Provisioning a New Linux Host

 

Topic

Setting up a host environment

Selections

Select a host: Linux

Select a Linux task: Provisioning a new host

Select a HBA type: QLogic

Select a switch type: Cisco

 

 

Contents

·             Installing QLogic HBA(s) on a Linux host 4

·             Updating QLogic HBA firmware for Linux. 4

·             Enabling an in-kernel QLogic driver for Linux. 9

·             Downloading latest QLogic driver for Linux. 10

·             Installing a QLogic HBA driver on Linux host 12

·             Determine WWN of new HBA for Linux. 16

·             Creating switch zones. 16

·             Adding Symmetrix Devices. 18

·             Mapping Devices within Symmetrix. 23

·             Masking Devices. 28

·             Creating a Device Group.. 32

·             Making LUNs or Devices available to Linux. 34

·             Installing PowerPath software on a Linux host 37

 


 

Provisioning a new Linux host

This document for Linux describes how to setup a new host for the Symmetrix storage management environment. A series of procedures take you from installing HBAs, HBA firmware and drivers, zoning, mapping/masking devices, to configuring kernel files and LUNs.

 

Install/Setup QLogic HBA/Driver Environment on Linux

Using various procedures, this document for Linux, featuring a QLogic HBA environment describes how to install and setup host bus adapters and other aspects of a host environment that deal with managing devices of a storage system. A series of procedures take you from installing HBAs, to installing and setting up HBA firmware and drivers.

 

 

·        Installing QLogic HBA(s) on a Linux host

 

Overview

This procedure describes how to install HBAs on a Linux host.

 

This procedure is based on content from the following EMC manuals:

  • EMC Host Connectivity Guide for Linux
  • EMC Support Matrix Linux (RHEL or SuSE)

 

You can download these guides from EMC Online Support (registration required): https://support.EMC.com

 

 

1.   

Install Adapter in card slot

With host system power removed, install the HBA(s) by following the instructions included with your adapter. The adapter installs into a single slot and has no user-configurable jumpers or switches.

 

Repeat the installation steps for each host in which you are installing Fibre Channel HBAs.

 

 

2.   

Apply power and boot

Reapply power and allow the system to boot normally.

 

 

 

7Table of Contents

 

·        Updating QLogic HBA firmware for Linux

 

Overview

This procedure describes how to update the QLogic HBA firmware on a Linux host.

 

This procedure is based on content from the following EMC manuals:

  • EMC Host Connectivity Guide for Linux
  • EMC Support Matrix Linux (RHEL or SuSE)

 

You can download these guides from EMC Online Support (registration required): https://support.EMC.com

 

Ensure that the HBA firmware (FCode) is current on the adapter before proceeding with other host setup tasks.

 

 

1.   

Verify Current Firmware

Verify the correct BIOS version is present after the HBA is installed:

Note:  Refer to the E-Lab Interoperability Navigator (ESM) for required BIOS versions for qualified HBAs.

a.     As the host is booting, watch for the QLogic banner.  As soon as the QLogic banner appears, press Qtrl-Q to enter Fast!Util.

b.    Select Configuration Settings from the Fast!Util Options menu.

c.     Select Adapter Settings from the Configuration Settings menu.

d.    Under Adapter Settings, note the BIOS version.

e.     Perform one of the following depending on the BIOS version:

·            If the version displayed is not the required version, proceed to the next step to upgrade the QLogic HBA BIOS and then continue to set the data rate and topology (if necessary).

·            If the version displayed is the required version, skip the next step to upgrade the QLogic HBA BIOS and then continue to set the data rate and topology (if necessary).

 

 

2.   

Upgrade the BIOS if required:

Upgrade the QLogic HBA BIOS (if necessary):

 

There are three methods that can be used to upgrade the HBA BIOS (instructions follow below):

 

·         QLogic SANsurfer

·         QLogic SANsurferCLI

·         NVRAM file on a DOS-bootable diskette

 

 

Using QLogic SANsurfer:

 

The SANsurfer GUI may be downloaded from the EMC-approved section of the QLogic website at http://www.qlogic.com.  Click Downloads on the left side of the screen.  Click the EMC link under OEM Models.  On the next page, select EMC SYMMETRIX, CLARiiON, & CELERRA supported software.  Select the HBA model in the Product column of the table.  On the next table, find the correct SANsurfer file and click the associated Download link.


To update the BIOS using the SANsurfer, refer to the QLogic provided documentation on their website for detailed instructions.

To invoke the GUI, run the following command from a terminal window:

/opt/QLogic_Corporation/SANsurfer/SANsurfer &

Note:  When upgrading the BIOS for QLogic 4GB HBAs, the minimum required version of the QLogic SANsurfer GUI is 2.0.30b52.

 

 

 

Using QLogic SANsurferCLI:

 

The SANsurferCLI (scli) is installed as part of the qlinstaller or may be downloaded from the EMC-approved section of the QLogic website at http://www.qlogic.com.  Click Downloads on the left side of the screen.  Click the EMC link under OEM Models.  On the next page, select EMC SYMMETRIX, CLARiiON, & CELERRA supported software.  Select the HBA model in the Product column of the table.  On the next table, find the correct SANsurfer file and click the associated Download link.


To update the BIOS using the SANsurferCLI, refer to the QLogic provided documentation on their website for detailed instructions.

To invoke the CLI, run the following command:

/opt/QLogic_Corporation/SANsurferCLI/scli

Note:  When upgrading the BIOS for QLogic 4GB HBAs, the minimum required version of the QLogic SANsurfer CLI is 1.06.16 build 23.

 

 

 

Using the NVRAM file on a DOS-bootable diskette:

 

a.      Obtain the latest version of the BIOS from the QLogic Website at http://www.qlogic.com.  Click Downloads on the left side of the screen.  Click the EMC link under OEM Models.  On the next page, select EMC SYMMETRIX, CLARiiON, & CELERRA supported software.  Select the HBA model in the Product column of the table.  On the next table, find the correct BIOS file and click the associated Download link.

b.      The file is a self-extracting .zip file that must be extracted to a bootable diskette drive by inserting a diskette into a Microsoft Windows 9x machine.  Open any DOS window and at the DOS prompt, format the diskette by entering the following command:

format /s a:

At the DOS prompt, change (cd) to the directory in which you saved the zipped file and extract the file to the diskette.

c.      Verify the QLogic HBA(s) are installed in the host before proceeding.

d.      Insert the BIOS upgrade installation diskette into the floppy drive.

e.      Reboot the host.

f.       After the host has rebooted, a DOS prompt appears.  Type flasutil /L /F and press Enter.

Note:  The HBA BIOS upgrade might take a few minutes.

After the upgrade is complete, remove the diskette and reboot the host.  During startup, the QLogic banner should display the required BIOS version.

 

3.   

Set BIOS/NVRAM parameters:

EMC requires configuring the QLogic BIOS settings with the EMC-approved NVRAM settings file.  This file contains all of the BIOS settings for the QLogic adapters that have been tested and approved for connection to EMC storage arrays. 

 

Note:  Refer to the E-Lab Interoperability Navigator (ESM) for required NVRAM versions for qualified HBAs.

 

For a complete list of the QLogic parameters and their pre-configured EMC settings, refer to the appropriate EMC Fibre Channel with QLogic Host Bus Adapters for the Linux Environment manual for your driver version (manuals are available on www.qlogic.com).

 

These settings are configurable in NVRAM using the Host Adapter Settings, Advanced Settings, and Extended Firmware Settings menus.

 

 

4.   

Modify HBA settings

The EMC default setting for the data rate on the QLA23xx and QLE23xx HBAs is Auto Select mode.  However, it is recommended to hard code the exact rate to the HBA port speed. For example, set the rate to 4 GB on 4GB-capable QLA246x/QLE246x series.

 

Note:  For any device connected to the HBA, set the device data rate (if applicable) before setting the HBA data rate.

