Front cover Implementing an Open IBM SAN Featuring the INRANGE Portfolio Discover the latest additions to the IBM SAN family Enhance your skills while using an easy-to-follow format Grow with the new technology Jon Tate Brian Cartwright Sven Eichelbaum Thomas Jahn ibm.
International Technical Support Organization Implementing an Open IBM SAN Featuring the INRANGE Portfolio December 2001 SG24-6413-00
Take Note! Before using this information and the product it supports, be sure to read the general information in “Special notices” on page 369. First Edition (December 2001) This edition applies to those IBM SAN hardware and software products described herein. Comments may be addressed to: IBM Corporation, International Technical Support Organization Dept.
Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix The team that wrote this redbook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix Special notice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi IBM trademarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii Comments welcome . . . . . . . . . . .
1.6.1 Attachment requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 1.6.2 Compaq Tru64 UNIX Version 4.0x host system . . . . . . . . . . . . . . . . 19 1.6.3 Verifying the Compaq configuration . . . . . . . . . . . . . . . . . . . . . . . . . 19 1.6.4 Operating system device recognition . . . . . . . . . . . . . . . . . . . . . . . . 19 1.6.5 Configuring AdvFS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 1.6.
1.10.9 Installing Emulex LP8000 adapter cards. . . . . . . . . . . . . . . . . . . . . 53 1.10.10 Downloading the current Fibre Channel adapter driver . . . . . . . . 54 1.10.11 Installing the Fibre Channel adapter drivers . . . . . . . . . . . . . . . . . 55 1.10.12 Parameter settings for the Emulex LP8000 on Windows NT . . . . 56 1.10.13 Verifying the configuration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 1.10.14 Verifying networking . . . . . . . . . . . . . . . . . . . . . . . . .
2.3.1 Performing the disk group configuration . . . . . . . . . . . . . . . . . . . . . 108 2.4 Creating and modifying host FC port definitions . . . . . . . . . . . . . . . . . . . 116 2.4.1 Creating host FC port definitions. . . . . . . . . . . . . . . . . . . . . . . . . . . 117 2.4.2 Modifying a host FC port definition . . . . . . . . . . . . . . . . . . . . . . . . . 123 2.5 Creating and assigning ESS logical volumes . . . . . . . . . . . . . . . . . . . . . 126 2.5.
3.4.5 Defining hard zoning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246 3.4.6 Defining name server zones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252 3.5 Cascading of directors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263 3.6 Monitoring and maintenance of an INRANGE SAN . . . . . . . . . . . . . . . . 266 3.6.1 Management communication protocols . . . . . . . . . . . . . . . . . . . . . 266 3.6.2 Microcode-loads . . . . .
5.4 Using SLIC Manager . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333 5.4.1 Drive properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334 5.4.2 Router properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334 5.4.3 Setting Router to master . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335 5.4.4 The SignOn drive . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Preface “Do everything that is necessary and absolutely nothing that is not.” In this IBM Redbook, we have tried to consolidate as much of the critical information as possible while covering procedures and tasks that are likely to be encountered on a daily basis. Each of the products described has much, much more functionality than we could ever hope to cover in just one redbook. The IBM SAN portfolio is rich in quality products that bring a vast amount of technicality and vitality to the SAN world.
the design and implementation of IBM Storage Area Networks across a wide customer base. Brian’s expertise includes SAN design, implementation, and management, and covers a wide range of disk and tape solutions across a number of different operating system platforms. Sven Eichelbaum is an I/T Specialist at StorAdvantage GmbH, an IBM Business Partner based in Germany which specializes in SAN solutions. He has 4 years of experience in designing and implementing storage architectures.
Richio Aikawa Jon Krueger Emulex Corporation Dave Burchwell Jack Consoli Richard Kurzban Mike Naylor INRANGE Technologies Corporation Chris Burke JNI Corporation Rob Jones QLogic Corporation Philipp Alexander StorAdvantage Special notice This publication is intended to help systems and storage administrators install IBM SAN portfolio equipment.
IBM trademarks The following terms are trademarks of the International Business Machines Corporation in the United States and/or other countries: e (logo)® IBM ® Redbooks Redbooks Logo AIX AS/400 BookMaster Current Early Enterprise Storage Server ESCON FICON FlashCopy iSeries Magstar Netfinity Notes NUMA-Q OS/390 OS/400 pSeries PTX RS/6000 S/390 SANergy SP StorWatch System/390 TotalStorage Wave xSeries zSeries Comments welcome Your comments are important to us! We want our IBM Redbooks to be as helpful
1 Chapter 1. Implementing Fibre Channel host adapter cards In this chapter we describe the steps involved in implementing Fibre Channel host adapter cards in an IBM TotalStorage Enterprise Storage Server (ESS) environment. Fibre Channel is a 100-MBps, full-duplex, serial communications technology to interconnect I/O devices and host systems that are separated by tens of kilometers. © Copyright IBM Corp.
1.1 Fibre Channel Fibre Channel transfers information between the sources and the users of the information. This information can include commands, controls, files, graphics, video, and sound. Fibre Channel connections are established between Fibre Channel ports that reside in I/O devices, host systems, and the network that interconnect them. The network consists of elements like switches, hubs, bridges, and repeaters that are used to interconnect the Fibre Channel ports.
1.1.2 Fibre Channel node-to-node distances You can order and have IBM install a maximum of 16 Fibre Channel adapters on ESS Models F10 and F20. Each adapter has a single host port. You can install both long-wave adapters and short-wave adapters in a single ESS. Table 1-1 lists the cables that are available for the long-wave and short-wave adapters. This table also lists the distances that the cables support. See your IBM field office about 62.5 micron cables or longer cables.
1.1.3 LUN affinity For Fibre Channel attachment, LUNs have an affinity to the host's Fibre Channel adapter through the world-wide port name (WWPN) for the host adapter. In a switched fabric configuration, a single Fibre Channel host could have physical access to multiple Fibre Channel ports on the ESS.
Feature code 3023 — This is a short wave laser adapter that includes a 31-m (100-ft) (50-micron) cable with duplex connectors. When you use the 3023 adapter to perform a FICON channel attachment, you must use the cables for FICON. The adapters cannot be shared with any Fibre Channel protocol attachments. Because the 3023 uses one of four slots in one of the four I/O bays, you can have a maximum of 16 adapters in the ESS. This allows you to have a maximum of 16 FICON interface attachments.
When you have eight paths in the path group, it minimizes the number of missed reconnections. Increasing the number of path groups does not minimize the number of missed reconnections substantially. If you use eight paths in path groups, you can increase the overall throughput. For FICON controllers, there is no synchronization between the lower DASD interface and the upper channel interface. The number of paths in the path group depend on the throughput requirement.
1.2.6 Attaching an ESS to FICON channels You can use the following FICON adapters with the IBM S/390 Enterprise Storage Servers. Feature code 2314 — This is the long-wave laser adapter. Feature code 2316 — This is the short-wave laser adapter. You can use the following FICON adapters with the IBM S/390 or zSeries system: Feature code 2315 — This is the FICON long-wave laser adapter. This adapter has two ports per adapter. This adapter is a 9-micron single mode cable, but you can use it with a 62.
The FICON architecture prohibits this capability. The reason for the restriction is because the base S/390 and zSeries I/O architecture uses a single byte for addressing the I/O devices. This one-byte I/O address is not compatible with the Fibre Channel, 3-byte port address. The FICON solution to this problem is to disallow switch cascading. 1.
1.3.2 Attachment requirements for the AS/400 or iSeries This section lists the requirements for attaching the ESS to your host system: 1. Ensure that you have all of the items in the equipment list. 2. Obtain the documents for the IBM AS/400 or iSeries host system from the following Web site: publib.boulder.ibm.com/pubs/html/as400/infocenter.htm 3. See the following Web site for details about program temporary fixes (PTFs) that you need to install on your AS/400 or iSeries host system: www.storage.ibm.
The following considerations need to be taken into account: You cannot specify a LUN size of 4.190-GB LUN for the SCSI Fibre Channel protocol (FCP) attachment. You can specify 1- 32 LUNs for each attachment to an AS/400 or iSeries Fibre Channel protocol adapter. Fibre Channel attached LUNS are identified as the 2105 device type on the iSeries host system. You can place an AS/400 or iSeries volume in the ESS storage arrays according to the selected host system attachment type.
1.3.5 Software requirements for the IBM OS/400 operating system Release V3R1.0 and later releases of the OS/400 or iSeries operating system support 4-GB and 8-GB capacity disk drives. In Table 1-3 we list the required program temporary fixes (PTFs) for OS/400 or iSeries versions that do not include the PTFs in the base code. Table 1-3 OS/400 program temporary fixes for V3R1.0 through V4R1.4 Operating System Version Level PTF V3R1.0 SF44131 V3R2.0 SF44132 V3R6.0 SF44126 V3R7.0 SF44127 V4R1.
Note: For an RS/6000 and pSeries host system, you can use two topologies: Point-to-point (switched fabric) topology Arbitrated loop topology The RS/6000 and pSeries host system does not support more than one host bus adapter on the loop. The RS/6000 and pSeries host system does support a direct connection of the RS/6000 and pSeries host system to an ESS using the Fibre Channel arbitrated loop protocol. 1.4.
4. Either you or an IBM SSR checks the LUN limitations for the RS/6000 and pSeries. See Table 2. Note: The IBM Subsystem Device Driver supports RS/6000 and pSeries host systems in a clustering environment. To have failover protection on an open system, the IBM Subsystem Device Driver requires a minimum of 2 Fibre Channel adapters. The maximum number of Fibre Channel adapters supported is 16 for a total of 16 Fibre Channel ports. 1.4.
