loaded, use the -d option with the device number.
Syntax
fio-update-iodrive [options] <firmware-path>
where <firmware-path> is the full path to the firmware archive file fio-firmware-dell_
<version>.<date>.fff available at http://dell.fusionio.com . If you downloaded the .fff firmware archive
file, then the firmware is most likely with the other downloaded packages. If you installed the firmware from the
firmware package, the default path is /usr/share/fio/firmware/. This parameter is required.

Option Description
Updates the specified devices (by fctx, where x is the number of the device shown in fio-
status). If this option is not specified, all devices are updated.
-d ATTENTION!
Use the -d or -s option with care, as updating the wrong Fusion ioMemory device
could damage your device.
Force upgrade (used when directed by Customer Support). If the Fusion ioMemory VSL driver
is not loaded, this option also requires the -s option.
-f
ATTENTION!
Use the -f option with care, as it could damage your card.
-l List the firmware available in the archive.
-p Pretend: Shows what updates would be done. However, the actual firmware is not modified.
-c Clears locks placed on a device.
-q Runs the update process without displaying the progress bar or percentage.
-y Confirm all warning messages.
Updates the devices in the specified slots using '*' as a wildcard for devices. The slots are
identified in the following PCIe format (as shown in lspci):
-s
[[[[<domain>]:]<bus>]:][<slot>][.[<func>]]
--split Split the Fusion ioMemory device into virtual devices.
--merge Merge the virtual devices of a Fusion ioMemory device.
If you arrived at this section from Upgrading the Firmware on page 23, you should return to that section.

Appendix B - Monitoring the Health of Devices
This section describes how the health of Fusion ioMemory devices can be measured and monitored in order to
safeguard data and prolong device lifetime.
Health Metrics
The Fusion ioMemory VSL software manages block retirement using pre-determined retirement thresholds. The
Dell ioSphere software and the fio-status utilities show a health indicator that starts at 100 and counts down
to 0. As certain thresholds are crossed, various actions are taken.
At the 10% healthy threshold, a one-time warning is issued. See Health Monitoring Techniques on page 71 for
methods for capturing this alarm event.
At 0%, the device is considered unhealthy. It enters write-reduced mode, which somewhat prolongs its lifespan so
data can be safely migrated off. In this state the Fusion ioMemory device behaves normally, except for the reduced
write performance.
After the 0% threshold, the device will soon enter read-only mode -- any attempt to write to the Fusion ioMemory
device causes an error. Some filesystems may require special mount options in order to mount a read-only block
device in addition to specifying that the mount should be read-only.
For example, under Linux, ext3 requires that "-o ro,noload" is used. The "noload" option tells the
filesystem to not try and replay the journal.
Read-only mode should be considered a final opportunity to migrate data off the device, as device failure is more
likely with continued use.
The Fusion ioMemory device may enter failure mode. In this case, the device is offline and inaccessible. This can be
caused by an internal catastrophic failure, improper firmware upgrade procedures, or device wearout.
NOTE-
For service or warranty-related questions, contact Customer Support at http://support.dell.com .
NOTE-
For products with multiple Fusion ioMemory devices, these modes are maintained independently for
each device.
Health Monitoring Techniques

fio-status -a: Output from the fio-status utility (using the -a option) shows the health percentage and
device state. These items are referenced as "Media status" in the sample output below.

Found 3 ioMemory devices in this system
Fusion driver version: 3.x.x build xxxx
Adapter: Single Adapter

Fusion-io ioDrive 1.30TB, Product Number:F00-001-1T30-CS-0001,
SN:1133D0248, FIO SN:1134D9565
...
Media status: Healthy; Reserves: 100.00%, warn at 10.00%; Data: 99.12%
Lifetime data volumes:
Physical bytes written: 6,423,563,326,064
Physical bytes read : 5,509,006,756,312
The following Health Status messages are produced by the fio-status utility:
l Healthy
l Read-only
l Reduced-write
l Unknown
Dell ioSphere software: In the Device Report tab, look for the Reserve Space percentage in the right column. The
higher the percentage, the healthier the drive is likely to be.
Software RAID and Health Monitoring

Software RAID stacks are typically designed to detect and mitigate the failure modes of traditional storage media.
The Fusion ioMemory device attempts to fail as gracefully as possible, and these new failure mechanisms are
compatible with existing software RAID stacks. A Fusion ioMemory device in a RAID group will fail to receive data at
a sufficient rate if: a) the device is in a write-reduced state, and b) it is participating in a write-heavy workload. In
this case, the device will be evicted from the RAID group. A device in read-only mode will be evicted when write
I/Os are returned from the device as failed. Catastrophic failures are detected and handled just as though they are
on traditional storage devices.