 

Perform these steps for each QLogic HBA to be modified:

 

a.     Boot the host and when the QLogic banner appears, press Ctrl-Q enter Fast!Util.

b.    Select Configuration Settings from the Fast!Util Options menu.

c.     Select Host Adapter Settings from the Configuration Settings menu.

d.    Select the Data Rate setting and press Enter.

e.     Select the appropriate speed for the device to which the QLogic HBA will connect.  The options are 0 for 1GBs, 1 for 2GBs, and 2 for Auto Select.

f.     Press ESC to return to the Fast!Util Options menu.

g.    When prompted to save the changes made to the current adapter, select Save Changes.

h.     Repeat the steps above (starting at step b) for each adapter.

i.      Press ESC to exit Fast!Util.

j.      Reboot the host.

 

 

5.   

Modify the HBA topology settings:

The EMC default setting for the topology is set to 2 (loop preferred; otherwise point to point).  For Linux environments, it is recommended that the Connection Options parameter be set to 1 when attached to a fabric and to 0 when attached to an EMC storage array directly.

 

If necessary, the topology value can be changed as follows:

 

a.     Reboot the host, if necessary, and when the QLogic banner appears, press Qtrl-Q to enter Fast!Util.

b.    Select Configuration Settings from the Fast!Util Options menu.

c.     Select Host Adapter Settings from the Configuration Settings menu.

d.    Select Connection Options and press Enter.

e.     Select the appropriate topology and press Enter:

·         For a direct connect environment: Set to 0 (loop only)

·         For a switch environment: Set to 1 (point to point only)

f.     Press ESC to return to the Fast!Util Options menu.

g.    When prompted to save the changes made to the current adapter, select Save Changes.

h.     If you have more adapters to configure, choose Select Host Adapter and repeat the steps above (starting at step b) for each adapter.

i.      Press ESC to exit Fast!Util.

Reboot the host.

 

 

7Table of Contents


 

·        Enabling an in-kernel QLogic driver for Linux

 

Overview

This procedure example for RHEL 4 Linux describes how to grab an appropriate QLogic driver from the in-kernel register and to enable and assign it to the specific HBA. Then build RAM disk including the driver. If driver entry is not available within the Linux kernel, then go to the next set of procedures.

 

IMPORTANT:  Refer to the E-Lab Interoperability Navigator (ESM) for specific qualified kernel versions and distributions.  The support stated in the E-Lab Interoperability Navigator (ESM) supersedes versions listed in this document.

 

This procedure is based on content from the following EMC manuals:

  • EMC Host Connectivity Guide for Linux
  • EMC Support Matrix Linux (RHEL or SuSE)
  • EMC Fibre Channel with QLogic Host Bus Adapters for Linux Environment

 

You can download these guides from EMC Online Support (registration required): https://support.EMC.com

 

 

1.   

Engage the appropriate HBA driver

To engage an HBA driver:

Ensure that the /etc/modprobe.conf file references an entry for each installed QLogic HBA.

 

Fore each installed QLogic HBA, add an entry:

 

alias scsi_hostadapterN qla2xxx

 

Where N is the sequential value of each QLogic HBA installed in the system, beginning with the number after the last host adapter number entry in the file. The first host bus adapter entry begins with zero and qla2xxx is the driver name for the adapter.

 

For example:

 

alias scsi_hostadapter1 qla2100

alias scsi_hostadapter1 qla2200

alias scsi_hostadapter1 qla2300

alias scsi_hostadapter1 qla2322

alias scsi_hostadapter1 qla2400

alias scsi_hostadapter1 qla6312

 

 

2.   

Create RAM disk:

Whenever /etc/modprobe.conf is modified, a new ramdisk should be created to reflect the changes made. Create a new ramdisk image to include the newly added references to the QLogic HBAs:

 

cd /boot

mkinitrd –v initrd-$1.img $1

 

Where: $1 is the v2.6.x kernel version currently running.

For example:

 

mkinitrd –v initrd- 2.6.9-55.ELsmp.img 2.6.9-55.ELsmp

 

 

3.   

Reboot

Reboot the host.

 

 

 

7Table of Contents

 

 

·        Downloading latest QLogic driver for Linux

 

Overview

This procedure describes how to grab the latest QLogic driver software from the QLogic website.

 

IMPORTANT:  Refer to the E-Lab Interoperability Navigator (ESM) for specific qualified kernel versions and distributions.  The support stated in the E-Lab Interoperability Navigator (ESM) supersedes versions listed in this document.

 

Note the following for QLogic v8.x-series drivers:

 

·         Only the QLogic v8.01.xx driver provides support for the 4 GB capable QLA24xx/QLE24xx-series HBAs, not the v8.00.xx driver.

·         The QLogic v8.xx.xx driver is supported on QLA23xx/QLE23xx-series HBAs for RHEL 4.0, RHEL 4.0 U1, and SLES 9 SP1 hosts.

The QLogic v8.01.06 driver introduces support for the 4 GB capable QLA24xx/QLE24xx-series HBAs. The v8.01.06 driver is supported only on RHEL 4.0 U2 and SLES 9 SP2 at this time.

When installing or upgrading drivers, be sure to use the latest versions specified by EMC. Review the E-Lab Interoperability Navigator (ESM) for the latest information on approved HBAs and drivers.

 

 

4.   

Download the HBA driver

To download the latest HBA driver:

a.   Create a directory to which you will save the driver.  For example:

mkdir /qlogic_driver

b.      Use a web browser to access the QLogic website at

http://www.qlogic.com.

 

c.      Click Downloads at the left side of the screen.

d.      Click the EMC link under OEM Models.

e.      On the next page, select EMC SYMMETRIX, CLARiiON, & CELERRA supported software. 

f.       Select the HBA model in the Product column of the table. On the next table, find the supported driver and click the associated Download link. 

g.      Extract the downloaded file to the directory created earlier.

 

 

 

7Table of Contents


 

 

Overview

This procedure describes how to install the QLogic driver on a Linux host. It uses, for illustration purposes, a version 8.0.x driver version for out of kernel (2.6) installation.

 

Refer to the release notes provided with the driver for information that might be unique to new driver revisions.

 

 

1.      

Install prep:

Prior to installing the QLogic driver, processes holding the driver open must be stopped so that the currently loaded driver can be removed from the running kernel.  Perform the following steps prior to the installation:

a.     Stop all I/O.

b.    Unmount all filesystems attached to the QLogic driver.

c.     If PowerPath is installed and enabled on the host, then the PowerPath service must be stopped.  To stop the PowerPath service, type one of the following commands:

 

/etc/init.d/PowerPath stop

or:

service PowerPath stop

d.    If the QLogic SANsurfer daemon qlremote is installed and enabled on the host, stop the qlremote service so the driver can be removed from the currently running kernel.  To stop the qlremote service, type one of the following commands:

 

/etc/init.d/qlremote stop

or:

service qlremote stop

 

e.      

Install the driver

Install the QLogic v8.xx.xx driver onto a Linux host platform using one of the following methods (instructions follow below):

 

·         Method One - To use the QLogic DKMS RPM to compile and install the modular driver for Dell servers running RHEL 4.0 systems and attached to EMC storage systems, refer to METHOD ONE:  Installing the QLogic v8.xx.xx driver via the QLogic DKMS RPM below.

This method requires no manual edits for Dell servers attached to EMC storage systems.  By installing the DKMS RPM, the necessary files will be edited and the driver will be compiled and installed automatically.

·         Method Two – To use the QLogic RPM to compile and install the modular driver for RHEL 4, SLES 9, and SLES 10, refer to METHOD TWO:  Installing the QLogic v8.xx.xx driver via the QLogic Installation Script below.

This method requires no manual edits for systems attached to EMC storage systems.  By installing the QLogic RPM, the necessary files will be edited and the driver will be compiled and installed automatically.

 

 

f.      