After you install the ibm2105.rte file and all of the 2105 devices are reconfigured, vary on the volume groups and remount the file systems. The data on the file systems should be available again. 1.4.4 Installing the 2105 host attachment package Perform the following steps by using SMIT to install the IBM 2105 host attachment on your system. Install the host attachment package from a compact disc or a diskette. You must have superuser authority to complete the instructions.
14.Remove the compact disc. 15.Shut down the host system. 16.Turn on the host system. 1.4.5 Verifying the ESS configuration To verify the configuration of the ESS on the AIX host system, type the following command: lsdev -Cc disk | grep 2105 A list of all IBM ESS devices is displayed as shown in the following example: hdisk3 Available 30-68-01 IBM FC2105F20 hdisk4 Available 30-68-01 IBM FC2105F20 hdisk5 Available 30-68-01 IBM FC2105F20 These indicate that the installation is successful.
The ESS also offers the following interim support for Fibre Channel attachment: Feature Code 3019 — With the feature code 3019, you can attach an ESS to an xSeries or an IBM NUMA-Q host system through the NUMA-Q Fibre Channel- toSCSI bridge. This feature code includes one SCSI adapter that you purchase and a no-cost loan of a NUMA-Q Fibre Channel to SCSI bridge. IBM requires that you sign a loan agreement for the bridge. Note: Feature code 3019 is not a standard feature.
1. Insert the Service Pack 3 compact disc into the CD-ROM or CD-R drive. 2. Open the README file for instructions on installing Service Pack 3. In Table 1-5 we show the NUMA-Q system requirements. Support for Copy Services on PTX V4.5.2 requires a special Technology Pack. You can obtain the Technology Pack through an IBM sales representative who handles your xSeries and NUMA-Q purchases.
1.6 Compaq host system FC attachment This topic describes the host system requirements and provides the procedure to attach a Compaq Alpha server to an ESS. The ESS supports the following Compaq Alpha server models: 2100 4100 8200 8400 1.6.1 Attachment requirements This section lists the requirements to attach the ESS to your host system. Ensure that you have all of the items listed in Equipment requirements. Check the logical unit number limitations for your host system. See Table 2.
1.6.2 Compaq Tru64 UNIX Version 4.0x host system In this topic we explain how to install and configure a Tru64 UNIX Version 4.0x host system. Before you configure the ESS volumes, ensure that the system and Fibre Channel adapter cards are at the minimum supported firmware revision. 1.6.3 Verifying the Compaq configuration The following procedures tell you how to verify the configuration for the host adapter cards and firmware for a Compaq host system: 1.
You do not need to perform any special operations on the Compaq system to view the ESS volumes. Ensure that the host Fibre Channel is already configured. If it is not configured, you must perform the instructions that come with the card to install the driver. Make all the hardware connections, configure the ESS, and restart the host system. If you can see the disks when you restart, all the special files are created automatically.
/dev/rrzf128c: character special (8/262466) SCSI #16 2105F20 disk #1029 (SCSI ID #0) (SCSI LUN #5) /dev/rrzg128c: character special (8/262530) SCSI #16 2105F20 disk #1030 (SCSI ID #0) (SCSI LUN #6) /dev/rrzh128c: character special (8/262594) SCSI #16 2105F20 disk #1031 (SCSI ID #0) (SCSI LUN #7) If a special file does not have a disk associated with it, the output of the #file command will only show the part concerning the major number. No disk model is displayed.
If there is no disklabel on the disk, type the following command to write the label to the disk: #disklabel -rw rzh128 shark Substitute your specific disk with the rzh128 command. Always specify the rz file without the partition. 1.6.5 Configuring AdvFS Before you create an AdvFS file system, you must design a structure by assigning a file domain and the file sets. Type the following commands to create an AdvFS file system with file sets: 1. # cd / 2. # mkfdmn -rw /dev/rzXc vol1_dom 3.
1.7 Hewlett Packard 9000 FC host system attachment This topic describes the host system requirements and provides procedures to attach an ESS to a Hewlett Packard 9000 host system with Fibre Channel adapters. 1.7.1 Attachment requirements This section lists the requirements for attaching the ESS to your host system: Ensure that you have all of the items listed in Equipment requirements. Ensure that you have the installation script files.
4. Either you or an IBM SSR configures the host system for the ESS by using the instructions in your host system publications. Note: The IBM Subsystem Device Driver 1.1.3 supports the Hewlett Packard host system in a clustering environment. To have failover protection on an open system, the IBM Subsystem Device Driver requires a minimum of two Fibre Channel adapters. The maximum number of Fibre Channel adapters supported is 16 for a total of 16 Fibre Channel ports. 1.7.
6. From the Software Selection window, click IBMis_tag. 7. From the Action menu, click Mark for Install. 8. When you see the word Yes next to the IBMis_tag product, go to the Action menu and click Install. 9. When the analysis completes with no errors (Status- Ready), click OK. 10.Click Yes in the Confirmation window to begin the installation. A window opens, notifying you that the installation is complete and that the system needs to be restarted. 11.Click OK to continue. 1.7.
3. Extend the volume group to include the c3t0d0 path. When you issue a vgdisplay -v command on the volume group, it lists c3t0d0 as an alternate link to the data. 1.8 Novell NetWare FC host system attachment This topic describes how to attach a Novell NetWare host system to an IBM ESS with the following adapter cards: QLogic QLA2100F QLogic QLA2200F Note: The IBM SAN Fibre Channel Switch 2109 S08 and IBM SAN Fibre Channel Switch 2109 S16 are supported for Novell NetWare.
6. From the Advanced Adapter Settings menu, press the Down Arrow to highlight LUNs per target. Press Enter. 7. Press the Down Arrow to find and highlight 256. Press Enter. 8. Press Esc. 9. To save the changes, click Yes. Press Enter. 10.Restart the server. 1.8.2 Installing the QLogic QLA2200F adapter card This section tells you how to attach an ESS to a Novell NetWare host system with the QLogic QLA2200F adapter card.
a. Execution throttle: 240 b. Fast command posting: Enabled c. >4 GB addressing: Disabled for 32 bit systems d. LUNs per target: 0 e. Enable LIP reset: No f. Enable LIP full login: No g. Enable target reset: Yes h. Login retry count: 20 (minimum) i. Port down retry count: 20 (minimum) j. Driver load RISC code: Enabled k. Enable database updates: No l. Disable database load: No m. IOCB allocation: 256 n. Extended error logging: Disabled (might be enabled for debugging) 7.
1.8.3 Loading the current Fibre Channel adapter driver Perform the following steps to load the current driver onto the QLogic adapter card. 1. Go to the following Web site: www.qlc.com 2. From the home page, click Driver Download. 3. Click Drivers. 4. Click Fibre-Channel Adapter Drivers. 5. Click QLA2xxx drivers. 6. Click Novell NetWare. 7. Click QLogic Vx.xxx where V is the version and x.xxx is the version level of the file name. 8. In the Save As window, find the current driver file, xxxxxxx.
5. Click “Select an additional driver.” 6. Press the Insert key. 7. Insert a floppy diskette with the QLogic drivers into the A:\ drive of the NetWare server. Press Enter. The available driver is displayed. 8. Select the driver for the QLogic card and press Enter. 9. Select Modify driver parameters and enter the slot number of the QLogic card into the slot number parameter. 10.Set the Scan All Luns parameter to Yes. 11.Press Tab and select Save Parameters and Load Driver. 12.
Solaris 2.6, Solaris 7, and Solaris 8 require patches to ensure that the host and the ESS function correctly. See Table 1-6 for the minimum revision level that is required for each Solaris patch ID. Table 1-6 Solaris 2.6, 7, and 8 minimum revision level patches for Fibre Channel Solaris 2.
4. Either you or an IBM SSR configures the host system for the ESS by using the instructions in your host system publications. Note: The IBM Subsystem Device Driver does not support the Sun host system in a clustering environment. To have failover protection on an open system, the IBM Subsystem Device Driver requires a minimum of two Fibre Channel adapters. The maximum number of Fibre Channel adapters supported is 16 for a total of 16 Fibre Channel ports.
6. Replace the computer case by tightening the screws on the case or use the clamp to secure the cover. 1.9.3 Downloading the current Emulex Fibre Channel adapter driver This section tells you how to install the Emulex driver: 1. Plug in and restart your host system. 2. Go to the following Web site: www.emulex.com 3. From the left navigation menu for Quick Links, click Documentation, Drivers and Software. 4. From the Fibre Channel menu, select the adapter model. 5. Click Drivers for Solaris. 6.
6. If the file is in the format of filename.tar.Z, type uncompress filename.tar.Z. 7. Type tar xvf lpfc-sparc.tar to “untar” the drive file from the temporary directory. 8. Type pkgadd -d pwd to install the package. Note: An installation script displays that prompts you to answer a number of questions. For each question, enter the appropriate response. Or, press Enter to each question to accept the default setting. 9. Specify the package number or press Enter to accept all packages. 10.
18.Update the parameter list. See Table 1-8 on page 42. 19.Restart the host system. Note: You must restart your system twice if IP is enabled. 1.9.5 Installing the JNI PCI adapter card This section tells you how to attach an ESS to a Sun host system with the JNI PCI adapter card: 1. Perform the following steps to install the JNI PCI adapter card. 2. Turn off and unplug the computer. 3. Remove the computer case. 4. Remove the blank panel from an empty PCI bus slot. 5.