Appendix C - Using Module Parameters
The following table describes the module parameters you can set by editing the
/etc/modprobe.d/iomemory-vsl.conf file and changing their values.
Each module parameter in the configuration file must be preceded by options iomemory-vsl. The
/etc/modprobe.d/iomemory-vsl.conf file has some example parameters that are commented out. You
may use these examples as templates and/or uncomment them in order to use them.
NOTE-
These changes must be completed before the Fusion ioMemory VSL software is loaded in order to take
effect.
Module Parameter Default (min/max) Description

1 (default) = Always attach the device on driver
load.
auto_attach 1 (0, 1)
0 = Don't attach the device on driver load.
Allows designated device(s) to draw the specified

amount (s) of power in Watts from the PCIe slot.
Where the <SN-value>:<W-value> for this
parameter is a comma-separated list of value pairs
with adapter serial numbers and the maximum
amount of power each device should pull (in
Watts).
external_power_
No devices selected
override
ATTENTION!
You can use a serial number wildcard (*)
to set this module parameter for all cards
in the system. Please see Enabling PCIe
Power Override on page 25 for more
information.
Number of seconds to wait for /dev/fio* files

to show up during driver load. For systems not
fio_dev_wait_timeout_
30 using udev, this should be set to 0 to disable the
secs
timeout and avoid an unneeded pause during
driver load.

Module Parameter Default (min/max) Description
1 = Force minimal mode on the device.
force_minimal_mode 0
0 = Do not force minimal mode on the device.
numa_node_forced_ 1 = Enable, 0 = Disable. See NUMA Configuration

0
local on page 75 for more information.
A list of <affinity specification>

couplets that specify the affinity settings of all
devices in the system. Each item in the couplet is
separated by a colon, and each couplet set is
separated by a comma. Where each <affinity
specification> couplet has the following
numa_node_override Nothing Selected
syntax:
<device-id>=<node-number>
See NUMA Configuration on page 75 for more

information on using this parameter.
1 = Enable parallel attach of multiple devices.

parallel_attach 1
0 = Disable parallel attach of multiple devices.
For the selected devices, pre-allocate all memory

necessary to have the drive usable as swap space.
preallocate_memory No devices selected
Where the <value> for this parameter is a
comma-separated list of device serial numbers.
Interval (microseconds) to wait between hardware

tintr_hw_wait 0 (0, 255) interrupts. Also known as interrupt coalescing. 0 is
off.
Linux only: 3 = use standard OS I/O elevators; 0 =

use_workqueue 0 (0 or 3)
bypass.
NOTE-
Other than external_power_override and preallocate_memory, module parameters are
global — they apply to all Fusion ioMemory devices in the computer.

Appendix D - NUMA Configuration
About NUMA Architecture

Servers with a NUMA (Non-Uniform Memory Access) architecture may require special installation instructions in
order to maximize Fusion ioMemory device performance. This includes most multi-socket servers.
On some servers with NUMA architecture, during system boot the BIOS will not associate PCIe slots with the
correct NUMA node. Incorrect mappings result in inefficient I/O handling that can significantly degrade
performance. Here are two methods for correcting the NUMA affinity:
l numa_node_forced_local Parameter: Forces I/O completions to happen on the same CPU that is
running VSL processes for a particular device. This parameter is simple to implement (enabled or disabled),
and is persistent.
l numa_node_override Parameter: Used to manually assign devices to specific NUMA nodes. This
parameter is more complex. For optimal implementation, you should understand the NUMA architecture
of the system in order to assign devices to the nodes that are closely linked to the device's PCIe slot. This
parameter is persistent until any new I/O devices are installed in the system (thus changing the PCIe bus
numbers in the system.
Using the numa_node_forced_local Parameter

The numa_node_forced_local parameter is either enabled or disabled, and therefore does not offer as
much user control as other options. It is persistent and it is enabled by modifying the
/etc/modprobe.d/iomemory-vsl.conf file and editing or adding the following line:
numa_node_forced_local=1
This parameter forces an I/O completion for a particular device to happen on a CPU within the local numa_node
that the other Fusion ioMemory VSL processes are running on (rather than trying to complete an I/O on the CPU
that the host issued it on). Because the I/O completions are grouped with other Fusion ioMemory VSL processes,
they are less likely to compete with processes from other device drivers.
This parameter may or may not improve performance depending on your configuration and workloads. You should
test this parameter with your use case to determine if it improves performance.
Using the numa_node_override Parameter