Select the directory

Installation instructions for QLogic v8.xx.xx driver:

 

METHOD ONE:  Installing the QLogic v8.xx.xx driver via the QLogic DKMS RPM:

 

This section guides you through the process of installing and utilizing the DKMS RPM for Dell servers running the RHEL 4.0 distribution.  The DKMS RPM will build and install the QLogic driver modules and will modify the /etc/modprobe.conf file with the entries similar to the following:

options qla2xxx ql2xfailover=0
alias scsi_hostadapter2 qla2100
alias scsi_hostadapter3 qla2200
alias scsi_hostadapter4 qla2300
alias scsi_hostadapter5 qla2322
alias scsi_hostadapter6 qla6312


The following are example steps to integrate the QLogic driver.  Also refer to the Readme file in the driver package.

a.     Boot into the qualified and supported kernel onto which the driver will be installed.

b.    Verify you have obtained the qla2xxx-v8.xx.xx1-2dkms.tgz package from the EMC-approved section of the QLogic website as instructed earlier.

c.     Uncompress and extract the source files from the tar archive:

[root@l82bi205 extra]# tar zxvf
qla2xxx-v8.xx.xx-2dkms.tar.gz


The initial decompression will provide you with the following:

qla2xxx-v8.xx.xx-2/
qla2xxx-v8.xx.xx-2/qla2xxx-v8.xx.xx-2dkms.noarch.rpm
qla2xxx-v8.xx.xx-2/dkms-2.0.5-1.noarch.rpm
qla2xxx-v8.xx.xx-2/README.dkms


d.    Install the DKMS RPM:

cd qla2xxx-v8.xx.xx-2
rpm -ivh dkms-2.0.5-1.noarch.rpm

Preparing...
   ########################################### [100%]
   1:dkms
   ########################################### [100%]


e.     Install the QLogic driver RPM:

rpm -ivh qla2xxx-v8.xx.xx-2dkms.noarch.rpm

Preparing...
   ########################################### [100%]
   1:qla2xxx
   ########################################### [100%]


The console output reported by the RPM driver installation will appear.

A new
ramdisk will be created automatically by the DKMS RPM installation.

f.     Reboot the host.

 

 

METHOD TWO:  Installing the QLogic v8.xx.xx driver via the QLogic Installation Script:

 

This section guides you through the process of installing and utilizing the QLogic installation script for RHEL 4.0 and SLES 9 SP1 distributions.  The script will build and install the driver and will modify the /etc/modprobe.conf file on RHEL 4.0 hosts and the /etc/modprobe.conf.local and /etc/sysconfig/kernel files on SLES 8 hosts.

NOTE:  By default, the installation script will install the QLogic HBA SNIA API libraries and the SANsurferCLI.

The following are example steps to install the QLogic driver via the installation script.  Also refer to the Readme file in the driver package.

a.     Boot into the qualified and supported kernel onto which the driver will be installed.

b.    Verify you have obtained the qlafc-linux-8.xx.xx-1-install.tgz package from EMC-approved section of the QLogic website as instructed earlier.

c.     Uncompress and extract the source files from the tar archive:

tar zxvf qlafc-linux-8.xx.xx-1-install.tgz

The initial decompression provides the following:

qlafc-linux-8.xx.xx-1-install/
qlafc-linux-8.xx.xx-1-install/scli-1.06.16-18.i386.rpm
qlafc-linux-8.xx.xx-1-install/set_driver_param
qlafc-linux-8.xx.xx-1-install/qla2xxx-v8.xx.xx-1.noarch.rpm
qlafc-linux-8.xx.xx-1-install/qlinstall
qlafc-linux-8.xx.xx-1-install/README.qlinstall
qlafc-linux-8.xx.xx-1-install/scli-1.06.16-18.ppc64.rpm
qlafc-linux-8.xx.xx-1-install/revision.notes
qlafc-linux-8.xx.xx-1-install/scli-1.06.16-18.ia64.rpm

d.    Install the QLogic driver via the installation script provided.  When using the -i option, the driver will be compiled, the current driver will be unloaded, and the newly compiled driver will be loaded into the kernel.

Note:  EMC does not support the persistent binding implementation contained within the QLogic driver.  As a result, it is recommended that the driver be compiled and installed without persistent binding enabled.  This can be accomplished by using the
dp option with the installation script.

Proceed with the installation.

cd qlafc-linux-8.xx.xx-1-install/./qlinstall -i –dp

The
qlinstall installation script provides the following features:

·         Installs the driver source RPM which installs the driver source code in the following path:

/usr/src/qlogic/<driver_version-rpm_release>

·         Builds and installs the QLogic driver and configuration module (qla2xxx_conf.o) for the QLogic HBA model(s) installed in the system.

·         Creates back-ups of important files and older drivers in the following directory:

/usr/src/qlogic/<driver version-rpm release>/backup

·         Automatically loads the driver for the HBA model present.

·         Builds and installs the ramdisk with the latest driver and configuration modules.

·         Installs the QLogic SNIA API Library V2.

·         Performs device discovery by default.

·         Provides the following command line options:

-    To invoke the device discovery at a later stage

-    To rebuild the ramdisk excluding or including QLogic HBA driver

-    To unload/load driver

-    To update the option ROM (BIOS, and so forth) on all HBAs

-    To update the NVRAM on all HBAs

-    To pass driver parameters.  For example, ql_port_down_retry (Overrides NVRAM default)

-    To invoke the SANsurferCLI (SCLI)
NOTE:  The SANsurferCLI (SCLI) is a command line interface that provides more flexibility for performing HBA configuration, device discovery, Option ROM and NVRAM updates, etc.

 

 

 

 

7Table of Contents

 

 

 

·        Determine WWN of new HBA for Linux

 

Overview

This procedure describes how to determine the WWN of the newly installed HBA.

 

 

1.   

Determine HBA WWN:

To determine the WWN of the newly installed HBA, at the command line use one of the following forms:

 

cat /proc/scsi/qla2300/# (instance number could be 0, 1, 2, 3,  …)

Or for Emulex:

cat /proc/scsi/lpfc/# (instance number could be 0, 1, 2, 3,  …)

This can also be verified by looking at the Fibre Channel Switch Port that the HBA is plugged in to.

 

 

 

7Table of Contents

 

·        Creating switch zones                                      

 

Overview

This procedure describes how to create switch zones using Cisco 9000 series switches.

For information on configuring a zone, refer to the following:

 

http://www.cisco.com/en/US/docs/storage/san_switches/mds9000/sw/rel_1_x/1_3/fm/configuration/guide/ZoneCnfg.pdf

 

For the entire Configuration guide, refer to:

http://www.cisco.com/en/US/docs/storage/san_switches/mds9000/sw/rel_1_x/1_3/fm/configuration/guide/FMCfgGuide.pdf

 

 

3.  

Create switch zones

Switch zones define paths between the host HBAs and the Symmetrix array and can be created using the switch management software (refer to the switch documentation).

 

Note:  If you are using an HP-UX server with QuickLoop zoning, then you must use port zoning.  Refer to EMC Knowledgebase article emc66355 for more information.

 

Definition -

Zoning is the process by which a single initiator is matched to a single storage array port (FA/SP) to be able to pass info to storage array device.

 

A Zone consists of 2 or more members, with only 1 initiator (HBA) per zone

A Zone is part of a zoneset .

A Zoneset is a collection of zones in a single fabric.

A Zoneset requires activation in order to be updated.

 

Important:  EMC Recommends single initiator zoning. To reduce the chance of data unavailability in the event that an HBA, switch, or FA Failure,  the host should have at least one primary path and one alternate path to each FA

 

 


 

4.   

Determine zoning configuration

Follow these steps to determine your zoning configuration:

a.      Determine the FA WWN by entering the following command:

symcfg –sid XXXX list –fa all

b.      Determine the HBA WWN by referencing the OS-specific commands from provisioning a host section.

c.      Determine the number of Paths needed.

d.      Determine the Active Zone set name.

The operations above must be performed on both fabrics for all FAs and HBAs.

 

5.   

Verify the HBA is logged in

Solutions Enabler offers a command line utility that will verify what is logged in to a specific Symmetrix FA. Use the following command:

 

Symmaskdb –sid SymmID –dir 3c –p 0 list database

 

Where SymmID is the last four numbers of the Symmetrix serial number, 3c is the FA and 0 is the port on the FA.

 

After verifying that the host is logged in, move on to configuring the host.

 

 

7Table of Contents

 

 

 

 

·        Adding Symmetrix Devices

 

Overview

This procedure explains how to add or create Symmetrix devices from the array’s free physical disk space.