11.If you downloaded the driver file from a Sun host system, go to Installing the JNI PCI Fibre Channel adapter driver. If you downloaded the driver file from a non-Sun host system, transfer the drive file to a Sun host system. 1.9.7 Installing the JNI PCI Fibre Channel adapter driver Perform the following steps to install the JNI PCI Fibre Channel adapter drivers: 1. Go to the following Web site: www.jni.com 2. From the navigation menu at the top of the page, click Drivers. 3.
1.9.9 Downloading current JNI SBUS Fibre Channel adapter driver This section tells you how to download the JNI SBUS Fibre Channel adapter driver: 1. Plug in and restart your host system. 2. Go to the following Web site: www.jni.com 3. From the navigation menu at the top of the page, click Drivers. 4. From the menu for Locate Driver by Product, click FC64-1063. 5. From the menu for FCI-1063, find the section for Solaris -JNI. Click fcw.pkg. 6.
1.9.11 Installing the QLogic QLA2200F adapter card This section tells you how to attach an ESS to a host system with the QLogic QLA2200F adapter card. Single-and dual-port Fibre Channel interfaces with the QLogic QLA2200F adapter card support the following public and private loop modes: Target Public initiator Private initiator Target and public initiator Target and private initiator Perform the following steps to install the QLogic QLA2200F adapter card: 1.
l. Disable database load: No m. IOCB allocation: 256 n. Extended error logging: Disabled (might be enabled for debugging) Note: The Enable LIP reset, Enable LIP full logon, and Enable target reset parameters control the behavior of the adapter when Windows NT tries to do a SCSI bus reset. You must perform a target reset to make cluster failovers work.Use the SCSI bus device reset option to clear SCSI reservations. 8. Press Esc to return to the Configuration Settings menu. 9.
2. From the home page, click Driver Download. 3. Click Use QLogic Drivers. 4. Click Fibre Channel Adapter Drivers and Software. 5. In table, click QLA22xx drivers. 6. From the menu for Software and Drivers available for, click Solaris. 7. From the table for QLA22xx Driver Download Page, Solaris Sparc 2.6/2.7/2.8, and Current released (Sparc) driver PCI to FC Adapter, click Link to Driver. 8. Download the file and save as required. 1.9.
1. Change to the directory by typing: cd /kernel/drv 2. Back up the sd.conf file in this subdirectory. 3. Edit the sd.conf file to add support for the target and LUN pairs that are configured on the host system. The following is an example of the lines that you would add to the file to access LUNs 0-7 on target 8.
See Table 1-7 for the recommended configuration settings for the host-bus-adapter for a JNI FC64-1063 and a JNI FCI-1063. Table 1-7 Recommended JNI HBA settings Parameters Recommended settings fca_nport 0: Default. Initializes on a loop 1: Recommended for fabric. Initializes as an N_Port. public loop 0: Default. Recommended. Initializes according to which fca_nport is set as disabled. ip_disable 0: Default. IP side of the driver is enabled. 1: Recommended for fabric.
Parameters Recommended settings scan-down 2: Recommended. Use inverted ALPA map and cause target assignment in private loop. topology 2: Recommended for fabric. Point-to-point mode only. 4: Recommended for non-fabric. Arbitrated loop mode only. zone-rscn 0: Default 1: Recommended for fabric; check name server for RSCNs. In Table 1-9 we show the recommended configuration file parameters for the host bus adapters for the QLogic QLA2200F adapter.
1.9.16 Installing the IBM Subsystem Device Driver The following instructions explain how to install the IBM Subsystem Device Driver from a compact disc. You can use the IBM Subsystem Device Driver in conjunction with the IBM Copy Services command-line interface program. 1. Type ps -ef |grep vold to ensure that the volume manager is running. This command displays the /usr/sbin/vold process. If it does not display, type: /etc/init.d/volmgt start 2.
1.9.17 Setting the Sun host system parameters The following sections contain the procedures to set the Sun host system parameters for optimum performance on the ESS with the following adapters: JNI Emulex QLogic JNI adapters The following sections contain the procedures to set the Sun host system parameters for optimum performance on the ESS with the JNI adapter: 1. Type cd /etc to change to the /etc subdirectory. 2. Back up the system file in the subdirectory. 3.
maxphys This parameter specifies the maximum number of bytes you can transfer for each SCSI transaction. The default value is 126976 (124 KB). If the I/O block size that you requested exceeds the default value, the request is broken into more than one request. The value should be tuned to the intended use and application requirements.
set maxphys=8388608 If you are using Veritas volume manager on the ESS LUNs, you must set the VxVM max I/O size parameter (vol_maxio) to match the maxphys parameter. If you set the maxphys parameter to 8388608, add the following line to the /etc/system file to also set the VxVM I/O size to 8 MB: set vxio:vol_maxio=16384 1.10 Windows NT 4.0 FC host system attachment In this topic we tell you how to attach an ESS to a Windows NT host system with the following Fibre Channel adapters.
3. Either you or an IBM SSR defines the Fiber Channel port configuration if you did not do it during the installation of the ESS or Fibre Channel adapters. Note: Use the information on the logical configuration work sheet in the IBM Enterprise Storage Server Configuration Planner that you should have previously filled out. You or an IBM SSR configures the host system for the ESS by using the instructions in your host system publications. Notes: 1. The IBM Subsystem Device Driver 1.2.
Perform the following steps to install the QLogic QLA2100F adapter card: 1. Install the QLogic QLA2100F adapter card in the host system. 2. Connect the cable to the ESS port. 3. Restart the host system. 4. Press Alt+Q to get to the FAST!Util menu. 5. From the Configuration Settings menu, select Host Adapter Settings. 6. From the Advanced Adapter Settings menu, press the Down Arrow to highlight LUNs per target; then press Enter. 7. Use the Down Arrow to find and highlight 256. Press Enter. 8. Press Esc. 9.
7. From the Advanced Adapter Settings menu, press the Down Arrow to highlight LUNs per target. Press Enter. Set the parameters and values from the Advanced Adapter Settings menu as follows: a. b. c. d. e. f. Execution throttle: 100 Fast command posting: Enabled >4 GB addressing: Disabled for 32 bit systems LUNs per target: 0 Enable LIP reset: No Enable LIP full login: No Note: In a clustering environment, set Enable LIP full login to Yes. g. h. i. j. k. l. m. n.
Note: If you connect the ESS directly to the host system, the option you select must match the port connections on the ESS. However, if you connect through a switch, the options do not need to match the port connections because the ESS is point-to-point. The appropriate HBA on the server must also support point-to-point connection on a direct connection. Currently, disparate vendors do not function properly in a direct point-to-point connection.
16.Click OK. 17.Click Close to close the window for WinZip Self-Extractor. 1.10.5 Installing the Fibre Channel adapter drivers Perform the following steps to install the Fibre Channel adapter drivers. Note: If you are installing the Fibre Channel adapter for the first time, you must specify the correct topology. You must also select the appropriate device mapping driver. 1. From your Windows NT desktop, double-click the icon for My Computer. 2. Double-click the icon for Control Panel. 3.
The host bus adapter uses the time-out parameter to bound its recovery actions and responses to the disk subsystem. The value exists in different places in the system configuration. You can retrieve and use it in different ways, depending on the type of host bus adapter. The following instructions tell you how to modify the value safely in either the Windows NT registry or in the device adapter parameters. 1.10.
Target Public initiator Private initiator Target and public initiator Target and private initiator The ESS supports increased connectivity with the use of Fibre Channel (SCSI-FCP and FICON) directors. Specific details on status, availability, and configuration options for the Fibre Channel directors supported by the ESS are available on the Web: www.storage.ibm.com/hardsoft/products/ess/supserver.htm Perform the following steps to install the Emulex LP8000 adapter card: 1.
9. From the File Download window, click the appropriate radio button and proceed as indicated: – To open this file from its current location, go to step 9. – To save this file to disk, go to step 16. 10.In the window for Winzip, click I agree. 11.In the window for WinZip Wizard - Welcome, click Next. 12.In the window for WinZip Wizard - Select Zip File xxxxxxxx.zip, where xxxxxxxx is the name of the file, highlight the file that you want to unzip. 13.Click Next. 14.
5. Click Add to create a list of drivers. A window opens that indicates the progress. When the window closes, you should see a window called Install Driver. 6. From the Install Driver window, click Have Disk. 7. Enter the path to the driver file that you downloaded and click OK. 8. To install the driver, highlight the line that lists the driver you want and click OK. Note: The driver affects every adapter in the system.
Parameters Recommended settings Use name server after RSCN Checked (enabled) if fabric attached using soft zoning Not checked (disabled) Lun mapping Checked (enabled) Automatic lun mapping Checked (enabled) Scan in device ID order Not checked (disabled) Enable class 2 for SCSI devices Not checked (disabled) Report unknown SCSI devices Not checked (disabled) Look for disappearing devices Not checked (disabled) Translate queue full to busy Not checked (disabled) Use bus reset status for retr
1.10.13 Verifying the configuration The following tells you how to determine whether or not your Windows NT 4.0 host system is configured for storage: 1. From the Windows NT desktop, right-click Start. 2. Partition new Fibre Channel drives with Microsoft Disk Administrator. See your Windows NT documentation for more instructions. 3. Click Explore and verify that you can see the Fibre Channel drives. 4. Select a large file (for example, a 9 MB file), and copy it to a Fibre Channel drive. 1.10.
Either you or an IBM SSR must perform the following tasks to install and configure an ESS: 1. The IBM SSR installs the ESS by using the procedures in the IBM Enterprise Storage Server Service Guide 2. You or an IBM SSR defines the Fibre Channel host system with the worldwide port name identifiers. 3. You or an IBM SSR defines the Fiber Channel port configuration if you did not do it during the installation of the ESS or Fibre Channel adapters. For the list of worldwide port names, see 1.