Use this parameter to map devices with specific NUMA nodes.
ATTENTION!
The example below shows the final implementation of custom affinity settings. This implementation
required an analysis of the specific system, including the system architecture, type and number of Fusion

ioMemory devices installed, and the particular PCIe slots that were used. Your particular circumstances
will require a custom analysis of your set-up. This analysis requires understanding of your system's NUMA
architecture compared to your particular installation.
Your actual settings may be different than the example below, depending on your server configuration. In order to
create the correct settings for your specific system, use fio-status to list all of the devices and determine the
<device-id> (see below). Then use the example below of setting the numa_node_override parameter as a
template and modify it for your particular system.
Determining the Device ID

You should present each <device-id> in the following format:
<domain>:<bus>:<device>.<function>
Typically the domain is 0000, consult your server documentation to determine your device ID. The remainder of
the device ID string is visible in fio-status output. For example:
# fio-status
Found 2 ioMemory devices in this system
...
PCI:04:00.0
...
PCI:15:00.0
In the example above the device IDs would be 0000:04:00.0 and 0000:15:00.0 on a system that had a
domain of 0000.
WARNING!
Note that the PCI device ID, including the bus number, may change if you change any of the PCI devices in
the system. For example, if you add a network card or another Fusion ioMemory device. If the device ID
changes, you will have to update the configuration.
numa_node_override Parameter
Configuring your Fusion ioMemory devices for servers with NUMA architecture requires the use of the numa_
node_override parameter by modifying the iomemory-vsl.conf file.
The numa_node_override parameter is a list of <affinity specification> couplets that specify the
affinity settings of all devices in the system. Each item in the couplet is separated by an equal sign (=), and each
couplet set is separated by a comma.
Syntax:
numa_node_override=<affinity specification>[,<affinity specification>...]
Where each <affinity specification> has the following syntax:

<device-id>=<node-number>
Simple Example:
numa_node_
override=0000:04:00.0=1,0000:1d:00.0=0,0000:05:00.0=2,0000:1e:00.0=3
Has the effect of creating :
<device-id> Node/Group Processor Affinity

0000:04:00.0 node 1 all processors in node 1
0000:1d:00.0 node 0 all processors in node 0
0000:1e:00.0 node 3 all processors in node 3
Advanced Configuration
If your server has multiple NUMA nodes and multiple Fusion ioMemory devices installed, you will need to make
sure that the Fusion ioMemory devices are spread out among the various nodes.
While it may be optimal to pair devices to nodes that are electronically closer each device's PCIe slot (which would
require an advanced understanding of your server's NUMA architechure and an analysis of the device installation),
just simply spreading out all of the devices' node affinity among the available nodes should result in improved
performance.
In the example above the device IDs would be 0000:04:00.0 and 0000:15:00.0 on a system that had a
domain of 0000.
WARNING!
Note that the PCI device ID may change if you change any of the PCI devices in the system. For example, if
you add a network card or another Fusion ioMemory device. If the device ID changes, you will have to
update the configuration.
<device-id> Node/Group Processor Affinity

0000:1d:00.0 node 0 all processors in node 0
0000:1e:00.0 node 3 all processors in node 3

Appendix E - Upgrading Devices from VSL 2.x to 3.x
This version of the Fusion ioMemory VSL software supports new features, including the latest generation of
ioMemory architecture and improved Flashback protection. These features require the latest version of the Fusion
ioMemory device firmware. Every Fusion ioMemory device in a system running 3.1.x or later must be upgraded to
the latest version of the firmware.
For example, if you have a system running 2.x Fusion ioMemory VSL software with Fusion ioMemory ioDrive
devices previously installed, and you want to install new Fusion ioMemory ioDrive2 devices (that require the latest
version of the firmware), then you will need to upgrade all of the existing devices to the latest firmware version.
WARNING!
You cannot revert a device's firmware to an earlier version once you have upgraded the device (without
voiding your warranty). If you experience problems with your upgrade, please contact Customer Support
at http://support.dell.com .
WARNING!
Upgrading devices (previously configured for VSL 2.x.x) to work with VSL 3.x.x will require a low-level
media format of the device. No user data will be maintained during the process. Be sure to back up all
data as instructed.
ATTENTION!
Upgrade Path
Depending on the current firmware version of your devices, you may need to upgrade your device's
firmware multiple times in order to preserve internal structures. Consult the Dell Fusion ioMemory VSL
Release Notes for the upgrade path. Visit http://dell.fusionio.com for all of the required software and
firmware versions.
For more information on upgrading from one version to the next, see the Dell Fusion ioMemory VSL Release Notes
(available at http://dell.fusionio.com ) for the version to which you will upgrade the device. Then follow the
upgrade instructions in that version's user guide for your operating system (including the firmware update
instructions).
Upgrade Procedure
While upgrading from VSL 2.x to 3.x, ensure that you follow the device upgrade path as described in the Dell Fusion
ioMemory VSL Release Notes. The device upgrade path outlines the specific order of firmware and VSL driver
upgrades required for the VSL 2.x to 3.x upgrade.