 

This procedure is based on content from the following EMC manual:

  • EMC Solutions Enabler Symmetrix Array Controls CLI Product Guide

 

You can download these guides from EMC Online Support (registration required): https://support.EMC.com.

 

 

1.

Look for free space

To check for available free space before you begin, enter the following command:

 

   symconfigure –sid SymmID list –freespace -units CYLINDERS | MB

 

where:

 

SymmID = The ID of a specific Symmetrix the symconfigure command is to be executed against.

 

For example, to see how much free space is available on a Symmetrix with an ID of  000000012345, execute one of the following commands:

 

symconfigure –sid 12345 list –freespace -units CYLINDERS

 

or

 

symconfigure –sid 12345 list –freespace -units MB

 

2.

Build a command file that will be used to add the device

The create dev command is used to add or create a Symmetrix device. The syntax for this command is:

 

create dev count=n, size=Cylinders,

  emulation=EmulationType,

  config=DeviceConfig,

  [, data_member_count=nn]

  [, remote_config=DeviceConfig, remote_data_member_count=nn, ra_group=n,

    [dynamic_capability=[dyn_rdf | dyn_rdf1_only |dyn_rdf2_only], ]

  [, attribute=ckd_meta | savedev

  [in pool PoolName] [member_state=ENABLE | DISABLE], ]

  [, disk_group_num=nnn, remote_disk_group _num=nnn];

 

where:

count = The number of devices to create.

 

size = The size of the device needed in number of cylinders. In Symmetrix DMX models, a cylinder of an FBA device contains 1920 512-byte blocks, or 0.98 megabytes. In earlier Symmetrix models, a cylinder of an FBA device contains 960 512-byte blocks or, 0.468 megabytes. To calculate the number of cylinders, use one of the following calculations:

 

Blocks ÷ 960

or

(Device size in megabytes) x 10242 ÷ (960 x 512 bytes)

or

(Size in megabytes) x 2.1333

Note: Devices to be used as BCV devices, SRDF devices, or metamembers will need to precisely match corresponding device sizes. Use symdev/sympd show or symdev/sympd list –cyl commands to see relevant device sizes.

 

For additional information about sizing devices in cylinders, refer to Chapter 1 of the Solutions Enabler Symmetrix Array Management CLI Product Guide.

 

emulation = The device emulation type, which currently must be fixed block architecture (FBA), CELERRA FBA, VME512 FBA, or CKD-3380, or CKD-3390 for z/OS environments.

 

config = The desired device configuration type. For possible values, see the Solutions Enabler Symmetrix Array Control CLI Product Guide.

 

data_member_count = The number of data members when creating RAID 5 or RAID 6 devices on a Symmetrix array with Enginuity version 5772 and higher.

The user should set the value to 3 or 7 for RAID 5 (3+1) and RAID 5 (7+1), or  6 or 14 for RAID 6 (6+2) and RAID 6 (14+2).

 

remote_config = The desired remote SRDF configuration (if any). Managing Configuration Changes

remote_data_member_count = The number of remote data members when creating RDF RAID 5 or RAID 6 devices on a Symmetrix array with Enginuity version 5772 and higher. Set the value to 3 or 7 for RAID 5 (3+1) and RAID 5 (7+1) or, 6 or 14 for RAID 6 (6+2) and RAID 6 (14+2).

 

ra_group = The RA group number in the SRDF environment.

 

dynamic_capability = The type of dynamic SRDF device to create. Possible values are:

 

dyn_rdf — Creates a dynamic SRDF device.

dyn_rdf1_only — Creates a dynamic R1 SRDF device.

dyn_rdf2_only — Creates a dynamic R2 SRDF device.

 

The specified dynamic capability will be applied to the local device and the corresponding remote device will be assigned a complementary dynamic capability, according to the following:

 

  Local device     Remote device

  dyn_rdf          dyn_rdf

  dyn_rdf1_only    dyn_rdf2_only

  dyn_rdf2_only    dyn_rdf1_only

 

attribute = The desired device attribute. Possible values are:

 

ckd_meta — When creating a device with an emulation type of CKD-3380 or CKD-3390, this indicates that the device should be a striped metadevice. CKD metadevices must be created in sets of four devices.

 

savedev — When creating a device, this indicates that the device should be a SAVE device. The device will become part of a pool of devices that are used with TimeFinder/Snap for virtual device Snap operations.

 

PoolName = The name of the SAVE device pool. It can be from 1 to 12 alphanumeric characters long and include hyphens ( - ), and underscore ( _ ) characters. The name DEFAULT_POOL is reserved to represent the container of all unpooled devices.

 

member_state = States whether the SAVE device(s) being added should be enabled or disabled in the pool.

Managing Configuration Changes

disk_group_num = When creating a device, this option allows you to specify a disk group. A disk group is a set of physical disks set aside to be used to create devices of the protection level. Disk group usage may improve the performance of some configurations.

 

To see how disk groups are organized in your configuration, use the following commands:

 

symdisk -sid SymmID list -by_diskgroup

 

and

 

symdev -sid SymmID list -diskgroup(nn)

 

remote_disk_group_num = When creating a device, this option allows you to specify a remote disk group.

 

Using a text editor, such as Notepad or the UNIX vi Editor, create a command file that contains one or more create dev commands. (For this procedure, a file named adddevices.cmd is used.)

 

For example, to add four new Symmetrix devices to Symmetrix array 000000012345 as two-way mirrored devices with a size of 1100 cylinders (516 MB) using FBA emulation, and to create the hypers to support the new devices in disk group 1, create a command file that contains the following line:

 

create dev count=4, size=1100, emulation=FBA, config=2-Way-Mir disk_group=1;

 

Note: If the Symmetrix array being used contains an EA, CA, or EF adapter, a value for the mvs_ssid parameter must be provided so any new FBA devices created are not seen as part of an existing subsystem ID group. If that is the case, a command line that looks more like this should be used instead:

 

create dev count=4, size=1100, emulation=FBA, mvs_ssid=1, config=2-Way-Mir disk_group=1;

 

 

3.

Commit the command file used to create the device(s)

Commit the command file created in the previous step (and activate the device creation command) by executing a symconfigure command that looks something like this:

 

symconfigure -sid SymmID -file CmdFile -v -noprompt commit

 

where:

 

SymmID = The ID of a specific Symmetrix the symconfigure command is to be executed against.

 

CmdFile = A file containing one or more create dev commands.

 

For example, to commit the commands in a command file named adddevices.cmd against a Symmetrix with an ID of 000000012345, use the following command:

 

symconfigure -sid 12345 -file adddevices.cmd -v -noprompt commit

 

When this command is executed, you should see output that looks something like this:

 

A Configuration Change operation is in progress. Please wait...

 

    Establishing a configuration change session...............Established.

 

 

    Processing symmetrix 000000012345

    {

      create dev count=4, size=1100 cyl, emulation=FBA, config=2-Way Mir,

        mvs_ssid=0, disk_group=1;

    }

 

    Performing Access checks..................................Allowed.

    Checking Device Reservations..............................Allowed.

    Submitting configuration changes..........................Submitted

    Validating configuration changes..........................Validated.

 

      New symdevs: 0918:091B

    Initiating PREPARE of configuration changes...............Prepared.

    Initiating COMMIT of configuration changes................Queued.

    COMMIT requesting required resources......................Obtained.

    Step 002 of 078 steps.....................................Executing.

    Step 004 of 078 steps.....................................Executing.

    Step 004 of 078 steps.....................................Executing.

 

...

 

    Step 111 of 151 steps.....................................Executing.

    Step 114 of 151 steps.....................................Executing.

    Step 151 of 151 steps.....................................Executing.

    Local:  COMMIT............................................Done.

    Terminating the configuration change session..............Done.

 

The configuration change session has successfully completed.

 

It is important to note that when creating devices there is no impact on I/O activity.

 

 

4.

Verify that the desired devices were created

To verify that the desired devices were created, enter the following command:

 

   symdev –sid SymmID list

 

where:

 

SymmID = The ID of a specific Symmetrix the symdev command is to be executed against.

 

For example, to see if new devices were successfully created on a Symmetrix with an ID of 000000012345, execute the following command:

 

symdev –sid 12345 list

 

The devices that were just created should be displayed at the bottom of the list produced.