Note: 1. The arbitrated loop topology is the only topology available for the QLogic QLA2100F adapter card. 2. When you install the QLA2100F adapter card on your host system, IBM recommends that you install the card on a host system with four processors. This ensures that you do not have a problem when you start the host system. Perform the following steps to install the QLogic QLA2100F adapter card: 1. Install the QLogic QLA2100F adapter card in the host system. 2. Connect the cable to the ESS port. 3.
4. Press Alt+Q to get to the FAST!Util menu. 5. From the Configuration Settings menu, select Host Adapter Settings. 6. From the Host Adapter Settings menu, set the following parameters and values: a. Host adapter BIOS: Disabled b. Frame size: 2048 c. Loop reset delay: 5 (minimum) d. Adapter hard loop ID: Disabled 7. From the Advanced Adapter Settings menu, press the Down Arrow to highlight LUNs per target. Press Enter. Set the parameters and values from the Advanced Adapter Settings menu as follows: a.
10.From the Extended Firmware Settings menu, scroll down to Connection Options to open the Option and Type of Connection window. 11.Select one of the following options: a. 0: Loop only b. 1: Point-to-point (preferred setting) c. 2: Loop preferred (If you cannot use arbitrated loop, then default to point-to-point) d. 3: Point-to point, otherwise loop (If you cannot use point-to-point, default to arbitrated loop).
11.Click Save. 12.When the download completes, click Open Folder. 13.From the A:\ window, double-click the icon for the driver file you downloaded to the floppy diskette. 14.In the window for WinZip Self-Extractor, type a:\ 15.Click Unzip. When the unzip process completes, you should see a message that says, x files unzipped successfully, where x equals the number of files you unzipped. 16.Click OK. 17.Click Close to close the window for WinZip Self-Extractor. 1.11.
1.11.7 Configuring for availability and recoverability This section describes how to ensure optimum availability and recoverability when you attach an IBM ESS to a Windows 2000 host system. You must set the time-out value associated with the supported host bus adapters to 240 seconds. The setting is consistent with the configuration for IBM SSA adapters and disk subsystems when attached to Windows 2000 host system.
1.11.9 Installing Emulex LP8000 adapter cards This section tells you how to attach an ESS to a Windows 2000 host system with Emulex LP8000 adapter cards. Single-and dual-port Fibre Channel interfaces with the Emulex LP8000 adapter cards support the following public and private loop modes: Target Public initiator Private initiator Target and public initiator Target and private initiator The ESS supports increased connectivity with the use of Fibre Channel (SCSI-FCP and FICON) directors.
8. From the File Download window, click the appropriate radio button and proceed as indicated: – To open this file from its current location, go to step 9. – To save this file to disk, go to step 16. 9. In the window for Winzip, click I agree. 10.In the window for WinZip Wizard - Welcome, click Next. 11.In the window for WinZip Wizard - Select Zip File xxxxxxxx.zip where xxxxxxxx is the name of the file, highlight the file that you want to unzip. 12.Click Next. 13.
5. Click Add to create a list of drivers. A window opens that indicates the progress. When the window closes, you should see a window called Install Driver. 6. From the Install Driver window, click Have Disk. 7. Enter the path to the driver file that you downloaded and click OK. 8. To install the driver, highlight the line that lists the driver you want and click OK. Note: The driver affects every adapter in the system.
68 Parameters Recommended settings Use name server after RSCN Checked (enabled) if fabric attached using soft zoning Not checked (disabled) Lun mapping Checked (enabled) Automatic lun mapping Checked (enabled) Scan in device ID order Not checked (disabled) Enable class 2 for SCSI devices Not checked (disabled) Report unknown SCSI devices Not checked (disabled) Look for disappearing devices Not checked (disabled) Translate queue full to busy Not checked (disabled) Use bus reset status for
1.11.13 Verifying the configuration The following tells you how to determine whether or not your Windows 2000 host system is configured for storage. 1. From the Windows 2000 desktop, right-click Start. 2. Partition new Fibre Channel drives with Microsoft Disk Administrator. See your Windows 2000 documentation for more instructions. 3. Click Explore and verify that you can see the Fibre Channel drives. 4. Select a large file (for example, a 9 MB file), and copy it to a Fibre Channel drive. 1.11.
If your host system uses more than one Fibre Channel adapter to connect to your ESS, you must add multiple entries to the host system list for this host, one for each Fibre Channel adapter. Each adapter will have its own unique WWPN. The format and content of the Fibre Channel port identifier are determined by the manufacturer of the link control facility for the applicable Fibre Channel port.
1.12.3 Locating the WWPN for an IBM eServer or IBM NUMA-Q host To locate the WWPN for a NUMA-Q host system with an IOC-0210-54 adapter, perform the following steps from the IBM Enterprise Storage Specialist: 1. From the Enterprise Storage Specialist Welcome panel, click Storage Allocation. 2. From the Storage Allocation graphical view panel, click Open System Storage. 3. From the Open System Storage panel, click Modify Host Systems. 4. In the Host Nickname field, type the nickname. 5.
Probing timeout item adapter WWN Cur.Topo Next Topo [0 ]pga0.0.0.7.1 1000-0000-c922-d469 FABRIC FABRIC [1 ]pgb0.0.0.8.1 2000-0000-c922-6a63 FABRIC FABRIC [9999 ]All of the above. If you receive the following error: wwidmgr available only prior to booting. Reinit system and try again. Type P00>>>init and repeat the wwidmgr command. If you receive the following error: wwidmgr: No such command If you see this message, type P00>>set mode diag and type the wwidmgr command again.
For example, to look for the device path name /dev/td1, type: fcmsutil /dev/td1 where /dev/td1 1.12.7 Locating the WWPN for a Novell NetWare host system To locate the WWPN for a Novell NetWare host system with a QLogic adapter, perform the following steps: 1. Restart the server. 2. Press Alt+Q to get the FAST!Util menu. If you have more than one Fibre Channel adapter installed, you will see a screen that displays all the Fibre Channel adapters. Scroll down to the adapter you want. 3. Press Enter. 4.
1.12.9 Locating the WWPN for a Windows NT host system To locate the WWPN for a Windows NT host system with a QLogic adapter, perform the following steps: 1. Restart the server. 2. Press Alt+Q to get the FAST!Util menu. If you have more than one Fibre Channel adapter installed, a panel displays all the Fibre Channel adapters. Scroll down to the adapter you want. Press Enter. 3. From the Fast!Util menu, scroll down and select Select Host Adapter. 4. Scroll up and highlight Configuration Settings. 5.
2 Chapter 2. IBM TotalStorage Enterprise Storage Server configuration One benefit of a SAN is to implement disk pooling. To do this successfully, we need an easy-to-manage storage server with a Fibre Channel attachment. This is the IBM TotalStorage Enterprise Storage Server, 2105-F20, shown in Figure 2-1. First we introduce the ESS Web Interface, including the functions to maintain, configure, and administer the ESS. Then we perform storage configuration.
For further information on the configuration of the ESS, especially for ESCON and SCSI attachment and the ESS Web Copy Services, refer to the ESS documentation, as described in “Related information: For more information on the ESS, see these references:” on page 172.
2.1 Introducing the ESS Web Interface The Java based Web interface consists of two consoles, the IBM TotalStorage Enterprise Storage Server Specialist (ESS Specialist) and the IBM TotalStorage Enterprise Storage Server Web Copy Services (ESS Web Copy Services). The ESS Specialist is used for querying the status of the ESS and for displaying, defining and modifying the operation, configuration and administration of the ESS.
2.1.1 Logging on to the ESS Specialist Here we will detail the steps to log on to the ESS Specialist. Viewing the Introduction panel Once connected to one of the ESS clusters, either reloading the Web page, using the Reload button of the Web browser, or clicking the Introduction button will take us to the introduction panel of the ESS Specialist, as shown in Figure 2-2.
To the left, we see the main button panel, from where we can access the viewing and configuration panels of the ESS Specialist. Also, from there we can connect to and troubleshoot the ESS Web Copy Service. On the main window we can see the machine type, machine model, and serial number displayed. Also, we see the identification of the Fibre Channel ESS in the Fibre Channel fabric in the form of its World Wide Node Name (WWNN).
Figure 2-4 New Site Certificate, information window More information on the certificate, including the certificate fingerprint and the time frame for which this certificate is valid, can be displayed by pressing the More Info button, as shown in Figure 2-5.
After closing the Information window, we are asked about our decision to accept the certificate or not, and if we would like to accept the certificate until it expires. This is shown in Figure 2-6. Figure 2-6 New Site Certificate, Accepting the certificate Logging in with the default user After completing the certification check, the Login Window, shown in Figure 2-7, will display. The default user is storwatch and the default password is specialist.
Figure 2-7 Site login window The session will be valid as long as any browser window is open. Therefore, when the work has been performed and security is important, we suggest closing the browser to log off from the ESS specialist. Be aware that a problem within the browser could keep the browser process running, even when all browser windows are closed. This could lead to the possibility of unauthorized personnel accessing the ESS with the Web interface.
Viewing the Status — Graphical View panel Using the Status button, we get to the Status — Graphical View panel, as shown in Figure 2-8. Figure 2-8 Status — Graphical View panel This panel allows us to quickly check for any errors in the ESS. To demonstrate the features of the Status — Graphical View panel and the Status — Problem Log panel we show them on a different ESS that is currently in service. What we see is a problem detected in the storage arrays, with access to data not affected.
Viewing the Status — Problem Log panel To get a detailed view on the problems encountered, we use the Problem Log button on the upper right of the screen and get the Status — Problem Log panel, as shown in Figure 2-9. Figure 2-9 Status — Problem Log panel Using the Graphical View button, we are able to go back to the graphical view.