Make sure that all previously installed Fusion ioMemory ioDrive devices are updated with the appropriate
firmware and VSL software versions required for this upgrade.
ATTENTION!
Make sure to upgrade all existing Fusion ioMemory ioDrive devices to VSL 3.x before installing a new
Fusion ioMemory ioDrive device in the same host.
If you install a new Fusion ioMemory ioDrive device before upgrading existing devices to VSL 3.x, the OS
may not recognize the new device. This means that when you enter the fio-status command, the OS
may return incomplete or incorrect information about the new device. In addition, the VSL 2.x driver
may put the new device into minimal mode until the upgrade process is completed for the existing
devices.
1. Prepare each existing Fusion ioMemory ioDrive device for upgrade.
a. Back up user data on each device.
WARNING!
The upgrade process will require a low-level media format of the device. No user data will
be retained during the process; be sure to make a complete backup.
Use a backup method of your choice. For best results, use software and backup devices that have
proven effective in the past. Do not back up the data onto another Fusion ioMemory device on the
same system. The backup must be to a local disk or to an externally attached volume.
b. Run the fio-bugreport utility and save the resulting file. This will capture the device information
for each device in the system. This device information will be useful in troubleshooting any upgrade
issues. Sample command:
fio-bugreport
c. Detach Fusion ioMemory ioDrive devices, for example:
fio-detach /dev/fct*
For more information, see fio-detach on page 60.
2. Unload the current Fusion ioMemory VSL driver, for example:
$ modprobe -r iomemory-vsl
3. Uninstall the 2.x Fusion ioMemory VSL software.

a. Uninstall the utilities:
l Sample RPM command:
$ rpm -e fio-util fio-snmp-agentx fio-common fio-firmware

iomanager-gui iomanager-jre libfio libfio-doc libfusionjni
fio-sysvinit fio-smis fio-snmp-mib libfio-dev
b. To uninstall the software, you must specify the kernel version of the package you are uninstalling.
Run the appropriate command to find the installed packages:
l RPM command:
$ rpm -qa | grep -i iomemory
l Sample output:
iomemory-vsl-2.6.18-194.el5-2.2.2.82-1.0
c. Uninstall the Fusion ioMemory VSL software by running a command similar to this example (specify
the kernel version of the package you wish to uninstall):
l Sample RPM command:
$ rpm -e iomemory-vsl-2.6.18-194.el5-2.2.0.82-1.0
4. Install the new Fusion ioMemory VSL software and related packages.
a. Download the Fusion ioMemory VSL software binary package for your kernel and all supporting
packages at http://dell.fusionio.com
NOTE-
If you don't see a binary for your kernel, follow the instructions in the Building the
ioMemory VSL from Source section of Installing RPM Packages on page 13. To see your
current kernel version, run: uname -r
b. Install the Fusion ioMemory VSL software and utilities using the appropriate commands:
l RPM commands:
rpm -Uvh iomemory-vsl-<kernel-version>-<VSL-version>.x86_

64.rpm
rpm -Uvh lib*.rpm
rpm -Uvh fio*.rpm

l See Installing RPM Packages on page 13for full instructions on installing those packages.
c. Reboot the system.
5. Update the firmware on each device to the latest version using fio-update-iodrive.
WARNING!
Prevent Power Loss
Take measures to prevent power loss during the update, such as using a UPS. Power loss during a
firmware update may result in device failure. For all warnings, alerts, and options pertaining to this
utility, see fio-update-iodrive on page 68.
Sample syntax:
fio-update-iodrive <firmware-path>
Where <firmware-path> is the full path to the firmware archive file (fio-firmware-dell_
<version>.<date>.fff) available at http://dell.fusionio.com . This command will update all of the
devices to the selected firmware. If you wish to update specific devices, consult fio-update-iodrive on page
68 for more options.
6. Reboot the system
7. Load the Fusion ioMemory VSL software, for example:
$ modprobe iomemory-vsl
For more information, see Loading the ioMemory VSL Driver on page 19.
NOTE-
If run, fio-status will warn that the upgraded devices are missing a lebmap. This is expected,
and will be fixed in the next step.
8. Destructive Step
WARNING!
Running fio-format in the next step will erase the entire device, including user data. Once this
format is started, the device cannot be downgraded to the 2.x driver without voiding your
warranty. If you experience problems with your upgrade, please contact Customer Support at
http://support.dell.com .
9. Format each device using fio-format, for example:
fio-format <device>