 

 

 

7Table of Contents

 

 

 

·        Mapping Devices within Symmetrix

 

Overview

This procedure describes how to map devices within the Symmetrix array. You can map devices to front-end director ports, or a range of devices to consecutive addresses by specifying a starting address.

 

This procedure is based on contents from the following EMC manual:

  • EMC Solutions Enabler Symmetrix Array Controls CLI Product Guide

 

You can download the guide from EMC Online Support (registration required): https://support.EMC.com

 

 

6.  

Identify Symmetrix devices that have not been mapped

Identify Symmetrix devices that have not been mapped to any front-end adapter ports by executing a symdev command that looks like this:

 

symdev –sid SymmID list -noport

 

where:

 

SymmID = The ID of a specific Symmetrix the symdev command is to be executed against.

 

For example, to obtain a list of Symmetrix devices that have not been mapped to a front-end adapter on a Symmetrix with an ID of 000000012345, execute the following command:

 

symdev –sid 12345 list -noport

 

Examine the output produced and identify the Symmetrix devices you want mapped.

 

 

7.  

Identify front-end adapters that are available

Identify the front-end adaptors that are available by executing a symcfg command that looks like this:

 

symcfg –sid SymmID list –dir all

 

where:

 

SymmID = The ID of a specific Symmetrix the symcfg command is to be executed against.

 

For example, to obtain a list of front-end adapters available on a Symmetrix with an ID of  000000012345, execute the following command:

 

symcfg -sid 12345 list -dir all

 

When this command is executed, you should see output that looks something like this:

 

Symmetrix ID: 000000012345

 

           S Y M M E T R I X    D I R E C T O R S

 

    Ident  Symbolic  Numeric  Slot  Type          Status

 

    DF-1A     01A       1       1   DISK          Online

    DF-2A     02A       2       2   DISK          Online

    DF-15A    15A      15      15   DISK          Online

    DF-16A    16A      16      16   DISK          Online

    DF-1B     01B      17       1   DISK          Online

    DF-2B     02B      18       2   DISK          Online

    DF-15B    15B      31      15   DISK          Online

    DF-16B    16B      32      16   DISK          Online

    FA-1C     01C      33       1   FibreChannel  Online

    FA-2C     02C      34       2   FibreChannel  Online

    FA-15C    15C      47      15   FibreChannel  Online

    FA-16C    16C      48      16   FibreChannel  Online

    RF-1D     01D      49       1   RDF-BI-DIR    Online

    RF-2D     02D      50       2   RDF-R1        Online

    RF-15D    15D      63      15   RDF-R1        Online

    RF-16D    16D      64      16   RDF-BI-DIR    Online

 

Examine the output produced and identify the front-end adapter to which you want  Symmetrix devices mapped.

 

 

8.  

Examine the front-end adaptor and port for available LUN Numbers

Identify the LUN numbers that are available for a specific front-end adapter and port by executing a symcfg command that looks like this:

 

symcfg –sid SymmID list –dir Adapter –p Port –address -available

 

where:

 

SymmID = The ID of a specific Symmetrix the symcfg command is to be executed against.

 

Adapter = The ID of the front-end adaptor that device information is to be collected for.

 

Port = The port number that device information is to be collected for.

 

For example, to obtain device information for adapter 16C, port 0 on a Symmetrix with an ID of 000000012345, execute the following command:

 

symcfg -sid 12345 list -dir 16c -p 0 -address -available

 

When this command is executed, you should see output that looks something like this:

 

Symmetrix ID: 000000012345

 

           Director                 Device Name    Attr     Address

    ---------------------- ----------------------- ---- --------------

    Ident   Symbolic  Port Sym   Physical               VBUS  TID  LUN

    ------  --------  ---- ----  -----------------      ----  ---  ---

    FA-16C  16C       0    0100  Not Visible               0   00  000

                           0101  Not Visible               0   00  001

                           0102  Not Visible               0   00  002

                           0103  Not Visible               0   00  003

                           0104  Not Visible               0   00  004

                           0105  Not Visible               0   00  005

...

                           08DC  Not Visible        (M)    0   00  0CC

                           -     AVAILABLE                 0   00  0CD *

                           0520  Not Visible        (M)    0   00  100

...

                           07CD  Not Visible               0   00  143

                           -     AVAILABLE                 0   00  144 *

                           07D1  Not Visible               0   00  200

...

                           0850  Not Visible               0   00  27F

                           -     AVAILABLE                 0   00  280 *

                           068C  Not Visible               0   00  68C

...

                           0697  Not Visible               0   00  697

                           -     AVAILABLE                 0   00  698 *

    Total                  ----

    Mapped Devices:         413

    Including Metamembers:  638

    Available Addresses:   3281 (s)

 

 

Legend for Available address:

 

(*): The VBUS, TID, LUN address values represent a gap in the

     address assignments or are the next available address in

     the run

(s): The Available Addresses for a director are shared among

     its ports (shared)

 

Note: An asterisk (*) in the LUN column represents an available LUN number or the next available range of numbers.

 

 

9.  

Build a command file that will be used to map devices in the Symmetrix array

The map dev command is used to map a Symmetrix device to a front-end director port. The syntax for this command is:

 

map dev SymDevName [:SymDevName] to dir DirectorNum:PortNum

[starting][target=ScsiTarget,] lun=ScsiLun

[, vbus=FibreVbus]

[, device_number=ckd_device_number]

[, awwn=awwn|wwn=wwn|iscsi=iscsi];

[, masking host_lun=lun | dynamic_lun];

 

where:

target = The SCSI target ID (hex value).

 

lun = The SCSI logical unit number (hex value).

 

dynamic_lun = The dynamic LUN addressing scheme. LUN addressing is assigned based on what may already be in use for that host HBA.

 

vbus = The virtual bus address for mapping to an FA port if using volume set addressing.

 

device_number = The CKD device number, when mapping a CKD device to a z/OS host.

 

awwn = The user-given name or alias WWN of a host HBA port, if updating a VCM database.

 

wwn = The unique 64-bit World Wide Name (WWN) identifier for an HBA port, if updating a VCM database.

 

iscsi = The iSCSI name, if updating a VCM database.

 

Note: For HP-UX operating systems, you must specify a target number, a virtual bus number, and LUN number. For Sun, AIX, and Windows, the target number should be 0 in a switched fabric environment.

 

Using a text editor, such as Notepad or the UNIX vi Editor, create a command file that contains one or more map dev commands. (For this procedure, a file named mapdevices.cmd is used.)

 

For example, to map Symmetrix device 000C to port 0 of front-end adapter 16C and assign it the LUN number 0CD, create a command file that contains the following lines:

 

map dev 000C to dir 16C:0 target=0, lun=0CD;

 

10.  

Commit the command file used to map devices

Commit the command file created in the previous step (and activate the map devices command specified) by executing a symconfigure command that looks something like this:

 

symconfigure -sid SymmID -file CmdFile -v -noprompt commit

 

where:

 

SymmID = The ID of a specific Symmetrix the symconfigure command is to be executed against.

 

CmdFile = A file containing one or more map dev commands.

 

For example, to commit the commands in a command file named mapdevices.cmd against a Symmetrix with an ID of 000000012345, use the following command:

 

symconfigure -sid 12345 –file mapdevices.cmd –v –noprompt commit

 

When this command is executed, you should see output that looks something like this:

 

A Configuration Change operation is in progress. Please wait...

 

    Establishing a configuration change session...............Established.

    Processing symmetrix 000000012345

    {

      map dev 000C to dir 16C:0 lun=0CD;

    }

 

    Performing Access checks..................................Allowed.

    Checking Device Reservations..............................Allowed.

    Submitting configuration changes..........................Submitted

    Locking devices...........................................Locked.

    Validating configuration changes..........................Validated.

    Initiating PREPARE of configuration changes...............Prepared.

    Initiating COMMIT of configuration changes................Queued.

    COMMIT requesting required resources......................Obtained.

    Step 003 of 012 steps.....................................Executing.

    Step 003 of 012 steps.....................................Executing.