Viewing the Problem Notification panel Using the Problem Notification button, we get the Problem Notification panel, shown in Figure 2-10. Figure 2-10 Problem Notification panel From within this panel we can configure the various types of problem notification provided with the ESS. The options available are e-mail, pager and Simple Network Management Protocol (SNMP). Chapter 2.
Viewing the Communications panel Using the Communications button, we get the Communications panel, as shown in Figure 2-11. Figure 2-11 Communications panel Here we can view the network information for which the ESS is configured, and verify if the Call Home feature and Remote Service Access is enabled. Also, using the Reset PE Password button, we can reset the PE password. Only users with administration or configuration levels of authority can modify these panels.
Generating a new PE password Resetting the PE password means actually that you generate a password for Product Engineers (PE), which they can use to access the ESS from remote for repair actions. By pressing the Reset PE Password button, a warning window will appear, and after clicking the Yes button, the window with the new PE password will be displayed. This password grants the PE unlimited access to the ESS LIC and the ESS configuration files for 168 hours (7 days).
Viewing the Licensed Internal Code panel Using the Licensed Internal Code button we get the “Licensed Internal Code” panel, as shown in Figure 2-13. Figure 2-13 Licensed Internal Code panel Here we can see the LIC levels of the ESS clusters and the licensed feature codes that include all additional ESS functionality like copy services.
2.1.3 Configuring the ESS Following, we introduce the panels that serve as entry points to the logical configuration of the ESS and link to the ESS Web Copy Services. Viewing the Storage Allocation — Graphical View panel Using the Storage Allocation button we get the Storage Allocation — Graphical View panel, shown in Figure 2-14. Figure 2-14 Storage Allocation — Graphical View panel This is the start panel for host and storage based configurations.
Using the Copy Services link Using the Copy Services button we link to the copy services server. We have to login to gain access, just like we did for the ESS Specialist. This opens the Java based Web interface for the ESS Web Copy Services server in a dedicated Navigator window, connects to it and loads the Java applets, as shown in Figure 2-15.
Figure 2-16 ESS Web Copy Services introduction panel Using the Tools help page Using the Tools button we get a help page with embedded action links for troubleshooting the ESS Web Copy Services. As mentioned before, the ESS Web Copy Services is a separate server program running on an ESS cluster, which is accessed through a different browser window than the ESS Specialist.
Figure 2-17 Tools help page, Web Copy Services trouble shooting 92 Implementing an Open IBM SAN
Figure 2-18 Tools help page, Web Copy Services trouble shooting, continued 2.1.4 Administer the ESS We will now introduce the administration of users on the ESS. Using the User Administration panel Using the Users button we get the User Administration panel, as shown in Figure 2-19. Chapter 2.
Figure 2-19 User Administration panel In our case there are no users configured, other than the default user that we logged on with. Using the Modify Users panel In the Modify Users panel, as shown in Figure 2-20, we can add users with different access levels.
Figure 2-20 Modify Users panel The access levels are: View Viewing the status and the configuration. Operation Viewing the status and the configuration and performing operation functions. Configuration Viewing the status and the configuration, performing operation and configuration functions. Administration Viewing the status and the configuration, performing operation, configuration and administration functions.
2.2 Starting the Open Systems Storage Configuration In the topics that follow, we will guide users through the storage and Fibre Channel environment related parts as we progress through and build the configuration. The tasks performed here will affect the way the ESS presents itself in our SAN and to the SAN attached Fibre Channel hosts. 2.2.1 Viewing the storage configuration status The panels described here will always show the status of the configuration of the ESS.
In this panel, we get the logical view and the status of the installed and configured storage related components of the ESS. These are the parts of the ESS that need to be configured for ESS logical volumes to be accessible from host FC ports. The panel is interactive, which means, as we select different entities on this panel, the panel will reflect the configuration of the ESS according to the selected entity.
Using the ESS interfaces row Below the host interface row, we see the row with icons for the ESS installed interface adapters, shown in Figure 2-24. Figure 2-24 Installed ESS interfaces row The row is divided into four columns to show the location in the ESS interface bays. Our ESS is equipped with four ESCON adapters and four SCSI adapters, with one of them in each bay, and 8 Fibre Channel adapters, with two of them in each bay.
Figure 2-26 ESS SCSI adapter with two ports Figure 2-27 ESS FC adapter with one port Viewing the association between host ports and ESS ports In the ESS, we do not tie a host FC port to a Fibre Channel port in the ESS, which was what we did with the SCSI adapters. Every host FC port within the Fibre Channel fabric will be identified by its WWPN and can access data through every ESS FC port unless other measures are taken.
Figure 2-28 FC host port selected In contrast to this, the host SCSI ports have an affinity to the ESS SCSI ports, which is visible by selecting a host SCSI port, as shown in Figure 2-29. Figure 2-29 SCSI host port selected Using the disk group section The two ESS clusters with the SSA device adapters and the SSA loops with the disk groups are shown below the ESS Fibre Channel ports. This is shown in Figure 2-30 and is how it looks if there is no disk group configured.
Figure 2-30 SSA device adapters without any disk group configured Again, we are anticipating the configurations that are to be made later. To graphically show the state the disk groups are in, we press the View All Storage button on the upper right of the panel. This is shown in Figure 2-31. Disk groups can also be selected to show the state of the selected storage.
Figure 2-31 Disk groups, view all storage mode To show the state of the disk groups, which contain ESS logical volumes assigned to a particular host FC port, we select the host FC port first and then one of the yellow highlighted ESS FC ports, which is shown in Figure 2-32.
There are different colors to indicate the state of the disk groups. At the upper right of the screen, shown in Figure 2-33, a legend is displayed, which explains what the different colors in the disk groups represent. Figure 2-33 Legend of the disk group status colors The colors indicate the following: Purple, Host Storage Disk storage that contains ESS logical volumes that are assigned to a specific (selected) host FC port. This displays only if a host FC port is selected first.
Figure 2-34 Tabular View, no ESS logical volumes configured As expected, the table that contains the detailed description of all volume assignments made, is empty right now. 2.2.2 Using the Open System Storage panel To start with the configuration of the ESS, we press the Open System Storage button, as shown in Figure 2-35. This is the door to the configuration of open systems storage. It will lead us to the Open System Storage panel, as shown in Figure 2-36.
Figure 2-36 Open System Storage panel, no host FC ports defined The configuration panel buttons at the bottom are the entry points to all of our storage related configuration tasks. The Open System Storage panel is like a hub, from where we start every open systems storage configuration. In this section, we will indicate our use of the configuration panel buttons by pointing the mouse on the button.
Configuring disk groups and defining host FC ports to be able to add ESS logical volumes Configuring the ESS FC ports for FC-SW or FC-AL Adding ESS logical volumes to disk groups and assigning them to host FC ports Accessing the assigned ESS logical volumes Figure 2-37 ESS configuration dependencies 106 Implementing an Open IBM SAN
One step that is independent from the others, is the configuration of the ESS FC ports for FC-SW or FC-AL. For this we have to know what type of FC topology we want to connect with a particular ESS FC port. Other steps rely on the completion of others. In order to add ESS logical volumes to disk groups, we need to perform two configuration steps. Those are the configuration of some disk groups for RAID or JBOD and the definition of one or more host FC ports by its host type and WWPN.
2.3.1 Performing the disk group configuration Pressing the Configure Disk Groups button takes us to the Fixed Block Storage panel, as shown in Figure 2-38. Figure 2-38 Fixed Block Storage panel, no disk groups configured Using the Fixed Block Storage panel The Available Storage table lists all of the available physical storage in the ESS with its location, storage type, track format, and capacity. As we can see, when we scroll down the table, no disk array has been formatted at this time.
Note: The disk groups should not be mistaken for the eight packs, which are the physical disk packages, that are installed in the ESS. The disk groups are logical groups that consist of disks from two eight packs, four disks from each of the two. This is the reason why it is only possible to upgrade the ESS in steps of two eight packs. The unformatted size of the disk groups is 254.8 GB. This is equivalent to seven disks of a capacity of 36.4 GB each. In the ESS, we have two spare disks per SSA loop.
Figure 2-40 Disk groups defined for RAID 5 Performing the configuration update To apply the changes made, we press the Perform Configuration Update button. A warning message appears, stating that this will be a time consuming action, as shown in Figure 2-41.
How time consuming this will be depends on how many disk groups are to be configured at once. Pressing the OK button executes the script which changes the configuration and gives us a progress window, where we can see which arrays are initializing, as shown in Figure 2-42. Figure 2-42 Progress window, RAID configuration After completing the configuration, the ESS Specialist informs us of the success, as shown in Figure 2-43.
Viewing changes on the Graphical View panel From here we can continue with our configuration. To see how the Storage Allocation — Graphical View panel now looks, we press the Storage Allocation button on the left of the screen. The panel shown in Figure 2-44 now illustrates the representation of the two disk groups between the SSA device adapter pair four in loop A. It is the view of the panel with the disk group in cluster one selected.
The rectangle representing the disk group is green to show that this is empty storage. We also see that the host FC port icon row is empty. When selecting a particular disk group, the row will only present and highlight those host FC port icons that have ESS logical volumes assigned to it in the selected disk group. Because there is no host FC port defined, there cannot be any ESS logical volume defined and assigned.
Defining two more disk groups for RAID 5 We will configure two more disk groups and so we therefore navigate back to the Fixed Block Storage panel. In Figure 2-46 we show what the Available Storage table looks like with the two disk groups in loop A of the fourth adapter pair for RAID configured. This is prior to performing the configuration update of the next two disk groups.