You will be prompted to confirm you wish to erase all data on the device.
ATTENTION!
The format may take an extended period of time, depending on the wear on the device.
10. Attach all Fusion ioMemory ioDrive devices, for example:
fio-attach /dev/fct*
11. Check the status of all devices using fio-status, for example:
fio-status -a
Your Fusion ioMemory ioDrive devices have now been successfully upgraded for this version of the Fusion
ioMemory VSL software. You may now safely install any new Fusion ioMemory ioDrive devices.

Dell Support
Contacting Dell
Dell provides several online and telephone-based support and service options. Availability varies by country and
product, and some services may not be available in your area. To get help with your ioDrive and ioDrive Duo,
contact your Dell Technical Service representative or access the Dell Support website at http://support.dell.com.
Choose the method of contacting Dell that is convenient for you.
NOTE-
The safety information that shipped with your system provides important safety and regulatory
information. Warranty information may be included within this document or as a separate document. To
access Dell documentation, go to http://support.dell.com/manuals.
® products.
Western Digital Technologies, Inc. is the seller of record and licensee in the Americas of SanDisk

Dell IoMemory VSL 3.2.15 User Guide For Linux 2017-04-11

Uploaded by

Copyright:

Available Formats

Dell IoMemory VSL 3.2.15 User Guide For Linux 2017-04-11

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Dell IoMemory VSL 3.2.15 User Guide For Linux 2017-04-11

Uploaded by

Copyright:

Available Formats

Dell Fusion ioMemory VSL 3.2.

Tuesday, April 11, 2017

Dell Fusion ioMemory VSL 3.2.15 User Guide for Linux 7

CAUTION— Safety Instructions 8

Protecting Against Electrostatic Discharge 8

SAFETY— When Working Inside Your System 9

About the ioMemory Platform 10

Fusion ioMemory VSL Software Installation 12

Installing RPM Packages 13

Loading the ioMemory VSL Driver 19

Controlling Driver Loading 20

Using the init Script 20

Mounting Filesystems when Using the init Script 20

On either udev or init script systems 22

Fusion ioMemory devices and Multipath Storage 22

Dell Fusion ioMemory VSL 3.2.15 User Guide for Linux 2

Upgrading the Firmware 23

Optional GUI - Dell ioSphere software 24

Setting the ioMemory VSL Options 25

Enabling PCIe Power Override 25

Enabling the Override Parameter 27

Determine Serial Number(s) 27

Setting the Parameter in VSL 3.2.15 28

Virtual Controller Configuration 29

Using the Device as Swap 32

Setting the Amount of Preallocated Memory 33

Using the Logical Volume Manager 34

Mounting Filesystems when Using the init Script 34

Configuring RAID using the Logical Volume Manager 34

Creating a Striped Volume (RAID 0) Using LVM 35

Creating a Mirrored Volume (RAID 1) Using LVM 35

Configuring RAID Using mdadm 36

Dell Fusion ioMemory VSL 3.2.15 User Guide for Linux 3

Building a RAID10 Across Multiple Devices 38

Discard (TRIM) Support 39

Discard (TRIM) on Linux 39

Performance and Tuning 41

Disable CPU Frequency Scaling 41

Limiting ACPI C-States 41

Setting ACPI C-State Options 41

C-States Under Linux 42

Setting NUMA Affinity 42

Setting the Interrupt Handler Affinity 42

Linux IRQ Balancing 42

Monitoring and Managing Devices 44

Example Conditions to Monitor 45

Health Reserves Percentage 46

Write (Health Reserves) Status 47

Device LED Indicators 47

Uninstalling the Software 49

Dell Fusion ioMemory VSL 3.2.15 User Guide for Linux 4

Uninstalling 2.x Support Packages 49

Uninstalling 3.x Support Packages 49

Uninstalling the Fusion ioMemory VSL RPM Package 49

Unloading the Software Driver 50

Upgrading the Kernel 50

Disabling the ioMemory VSL Software 50

Unmanaged Shutdown Issues 51

Improving Rescan Times 51