    Step 009 of 012 steps.....................................Executing.

    Local:  COMMIT............................................Done.

    Terminating the configuration change session..............Done.

 

The configuration change session has successfully completed.

 

 

11.  

Rediscover the SYMAPI database

After performing the proper host procedures to update the host view, you need to complete host addressing by making sure that the host address is recognized in the SYMAPI view. To update the SYMAPI database, execute the following command:

 

symcfg discover

 

This may takes some time; please wait for completion.

 

 

12.  

Verify that the device mapping was successful

To verify that the device specified was successfully mapped to the front-end adapter, port, and LUN specified, execute the following command:

 

symdev –sid SymmID show SymDevName

 

where:

 

SymmID = The ID of a specific Symmetrix the symdev command is to be executed against.

 

SymDevName = The specific Symmetrix device that detailed information is to be displayed for.

 

For example, to display detailed information about Symmetrix device 000C on a Symmetrix with an ID of 000000012345, execute the following command:

 

symdev –sid 12345 show 000C

 

Scan the output produced and examine the section titled “Front Director Paths”. If the Symmetrix device was successfully mapped, you should see something that looks like this:

 

    Front Director Paths (1):

        {

        ------------------------------------------------------------------

                             POWERPATH  DIRECTOR   PORT             LUN

                             --------- ----------  ---- -------- ---------

        PdevName             Type      Type Num    Sts  VBUS TID SYMM Host

        ------------------------------------------------------------------

        Not Visible          N/A       FA   16C:0  RW   000  00  0CD  N/A

        }

 

 

 

 

 

·        Masking Devices

 

Overview

Masking is an operation that is used to present devices to a specific host bus adapter (HBA) and is typically utilized when control is needed over which devices are made available to hosts. In a Symmetrix environment, masking defines a relationship between one or more Symmetrix devices, front-end director ports, and HBAs that reside at a host.

 

To perform masking, you will need access to a host that has the Solutions Enabler base installed along with the device masking portion licensed. This host will need to have a fiber path to the Symmetrix and be able to see the VCM device and a gatekeeper.

 

This procedure is based on contents from the following EMC manual:

  • EMC Solutions Enabler Symmetrix Array Controls CLI Product Guide

 

You can download these guides from EMC Online Support (registration required): https://support.EMC.com

 

 

 

 

 

13.  

Verify that the device(s) to be masked have been mapped

To verify that the device(s) to be masked have been mapped to a front-end director, port, and LUN, execute the following command:

 

symdev –sid SymmID show SymDevName

 

where:

 

SymmID = The ID of a specific Symmetrix the symdev command is to be executed against.

 

SymDevName = The specific Symmetrix device that detailed information is to be displayed for.

 

For example, to display detailed information about Symmetrix device 000C on a Symmetrix with an ID of 000000012345, execute the following command:

 

symdev –sid 12345 show 000C

 

Scan the output produced and examine the section titled “Front Director Paths”. If the Symmetrix device has been mapped, you should see something that looks like this:

 

Front Director Paths (1):

    {

    ----------------------------------------------------------------------

                             POWERPATH  DIRECTOR   PORT             LUN

                             --------- ----------  ---- -------- ---------

    PdevName                 Type      Type Num    Sts  VBUS TID SYMM Host

    ----------------------------------------------------------------------

    \\.\PHYSICALDRIVE7       N/A       FA   16C:0  RW   000  00  016  000

    }

 

Repeat this step for every device to be masked. If a device to be masked has not been mapped, you must map it to a front-end director, port, and LUN before continuing with this procedure.

 

 

14.  

Get the World Wide Name (WWN) for the Host Bus Adapter (HBA) the device(s) are to be masked to

There are several tools that can be used to obtain the WWN of installed HBAs. For example, Emulex provides a utility called HBAnywhere that can be used to obtain this information. If Solutions Enabler has been installed on a host, the syminq hba command can be used as well.

 

 

Using either the appropriate tools on the host or the configuration documentation where this information was recorded earlier, obtain the WWN for the HBA that one or more devices are to be masked to.

 

 

15.  

Mask the appropriate devices in the Symmetrix array

The symmask command is used to mask one or more devices to a specified HBA/director-port channel. The syntax for this command is:

 

symmask -sid SymmID -wwn wwn|-awwn awwn|-iscsi iscsi |-aiscsi aiscsi

add devs StartSymDev:EndSymDev|SymDevName|SymDevName,,,...

-dir DirNumber -p PortNumber [-noprompt]

 

where:

 

SymmID = The ID of a specific Symmetrix the symmask command is to be executed against.

 

wwn= A valid World Wide Name (WWN).

 

awwn = A user-given name in an ASCII WWN format.

 

iscsi = A valid iSCSI name.

 

aiscsi = A user-given name or an alias iSCSI name.

 

StartSymDev = The first Symmetrix device name in a range.

 

EndSymDev = The last Symmetrix device name in a range.

 

SymDevName = One or more specific Symmetrix device names.

 

DirNumber = A specific director number designation.

 

PortNumber = A specific port number designation.

 

Note: When masking metadevices, only the Symmetrix device name that is designated as the metadevice head should be specified (via the SymDevName parameter).

 

For example, to mask Symmetrix devices 000C and 000D, (which have already been mapped to port 0 of front-end adapter 16C on a Symmetrix with an ID of 000000012345), so that they can only be accessed by an HBA that has the World Wide Name 10:00:00:00:c9:6c:c8:12, execute the following command:

 

symmask –sid 12345 –wwn 10000000c96cc812 add devs 000C:000D -dir 16C -p 0

 

If Symmetrix devices 000C and 000D were masked to a different HBA earlier, the following message will be displayed when this command is executed:

 

The following devices are already assigned in at least one entry:

 

000C 000D

 

Would you like to continue (y/[n])?

 

If you indicate that the mask operation is to continue, the devices will be masked for multiple HBAs. (If the -noprompt option is specified, the masking operation will execute automatically without requiring user intervention.)

 

 

16.  

Verify that the device masking was successful

To verify that the device(s) specified were successfully masked, execute the following command:

 

symmaskdb –sid SymmID list database –dir DirNumber –p PortNumber

 

where:

 

SymmID = The ID of a specific Symmetrix the symmask command is to be executed against.

 

DirNumber = The specific director number the Symmetrix device was masked for.

 

PortNumber = The specific port number the Symmetrix device was masked for.

 

For example, to display a list of devices that have been masked for port 0 of front-end adapter 16C on a Symmetrix with an ID of  000000012345, execute the following command:

 

symmaskdb -sid 12345 list database –dir 16C –p 0

 

When this command is executed, you should see output that looks something like this:

 

Symmetrix ID            : 000000012345

 

Database Type           : Type6

Last updated at         : 03:55:26 PM on Tue Mar 31,2009

 

Director Identification : FA-16C

Director Port           : 0

 

                               User-generated

Identifier        Type   Node Name        Port Name         Devices

----------------  -----  ---------------------------------  ---------

10000000c96cc812  Fibre  10000000c96cc812 10000000c96cc812  000C:000D

 

 

17.  

Back up the VCMDB

The symmaskdb backup command is used to create a backup copy of the Device Masking VCMDB. The syntax for this command is:

 

symmaskdb -sid SymmID backup –file BkupFileName -noprompt

 

where:

 

SymmID = The ID of a specific Symmetrix the symmaskdb command is to be executed against.

 

BkupFileName = The name of the file to be used to store the backup copy of the Device Masking VCMDB created. (You cannot reuse backup file names; a new file name must be provided each time this command is executed.) The backup copy will be stored in the file specified and that file will be written to the current directory (i.e., the directory that this command is executed from).

 

For example, to store a backup copy of an existing Device Masking VCMDB for a Symmetrix with an ID of 000000012345 in a file named BackupDevMaskDB_1, execute the following command:

 

symmaskdb –sid 12345 backup –file BackupDevMaskDB_1 –noprompt

 

When this command is executed, you should see a message that looks something like this:

 

Symmetrix SymMask database backed up to file BackupDevMaskDB_1

from Symmetrix 000000012345

 

 

18.  