Viewing changes on the Graphical View panel Device adapter pair number four and the connected SSA loops now look like that shown in Figure 2-47. Figure 2-47 Four disk groups configured Undefining disk groups or redefining disk groups It may be necessary to reconfigure disk groups. They can be unconfigured or configured to another storage type. If there is a disk group configured for RAID 5, it can be reformatted to JBOD or unconfigured. Doing this is the same process as we did before.
Figure 2-48 Status — Graphical View, four disk groups in the process of formatting This discussion was just intended to show how it is done. Now we continue from the point where we defined the disk groups. 2.4 Creating and modifying host FC port definitions To fulfil the second condition to be able to define and assign ESS logical volumes, we will define the host FC ports that are installed in the Fibre Channel attached hosts.
2.4.1 Creating host FC port definitions From the Open System Storage panel, we press the Modify Host Systems button, as shown in Figure 2-49. Figure 2-49 Entry to the Modify Hosts Systems panel Viewing the Modify Host Systems panel We are now taken to the Modify Host Systems panel, as shown in Figure 2-50. This panel consists of the Host Attributes entry fields and The Host Systems List table. The Host Systems List represents each defined host FC port.
Creating two host FC port definitions In the Host Attributes entry fields, we identify the host FC port that will access the ESS. The first entry is a nickname to identify the port in the ESS. Also, we have to choose the host type, which is a PC server; the type of the host attachment, which is Fibre Channel, and the WWPN of the host FC port, as shown in Figure 2-51 and Figure 2-52. We are doing this for the two host FC ports that we want to access ESS logical volumes, nicknamed PC1_1 and PC1_2.
Figure 2-52 Host Attributes frame, host interface type selection The IP Address is only used by the StorWatch Enterprise Storage Server Expert. It is used to show which host FC ports are in a specific IP host. We will not be using the Expert for our example, so we leave the entry field empty. By pressing the Add button, we add the specified host FC port to the Host Systems List. We do the same for a second host FC port, as shown in Figure 2-53. Chapter 2.
Figure 2-53 Host Systems List, host FC ports defined This host FC port can belong to the same system, or we can define ports for other systems. Performing the configuration update Adding ports to the list puts together a script in the background which is executed by pressing the Perform Configuration Update button. A progress bar informs us about the steps that are being executed, as shown in Figure 2-54.
A window appears which indicates that the configuration update was successful, as shown in Figure 2-55. If the update had failed for any reason, we would get an error message. Figure 2-55 Successfully added host FC port Pressing the OK button on this window will immediately take us back to the Open System Storage panel. Viewing changes on the Open System Storage panel We now see the newly configured host FC ports in the Host Systems Table, as shown in Figure 2-56.
As we select one of them, we see that there are no ESS logical volumes currently assigned to the host FC ports. Viewing changes on the Graphical View panel Now, we would like to see what the Storage Allocation — Graphical View panel looks like. This is shown in Figure 2-57. Figure 2-57 Four disk groups, two host FC ports The two host FC ports we just have configured, now show up with their nicknames in the host interface row.
Clicking on the host FC port icon also shows that the port would be able to access volumes through each of the eight installed ESS FC ports if they were all connected to the SAN and no zoning in the fabric would prevent them from accessing the ESS FC ports. This is illustrated in Figure 2-59, with lines drawn from the selected host FC port to each installed ESS FC port. The involved components also get highlighted yellow. Figure 2-59 Selected host FC port, no ESS FC port restrictions 2.4.
Figure 2-60 FC port PC1_2 selected for modification As we show here, the Fibre-Channel Ports list indicates the current setup, which is not restricted in ESS FC ports. It lists all ESS FC ports that are available as shown in Figure 2-61.
We scroll to and select the adapter tree in bay 1 to limit the host FC port PC1_2 to this ESS FC port. We could also select any other port in addition to the one we have chosen. To update the Host Systems List table, we press the Modify button. The restriction is now reflected in the FC Ports column, shown in Figure 2-62. Figure 2-62 Host Systems List with restricted host FC port PC1_2 Now we perform the configuration update, which makes the changes we made active.
Figure 2-63 Selected host FC port, restricted to one ESS FC port It shows us also that there are no ESS logical volumes currently assigned to the selected host FC port. This is indicated by the color of the disk group which indicates that no volume exists. This is going to change, as we are now prepared to add volumes. 2.5 Creating and assigning ESS logical volumes As mentioned before, adding volumes to the disk groups and assigning them to the newly defined host FC ports is done in one process.
This takes us to the Add Volumes (1 of 2) panel, as shown in Figure 2-65. Figure 2-65 Add Volumes (1 of 2) panel with host FC port PC1_1 selected Selecting host FC port for the volume assignment From here, we select the host FC port PC1_1 that we just have defined. We also have to select one of the ESS FC ports (Figure 2-66), before we can go to the next panel.
Selecting the disk groups for the volume placement By clicking on a port, we can now see the storage that we can use to define volumes on. This is shown in Figure 2-67. Figure 2-67 Add Volumes (1 of 2) panel with disk group status shown There are different colors to indicate which state the disk groups are in. In our case, none of the four disk groups that we configured contains an ESS logical volume. This is indicated by a green color.
2.5.2 Adding ESS logical volumes — step two Pressing the Next button, we are taken to the second step of defining the volumes. This is the Add Volumes (2 of 2) panel, shown in Figure 2-69. This is where we define how many volumes of what size we want to place sequentially on, or spread across all prior selected disk groups. Figure 2-69 Add Volumes (2 of 2) panel, two RAID 5 disk groups to be used At the top of the panel, we see the Available Free Space area.
Note: An ESS logical volume can be as big as we want it to be within the boundaries of one disk group. In our case, using 36.4 GB disks, we can have a maximum volume size of 210.45 GB. Looking at the Information window on the Storage Allocation — Graphical View panel we see 210.48 GB. Selecting the storage type To create RAID 5 volumes, we have to click on the RAID-5 Array row shown in Figure 2-70.
Instead, we use a volume size of 16.1 GB, which allows us to place 26 of them in our selected disk groups. To illustrate the process we only want to define two volumes here. We enter the number of volumes in the entry field, as shown in Figure 2-72. Figure 2-72 Volume Attributes frame, ESS logical volume size of 16.1 GB selected Adding volume definitions to the New Volumes frame We press the Add button to place the two volumes in the New Volumes frame.
Placing the volumes in selected disk groups Now, as we have defined how many volumes and of what size we want to add, we can also decide how the selected space should be used for the placement of the volumes. The two choices we have are shown in the Volume Placement frame. This is shown in Figure 2-74.
Figure 2-75 Add Volumes (2 of 2) panel, ready to perform the configuration update Again, we get a window warning us about a time intensive action, and we are asked if we want to continue with the update. This is shown in Figure 2-76. Figure 2-76 Warning window, time consuming action Chapter 2.
As with every run of a configuration update, there is a progress window, as shown in Figure 2-77. Figure 2-77 Progress window, volume creation After the update has finished successfully, we are presented with the message shown in Figure 2-78. Figure 2-78 Information window, successful volume creation Pressing OK leads us back to the Adding Volumes (1 of 2) panel, from where we can start the process again and define other volumes.
Viewing changes on the Graphical View panel To see the updates we have made, we press the Storage Allocation button. This takes us back to the Storage Allocation — Graphical View panel. There, we select the host FC port that we configured. This is shown in Figure 2-79. Figure 2-79 Host FC port selected This shows the volumes defined for our host FC port. The small purple parts of both disk arrays represent the space that is filled with volumes that are assigned to the selected host FC port.
Figure 2-80 Host FC port and ESS FC port selected The Information field now displays information relating to the configuration of the ESS FC port, which is currently undefined. This information is useful later to see which ESS FC port is configured for the appropriate topology. We show how to configure this in 2.7, “Configuring ESS FC ports” on page 158. After selecting one of the disk groups, the display changes again. Now the disk group has a thick border, and the Information frame has changed too.
The items that have changed on the disk group information since our last visit, are the Host Storage row and the Not Allocated row. Host Storage is the space that now has an ESS logical volume of 16.1GB assigned to the selected host FC port. Not Allocated means the space that can be used to define other ESS logical volumes on. Last but not least, we can also click on the disk array without selecting a host FC port first. The panel’s look changes again when only the disk group is selected.
Again, the Information frame changed the look to represent the general information of the disk group and is still not related to a specific host FC port. The used space now shows up in the Assigned row instead of the Host Storage row. This is shown in Figure 2-83. Figure 2-83 Information frame, disk group selected Now we want to go on and define more volumes and assign them to the host FC port PC1_2.
Figure 2-84 FC port PC1_1 with two assigned volumes Now when we select the host FC port PC1_1 in the Host Systems field, the associated volumes appear in the Assigned Volumes field with all the relevant details. For host FC port PC1_2, the Assigned Volumes table is still empty. Creating two more ESS logical volumes Now we do the same as we did before with the host FC port PC1_1, for PC1_2.
Viewing changes on the Graphical View panel Back on the Storage Allocation — Graphical View panel we see how the panel now looks using the View All Storage mode, shown in Figure 2-85. Figure 2-85 View All Storage mode The four areas in the four disk groups now represent the space that is used for ESS logical volumes. We cannot see to which host FC port they are assigned. To discover what volumes are assigned to PC1_2, we select this host FC port.
We select the host PC1_2, and we see that lines now appear connecting the host port, ESS FC port, and storage volumes. This is shown in Figure 2-86. Figure 2-86 Host FC port and ESS FC port selected Viewing changes on the Tabular View panel Now we will look at the Storage Allocation — Tabular View panel. We reach this from here using the Tabular View button on the top right of the panel.
Figure 2-87 Displays of all four volume assignments We see all the information about the location and the current assignments here, as well as the ESS logical volume ID. This is the ID that will help us later to identify the ESS logical volumes on our host system. Right now, there are not too many entries in the table. Later on, as we define more volumes, the table will become large. To make it easier to view the table, we can also perform a sort on the table.
For example, we can share volumes between host FC ports. These ports can be in the same host which uses, for example, the Subsystem Device Driver (SDD), to increase the bandwidth and availability, or in different hosts, for example, for clustering. Or, we might want to unassign volumes that were shared before, or we might want to unassign volumes to keep them unused for later reuse. We cover all this in the following topics. 2.6.
The Volume Assignments table provides a comprehensive overview of all configured volume associations. We can also sort the table, specifying the column which we want to use as sort criteria. The rows do not represent the volumes. Instead there will be a row for each association that a volume has as we will see later on. There is an Action field, which is grayed out, and a Target Hosts field, which is empty, when no volume association is selected. Now we show their interaction.
The second case is if the assignments to be selected are adjacent. Of course to avoid the application of the first selection method, we can sort the assignments to show up adjacent. We sort the assignment here by its host FC port. For adjacent assignments, we click on the first assignment to be selected, press and hold the Shift key and click on the last assignment to be selected. This selects all assignments in between.
Performing the configuration update To apply the changes, we press the Perform Configuration Update button. The progress bar indicates that the configuration update is taking place, as shown in Figure 2-92. Figure 2-92 Progress window, creating volume assignments If successful, we get the message shown in Figure 2-92.
If we press OK, we do not change back to the Open System Storage panel. Instead, we see the new volume assignments to the host FC port PC1_1, as shown in Figure 2-94. Figure 2-94 Assignments for the host FC port PC1_1 Viewing changes on the Graphical View panel Back on the Storage Allocation — Graphical View panel, we now see that the host FC port PC1_1 has volumes assigned to it on all four disk groups as shown in Figure 2-95. Chapter 2.
Figure 2-95 PC1_1 and ESS FC port selected In fact there is no volume configured on the ESS which is not assigned to this port. This is indicated by the lack of any colored space on the disk group, apart from the color of purple for the selected host FC port assigned volumes, or green for unused space. When a host FC port is selected, there is no indication that two of the volumes are shared between both the defined host FC ports. To see if ESS logical volumes are shared, we have to select a disk group.
Figure 2-96 Disk group containing shared ESS logical volume(s) selected The red color indicates that it contains volumes that are assigned to the host FC ports that show up highlighted yellow in the host interfaces row. When we compare this with Figure 2-95, we see that the amount of space used for this disk group is the same. Also, when we select the host FC port PC1_2, we will see the same amount of storage assigned to it, which tells us that the volumes must be shared between the two.
Figure 2-97 Tabular View, two volumes shared, two not shared Converting volume assignments table in HTML format To view the table offline, or to print it, this panel features the Print Table button. Then a window as shown in Figure 2-98 will appear, advising that the information is going to be displayed in another browser window. Figure 2-98 Information window, opening another browser window.
Now the content of the table is displayed in HTML format, as shown in Figure 2-99. Figure 2-99 HTML file, List of Assigned Volumes table, list of existing assignments It can now be printed from within the browser or it can also be saved as an HTML file for later viewing. 2.6.
Removing the assignments of one ESS logical volume First as we did in, “Selecting ESS logical volume assignments for modification” on page 144, we select the volume assignments for both host FC ports. Then, when the Action field becomes available, we select the radio button to Unassign selected volume(s) from target hosts, which means, remove selected ESS logical volume assignments for the selected host FC ports. The Target Hosts field now lists the host FC ports to which the assignments belong to.
Figure 2-103 Warning window two, volumes to be isolated from host FC ports We press the Yes button to proceed with the update and get a progress bar and a message window as shown in Figure 2-104 and, as we already know from the other actions we have done using this panel. Figure 2-104 Progress window, applying assignment changes In Figure 2-105 we are told that the volume assignments have been successfully removed.
Figure 2-106 Placeholder for unassigned ESS logical volume selected Viewing changes on the Open System Storage panel If we use the Back button of our browser we return to the Open System Storage panel. There we see that the volume is not displayed any more on any of our host FC ports. This is shown in Figure 2-107 for the host FC port PC1_2, which now has only one volume assigned to it.
Viewing changes on the Graphical View panel Now we want to see the changes on the Storage Allocation — Graphical View panel. Figure 2-108 and Figure 2-109 show the selected host FC ports PC1_1 and PC1_2 with the disk groups that contain volumes that are assigned to it. Figure 2-108 No volume assignment to PC1_1 in disk group 2 of cluster 1 loop B Figure 2-109 No volume assignment to PC1_2 in disk group 2 of cluster 1 loop B Chapter 2.
The space of the disk groups that contains ESS logical volumes that are assigned to the selected host FC port show up in purple, and unused space in green. Both host FC ports have no ESS logical volumes assigned to them in the disk group in cluster 1, adapter pair four, loop B. This is indicated by the absence of colors. However, on the particular disk group space is allocated for volumes. This means that there is an existing volume (or more), which is not yet assigned to any host FC port.
Another way to find out about the state of the space in a disk group is to select it. In our example, shown in Figure 2-111, we select the disk group in which we have an unassigned volume. Figure 2-111 Disk group selected, unassigned volume(s) in selected disk group Again, the space used for unassigned volumes is displayed in yellow. Also, we see in the host interfaces row the interfaces that have volumes assigned to them in the selected disk group, which is empty.
Figure 2-112 Tabular View, four volume assignments, three volumes Here we see that there is no longer any assignment for the volume we unassigned. Unassigned volumes only show up in the Volume Assignments table in the Modify Volume Assignments panel. Now, as we have ESS logical volumes assigned to host FC ports, and we have also made adjustments to the configuration, there is only one step left to be performed. 2.
As we already know, the selection of an entity in the Storage Allocation — Graphical View panel changes the information in the Information frame of the panel. Clicking on an icon of an ESS FC port in the ESS interfaces row displays information about the adapter. There we see the physical location in the ESS, the topology the port is configured for and the WWPN. For any ESS FC port that we have not configured, the window looks like that shown in Figure 2-113.
Selecting the interface port to be configured The Configure Host Adapter Ports panel is shown in Figure 2-115. Figure 2-115 Configure Host Adapter Ports panel, ESCON port selected From this panel, we select the port of the ESS interface adapters to be configured, depending on the type of the adapter. We can select the ports by clicking on the icons of the adapter (left or right on any adapter with two ports), or by selecting the port from the drop down list as shown in Figure 2-116.
Figure 2-116 Port selection drop down list, SCSI port selected There is no configuration necessary for ESCON. There are different configuration options for SCSI and Fibre Channel. However, we will focus on the FC part. By selecting an FC port, we are presented with the panel, as shown in Figure 2-117. Chapter 2.
Figure 2-117 Configure Host Adapter Ports panel, FC port selected The port we used is in bay 1, adapter 3, and is port A. It is called ‘A’ even if it is the only port on the adapter. Understanding the Storage Server Attributes field As shown in Figure 2-118, the Storage Server Attributes field — also known as the logical unit number (LUN) access mode — which specifies the channel access mode, can only be changed by an IBM SSR using the service terminal.
There are two Fibre Channel Access Modes to which the ESS can be configured: Access_Any In Access_Any mode, any host FC port that is not defined in the ESS Specialist, can access all non-AS/400, ESS logical volumes. In Access-Any mode, the ESS Specialist displays an anonymous pseudo-host FC port icon, which is shown in Figure 2-119 and which is taken from another ESS, because the one we used here is configured for Access_Restricted.
Adding new ESS logical volumes to the ESS, the new volumes go to the host FC port that is selected. Assigning ESS logical volumes to host FC ports, means adding them to the access profile of the selected host FC port and to pre-existing ESS logical volumes. Removing a host FC port from the ESS, means deleting the host FC port definition and its access profile. 2.7.
These actions are host dependent. With respect to the fabric components, if WWPN zoning has been implemented, then it will be necessary to update every instance of the WWPN in the switches and/or directors. The WWPNs of the adapter are built from the WWNN of the ESS and the location of the adapter in the ESS interface bays. This is done using the pattern shown in Figure 2-120.
The Information frame on the Storage Allocation — Graphical View panel reflects the new WWPNs of the ESS we have used for this illustration, which is shown in Figure 2-121. Figure 2-121 Information Frame, reflecting the new locally administered WWPN In Figure 2-122 we show the WWNN of the ESS that is used and modified when this feature is enabled.
2.7.3 Configuring host FC ports for FC-SW and FC-AL Now, as we have selected the first ESS FC port, we are going to configure this and two other ports to be able to communicate with our host FC ports. The first two ports we configure for use in a switched fabric environment and the second are for use in an arbitrated loop environment.
As we have selected the topology for the selected ESS FC port, we see a small red line directly below the icon of the just-defined ESS FC port, as shown in Figure 2-124. Figure 2-124 ESS FC port for Point to Point defined This indicates that the port definition has been changed. We can now go on and select and define the other ports. The red lines will remind us where we made definition changes. Now we go on and define port three in bay four for FC-SW and adapter three in bay two for FC-AL.
Figure 2-125 ESS FC port for Arbitrated Loop defined Resetting changes made to a selected ESS FC port Using the Reset Selected Port button on the upper right of the screen, we can reset the definition for the selected port to the status the port was in before we made any changes. This brings up a warning window as shown in Figure 2-126, stating that the configuration changes made are going to be discarded.