Refresh the Symmetrix director with VCMDB data

Update the Symmetrix array with the configuration changes by performing a director refresh operation. This causes the Symmetrix director to refresh its WWN/iSCSI-related profile tables in cache with the content data from the device masking VCMDB.

 

The symmaskdb refresh command is used to force a Symmetrix director to refresh its profile tables with the contents of the Device Masking VCMDB. The syntax for this command is:

 

symmask -sid SymmID refresh -noprompt

 

where:

 

SymmID = The ID of a specific Symmetrix the symmaskdb command is to be executed against.

 

For example, to force a Symmetrix director to refresh its profile tables with the contents of the Device Masking VCMDB for a Symmetrix with an ID of  000000012345, execute the following command:

 

symmask –sid 12345 refresh –noprompt

 

When this command is executed, you should see a message that looks something like this:

 

Symmetrix FA directors updated with contents of SymMask Database 000000012345

 

! Caution: Before executing this command, make sure no HBAs are accessing devices in the masked channel (in other words, make sure that no application or user activity is taking place on the HBAs that are about to be refreshed). !

 

 

19.  

Run the appropriate configuration method on all affected hosts or reboot all affected hosts

Many host operating systems provide a utility that can be used to automatically configure devices that have been made available to a system. For example, in Windows you can use the Device Manager to scan for hardware changes; with AIX you can execute the cfgmgr command. Before masked devices will be made visible to a host, this utility must be run on all hosts that have had devices added or removed (via masking changes). If no such utility exists, the affected hosts should be rebooted.

 

Refer to the online help available with each host operating system for information on how to perform online configuration changes or a reboot operation.

 

 

20.  

Rebuild the Symmetrix configuration database on all appropriate hosts/management servers

Any time you reconfigure a Symmetrix array by adding or removing device masking information, the SYMAPI configuration database should be updated on all affected hosts/management servers that are running Solutions Enabler. This is done by executing the following command:

 

symcfg discover

 

When executed, this command interrogates all SCSI devices available; therefore, it can take a significant amount of time to complete.

 

 

 

 

 

 

·        Creating a Device Group

 

Overview

A collection of devices can be assigned to a named group to provide a more manageable single object to query status and impart blanket control operations. Device groups and composite group definitions are created locally through an attached host.

 

This procedure is based on content from the following EMC manual:

  • EMC Solutions Enabler Symmetrix Array Management CLI Product Guide

 

You can download the guide from EMC Online Support (registration required): https://support.EMC.com.

 

 

 

21.  

Examine device groups you already have

Since device group names must be unique, before you attempt to create a new one you should find out what device groups have already been created. To view a list of device groups that have already been created, execute the following command:

 

symdg list

 

 

22.  

Create a new device group

The symdg create command is used to create a new device group. The syntax for this command is:

 

symdg create DgName [-type REGULAR | RDF1 | RDF2 | RDF21]

 

where:

 

DgName = The name to assign to the device group.

 

For example, to create a device group, assign it the name TFgroup, and specify that it is a REGULAR device group, execute a symdg create command that looks like this:

 

symdg create TFgroup –type REGULAR

 

 

23.  

Add Symmetrix devices to the group

The symld command is used to add devices to an existing device group. Individual devices can be added by specifying either the physical device name (add pd) or the Symmetrix device name (add dev), as follows:

 

symld -g TFgroup add pd /dev/rdsk/c4t0d0s2
symld -g TFgroup -sid 12345 add dev 000C

A contiguous set of devices can be added to a device group by executing a symld command that looks more like this:

 

   symld -g TFgroup –sid 12345 addall –RANGE 000C:000E

(In this example, Symmetrix devices on Symmetrix array 000000012345 were used.)

 

On the other hand, BCV devices can be added to a device group by executing a symbcv command that looks like this:

 

symbcv -g TFgroup -sid 12345 associate dev 001C

 

Keep in mind that all devices added to a device group must be part of the same Symmetrix array. (Composite groups are used to group devices that reside on different Symmetrix arrays.)

 

24.  

Verify the device group creation

To verify that the device group was successfully created and that the appropriate devices were added to it, execute the following command:

 

symdg show DgName

 

where:

 

DgName = The name that has been assigned to the device group.

 

For example, to verify that a device group named TFgroup was created and the appropriate devices were assigned to it, execute the following command:

 

symdg show TFgroup

 

When this command is executed, you should see output that looks something like this:

 

Group Name:  TFgroup

 

    Group Type                                   : REGULAR

    Device Group in GNS                          : No

    Valid                                        : Yes

    Symmetrix ID                                 : 000000012345

    Group Creation Time                          : Mon Mar 23 08:50:17 2009

    Vendor ID                                    : EMC Corp

    Application ID                               : SYMCLI

 

    Number of STD Devices in Group               :    3

    Number of Associated GK's                    :    0

    Number of Locally-associated BCV's           :    0

    Number of Locally-associated VDEV's          :    0

    Number of Locally-associated TGT's           :    0

    Number of Remotely-associated VDEV's(STD RDF):    0

    Number of Remotely-associated BCV's (STD RDF):    0

    Number of Remotely-associated TGT's(TGT RDF) :    0

    Number of Remotely-associated BCV's (BCV RDF):    0

    Number of Remotely-assoc'd RBCV's (RBCV RDF) :    0

    Number of Remotely-assoc'd BCV's (Hop-2 BCV) :    0

    Number of Remotely-assoc'd VDEV's(Hop-2 VDEV):    0

    Number of Remotely-assoc'd TGT's (Hop-2 TGT) :    0

 

    Standard (STD) Devices (3):

        {

        --------------------------------------------------------------------

                                                      Sym               Cap

        LdevName              PdevName                Dev  Att. Sts     (MB)

        --------------------------------------------------------------------

        DEV001                N/A                     000C      RW      1031

        DEV002                N/A                     000D      RW      1031

        DEV003                N/A                     000E      RW      1031

        }

 

 

 

 

·        Making LUNs or Devices available to Linux

 

Overview

This procedure describes how to make the devices or LUNs available or visible to a Linux host. It describes how to use the fdisk command to label and partition devices.

 

The Symmetrix devices will be represented in /proc/scsi/scsi and can be viewed by using:

 

more /proc/scsi/scsi.

 

Additional sd and sg filenames may need to be created because the number created by default in the kernel is usually less than 128 device filenames. 

 

Note and Caution: If your host system cannot see the devices, it may necessary to do a disk or device scan of the I/O bus with this procedure. If you must, at an appropriate low-peak time in your production environment, apply the necessary operating system native commands, to perform a device scan (discover) to make the devices visible to the host workstation. This might take a few minutes or so to execute. Be sure you follow your operating system specific administrators guide and/or known site procedures for this operation, as scanning devices can cause pauses and crashes to other operational processes running on this host. If you can and prefer, a reboot will also accomplish the same thing (device discovery). The following steps are provided only as a general or possible guide and not meant to supersede your site preferences.

 

 

25.  

Partition devices:

The fdisk command allows a single disk device to be partitioned. To partition the Symmetrix devices, verify that you are logged in as root and perform the steps below for each new device that requires partitioning as follows:

 

a.      Enter the following command:

fdisk /dev/<device_name>

where <
device_name> is the LUN device-name

For example:  fdisk /dev/sdb

b.      Type m (for help) and press Enter to display a menu of options.

Type
x (for expert functionality) and press Enter.  Type m and press Enter to see another menu of options.

c.      From the main fdisk menu, type n and press Enter to add a new disk partition.

d.      Type t and press Enter to change a partition’s system ID.

e.      Type L and press Enter for a list of hex codes for system Ids.  (Linux is type 83; Linux swap is 82.)

f.       Type p and press Enter to print the partition table so you can verify that the selections look correct.

g.      Type w and press Enter to write the table to disk and exit fdisk.

h.      Repeat steps c through g for each new Symmetrix device.

i.        If you want to display the partition numbers created after Linux recognizes them, view the file:

more / proc/partitions

 

 

26.  

Create filesystems on partitions:

Once each Symmetrix device has been formatted, partitions, and labeled, you must create a filesystem in order to make the partition useable.

 

To create a new filesystem:

 

a.   Enter the following mkfs command:

mkfs /dev/<device_name>/<partition_number>

where <device_name> is the LUN device name and <partition_number> is the partition number

Examples:

mkfs /dev/sdb1
mkfs –j/dev/sbd1
mkreiserfs –h r5 /dev/sdb1

b.      When prompted to confirm the construction of a new filesystem, type Y and press Enter.

 

 

27.  

Create mount directory:

Once the filesystem for each Symmetrix disk device has been created, you must create a mount directory for each device as follows:

 

mkdir /<mount_point>

For example:  

 

mkdir /mnt/sdb

 

where /mnt/sdb is the complete path for the new filesystem

 

 

28.  

Mount the file systems:

To mount each filesystem, enter the mount command as follows:

 

mount -t <filesystem type> /dev/<device_partition_name> <mount_point>

 

Examples:

 

mount –t ext2 /dev/sdb1 /mnt/sdb

mount –t ext3 /dev/sdb1 /mnt/sdb

mount –t reiserfs /dev/sdb1 /mnt/sdb

 

This mounts the partition /dev/sdb1 onto the mount point /mnt/sdb as the filesystem type specified.

 

This mount command is required every time the host is rebooted.  If preferable, the filesystem can be mounted automatically at boot time by adding this information to the /etc/fstab file.  Once the filesystems are mounted, data can be written to and read from them.

 

 

 

7Table of Contents

 

 

 

 

·        Installing PowerPath software on a Linux host

 

Overview

This procedure describes how to install PowerPath on a Linux host.

 

Note: This procedure is specific to the installation of PowerPath version 4.5.x, which is the current shipping version for Linux hosts.  For instructions on installing other versions of PowerPath, refer to the applicable PowerPath installation manual available on PowerLink.

 

This procedure is based on content from the following EMC manual:

  • PowerPath for Linux Installation and Administration Guide

 

You can download the guide from EMC Online Support (registration required): https://support.EMC.com

 

Verify the following prior to the installation of PowerPath:

 

·         Review the patch ReadMe files to determine which patches (if any) you want to install after PowerPath, and whether those patches have any added prerequisites that must be met before you install PowerPath.

·         Determine if the PowerPath software you are installing requires the removal or presence of a previous version of PowerPath.  Some full versions require the previous version to be removed while others do not.  Also, some patches require the full version to be present while others require it to be removed.  Refer to the PowerPath Release Notes and/or PowerPath patch readme files for your specific version to determine what needs to be present/removed and if and when a reboot is necessary in order to install your specific PowerPath software version and/or patch.  These documents are available on http://Powerlink.EMC.com.

 

·         For RHEL 3.0 and for all SuSE SLES 8 distributions, verify you have 256 sd and sg devices in the /dev directory.  This prerequisite is required even if the devfs file system is mounted over /dev.

·         Ensure that the PowerPath driver’s major numbers (232-247) are not already in use. If the major numbers are already in use, the driver fails to install at system boot time and issues the following error:

emcp: Device or resource busy


·         Ensure that there is only one path (physical connection) per logical device.

 

 

29.  

Mount the PowerPath CD-ROM or grab from  website:

Install PowerPath software as follows:

 

a.     Verify that you are logged in as root.

b.    Perform the applicable substep:

·         To install from a CD-ROM: 

Insert the PowerPath installation CD-ROM and create the directory /cdrom to be the mount point for the CD-ROM by entering:

mkdir /cdrom

Mount the PowerPath CD on /cdrom by entering:

mount -o ro /dev/cdrom /cdrom

Change to the directory for your operating system:

For SLES, enter: cd /cdrom/LINUX/2.6/pp4.5.x/sles

For RHEL and Asianux, enter: cd /cdrom/LINUX/2.6/pp4.5.x/rhel

·         To installing from a compressed archive:

Download the PowerPath 4.5 archive from http://powerlink.emc.com Resources/Tools CS Support Downloads and Patches Downloads D-R PowerPath for Linux.  Untar the PowerPath archive by entering:

tar -xzf EMCpower.LINUX.4.5.<
release>.tar.gz

 

 

30.  

Install PowerPath

Install the PowerPath software as follows:

 

c.      Install PowerPath by entering the following command:

rpm –i EMCpower.LINUX-4.5.<release>.<package>.rpm

where <package> is:

Package Name

 

Description

sles.x86_64

 

PowerPath 4.5.x on SLES x86_64 platforms.

sles.ia64

 

PowerPath 4.5.x on SLES IA64 platforms.

sles.i386

 

PowerPath 4.5.x on SLES i386 platforms.

rhel.x86_64

 

PowerPath 4.5.x on RHEL and Asianux x86_64 platforms.

rhel.ia64

 

PowerPath 4.5.x on RHEL IA64 platforms.

rhel.i386

 

PowerPath 4.5.x on RHEL and Asianux i386 platforms.

 

 

 

31.  

Register PowerPath

After the installation completes, you must register PowerPath on the host:

 

a.      Locate the PowerPath License Key Card, delivered with the PowerPath media kit, which has your license registration key printed.  

b.   Enter the following command:

/etc/emcpreg –install

c.      At the prompt, type y and press Enter.

d.      At the registration key prompt, enter the PowerPath registration key and press Enter. 

If you enter a valid registration key, you see a success message.

WARNING:  Failing to enter a registration key will result in the installation of PowerPath SE, which is the default policy without a valid PowerPath license.  If you failed to enter the registration key during installation, it can be entered at a later point.

e.   Press Enter and you will see the following output:

1 key(s) successfully registered.

f.            Start PowerPath by entering:

/etc/init.d/PowerPath start

Note:  Use only the above script for EMC PowerPath module loading and unloading.


 

32.  

Post activities

After PowerPath is registered on the host, perform the following:

 

a.      Unmount the CD-ROM by entering:

cd /
unmount /cdrom


b.      Remove the CD-ROM from the CD-ROM drive.

 

c.      Verify the PowerPath capabilities are correct:

 

Enter the following command and verify the output displays the correct load balancing and failover policies.

 

powermt display dev=all

 

Note:  If the load balancing policy does not match your expectations for your license level, check the license registration.  Refer to the man pages for the emcpreg and powermt utilities for clarification on licenses and load balancing.

 

 

33.  

Check for patches

If necessary, install any PowerPath patches from the following URL:

 

    http://Powerlink.EMC.com

 

Note:  A readme file that explains how to install the patch accompanies every patch release.  This file will also state whether you need to reboot the host after the installation of the patch.

 

 

34.  

Make paths visible

The operating system must recognize new paths before PowerPath can. Enter the command:

The operating system must recognize new paths before PowerPath can do so.

 

a.     Stop all running applications, unmount any mounted filesytem.

b.    Run /etc/init.d/PowerPath stop to stop PowerPath.

c.     Unload the HBA driver module. Run rmmod <module_name>, where <module_name> is qla2300, lpfc, or some other value depending on the HBA used.

d.    Reload the HBA driver module. Run modprobe <module_name>, where <module_name> is qla2300, lpfc, or some other value depending on the HBA used.

e.     Run /etc/init.d/PowerPath start to start PowerPath.

 

 

35.  

Configure PowerPath

Configure PowerPath by running:

 

powermt release

 

then

powermt config

 

 

 

36.  

Check the configuration

Inspect the new PowerPath configuration as follows:

powermt display dev=all.

 

The new path(s) should be displayed with a state of alive.

 

Verify that each HBA sees only the targets (FAs) to which it is zoned.

 

Scan operating system error logs to ensure no errors are logged against the new paths.

Correct any issues detected.

 

37.  

Save the configuration

Save the configuration by running the following command:

 

powermt save

 

 

38.  

Check for missing PowerPath devices

If you must, check for missing PowerPath devices using the following command:

 

powermt display dev=all

 

 

39.  

Configure missing devices

Examine the output to determine if some storage system logical devices are not configured as PowerPath devices.

 

Enter the following commands to configure any missing devices:

 

powercf –q
powermt config

Rerun the powermt display dev=all command to confirm the following:
 

· the logical devices are configured as emcpower devices and the correct failover and load balancing policy is set.

· the expected LUNs are visible and each has the expected number of paths.