Performing the configuration update Click on the Perform Configuration Update button. During the process of performing the update we get a progress window as shown in Figure 2-127. Figure 2-127 Progress window, changing the topology Figure 2-128 shows a successful topology change of the ports. Figure 2-128 Information window, successful topology change Pressing the OK button takes us back to the Open System Storage panel.
For port three in bay one, which is now configured for use with FC Point to Point, this is shown in Figure 2-129. For port three in bay two, which is now configured for use with FC Arbitrated Loop, this is shown in Figure 2-130. Figure 2-129 Information frame, ESS FC port topology Point to Point Figure 2-130 Information frame, ESS FC port topology Arbitrated Loop We have now prepared the ESS to provide storage to our hosts.
Figure 2-131 Storage allocation now completed We see that the host FC port PC1_2 is restricted to see the one volume it has left through one FC port which is configured for point to point. Related information: For more information on the ESS, see these references: IBM Enterprise Storage Server Introduction and Planning Guide, 2105 Models E10, E20, F10 and F20, GC26-7294 Introduces the product and lists the features you can order.
IBM Enterprise Storage Server SCSI Command Reference, 2105 Models E10, E20, F10 and F20, SC26-7297 Describes the functions of the ESS. It provides reference information for UNIX and AS/400 hosts, such as channel commands, sense bytes, and error recovery procedures.
174 Implementing an Open IBM SAN
3 Chapter 3. Implementing the INRANGE FC/9000 Fibre Channel Director The IBM machine type 2042 model number 001, INRANGE FC/9000 Fibre Channel Director, is the core product of a reseller agreement between IBM and INRANGE Technologies, which adds the INRANGE FC/9000 Fibre Channel Director to IBM’s growing list of enterprise-class SAN fabric offerings. © Copyright IBM Corp.
3.1 Introduction to the INRANGE FC/9000 director In this chapter we introduce the INRANGE FC/9000 Director and cover the following topics: Product architecture Zoning methodologies Management features Supported topologies Supported servers Supported devices In Figure 3-1 we show the front view of an FC/9000.
3.1.1 Product overview The FC/9000 Fibre Channel Director is INRANGE’s flagship SAN product.
3.1.2 Currently supported zoning methodologies Zoning is one of the key functions provided by SAN switches or directors. Basically it provides methods to limit possible communication between ports. This can be used to subdivide a physical SAN into separated logical ones. Zoning can be seen as an important security feature of SAN products.
Figure 3-2 shows an example of a director view with IN-VSN Manager. Figure 3-2 Example of IN-VSN management view The most commonly used interface with the INRANGE FC/9000 director is the IN-VSN software tool. Therefore, in the following topics we will focus on how to use IN-VSN. 3.1.4 Supported protocols In most environments, a homogeneous landscape of servers and storage is hard to find.
INRANGE does just that, by supporting these protocols: Open Systems: – Arbitrated Loop • Public loop • Private loop – FC-SW (Fibre Channel Switched Fabric) S/390 systems (zSeries): – FICON All these can attached to a single INRANGE director at the same time. Support of cascading To create even larger fabrics FC/9000 directors can be cascaded. By doing this you can create fabrics with more than a thousand external ports. Today, IBM supports fabrics with up to 8 cascaded INRANGE FC/9000 directors.
3.1.
In Figure 3-3 we show our target ITSO environment. Hewlett Packard servers running HP-UX 11.0 HP L Class 2000 HP L Class 2000 HP D Class 280 64 HP K Class 360 Inrange FC/9000 ShortWave fibre optics (50micron) ESS 2105 - F20 Figure 3-3 Initial INRANGE environment For management purposes we have one dedicated PC server attached to the director. This management station runs the server portion of the INRANGE IN-VSN Manager to control all fabrics. 3.2.
3.2.3 Initial setup of INRANGE FC/9000 IP settings The INRANGE Director is delivered with the current supported level of firmware and a default TCP/IP setting of 10.1.1.51 and 10.1.1.52, and a Subnet mask setting of 255.255.255.0. There is also a default chassis ID, switch ID and fabric ID. Without doing any kind of cascading, there is no need to change these IDs.
Leaving all IP settings as a private network For using the IN-VSN software from this local private network, the supplied 3Com/US Robotics hub is sufficient enough. To enable all IP based components to talk together, plug in the Management PC and both INRANGE IP ports to this 3Com Hub. So initially, our network setup looks like that shown in Figure 3-4.
After attaching our INRANGE setup to IBM’s corporate network, as shown in Figure 3-5, we are now able to access the IN-VSN software from wherever we are in the corporate network. Management PC Ethernet-Hub FC/9000 Corporate network Figure 3-5 INRANGE setup attached to a corporate network To gain actual access to the IN-VSN management tool, you will need to install an IN-VSN client that can then communicate with the IN-VSN server over the corporate network.
Referring to Figure 3-6, we see that now only the Management PC can access the directors’s IP ports. other directors ITSO director Ethernet Ethernet-Hub IN-VSN server on Management PC (with two network interfaces) Corporate network ITSO Workstation with IN-VSN client ITSO Workstation with IN-VSN client Figure 3-6 INRANGE setup with secure director access Direct IP access from the corporate network to directors is now impossible. The only way to gain access is using the IN-VSN server.
3.2.5 Installing the IN-VSN Enterprise Manager software There are a number of different ways in which the IN-VSN software can be delivered. If software feature code 7600 is configured, the workstation pack will include a PC, monitor, Ethernet Hub, and Modems. The management software comes pre-installed on the PC. If you require more than one version of the IN-VSN client code, this is achieved by using feature code 7201.
Running the setup procedures of IN-VSN server and client Both the IN-VSN server and client require the following two programs to be started from the installation CD: Use j2re1_3_0-win.exe to install the appropriate Java programs. Use setup.exe to actually install the IN-VSN server and/or the IN-VSN client software. For systems on which the IN-VSN server should run, an additional odbc driver has to be installed: Use odbcinst.exe to accomplish this.
The setup program then checks whether you have the appropriate level of Java installed on this system or not, as shown in Figure 3-8. Figure 3-8 Running the IN-VSN setup: Java VM search Click I Agree and Next in the Licence Agreement Window. Chapter 3.
Then you must decide which part of the IN-VSN software should be installed, server and/or client, as shown in Figure 3-9. Figure 3-9 Running the IN-VSN setup: Feature selection After selecting the features that you wish to install, you may either accept the default installation path or enter a new one. The default installation path for both client and server is C:\INRANGE\Fc9000. The setup program automatically creates the folder if you agree to its creation.
Before actually starting the installation a list of all installation parts is shown in a list with its needed disk space. Click Install Now to accept these settings, as shown in Figure 3-10. Figure 3-10 Running the IN-VSN setup: Verifying all packages to install After installing these IN-VSN packages, you are given an installation summary which tells you that all parts were installed successfully.
Leave all components selected in the odbc installation window, as shown in Figure 3-11. Figure 3-11 Running obdc 3.2.6 Accessing the management tool Make sure that the IN-VSN server is running either on the local machine or elsewhere in the network. Since the IN-VSN management tool is not browser based, it is necessary to have the client piece of software installed on your workstation.
We installed the IN-VSN client software on two different workstations in our target environment, as shown in Figure 3-12. Management PC with IN-VSN server Ethernet-Hub FC/9000 Corporate network ITSO Workstation with IN-VSN client ITSO Workstation with IN-VSN client Figure 3-12 Final IN-VSN software setup This layout makes it possible to have multiple users logged in to the IN-VSN software at the same time. IN-VSN clients do not talk directly to the FC/9000.
Once the IN-VSN server is up and running, an IN-VSN server window should appear with server running status, as shown in Figure 3-13. Figure 3-13 IN-VSN server up and running Another indicator of the running status is the permanent change of colors of the text in the middle of the IN-VSN server window. Once the IN-VSN server is running, you can then start multiple IN-VSN client sessions pointing to the servers IP address or DNS name.
After starting the client, a login window will appear. Enter user name (default admin) and password (default admin) as shown in Figure 3-14. Figure 3-14 Starting an IN-VSN client session Additionally, you will have to enter the IN-VSN server destination. This is where the client software should find the IN-VSN server running. If the server is running on a different machine, then enter the IP address of the IN-VSN server. If both pieces of software run on the same machine, just select localhost.
When logged into the IN-VSN server as the first client ever used, then all settings are still default values and no specific fabric or director information is available at all. This is shown in Figure 3-15. Figure 3-15 Initial screen of an IN-VSN client session Once you are logged in to an IN-VSN client session, you can then customize and set up your fabric layout.
3.2.7 Defining users Different user-levels are provided by INRANGE. You can have multiple users with different levels defined at a time. Default user names are the same as their level, and are as shown in Table 3-1.
Adding users To enter the user definition screen, click the Users tab as shown in Figure 3-16. Figure 3-16 Adding an IN-VSN user: Re-enter password For security reasons, you will have to re-enter your password. After re-entering the password, the current user definitions are displayed.
Click Add to enter the information for a new user as shown in Figure 3-17. Figure 3-17 Adding an IN-VSN user: Overview After clicking Add, the Details fields become active and you are able to enter the data needed to create a new user. You can give all rights to this new user apart from Maintenance level (maint). Chapter 3.
In Figure 3-18 we create a new user called itso_1 with admin rights. Figure 3-18 Adding an IN-VSN user: Creating an admin user Click Accept to add this newly created user into the list of already known users.
However, this new user is not saved until you click Save, as shown in Figure 3-19. Figure 3-19 Adding an IN-VSN user: Saving new user configuration You must take into account that the default users will not be deleted after adding new users. This may be different to other products that you are familiar with. To delete the default users, you can use the same procedure as for all other users. How to delete users is described in the topic that follows. Chapter 3.
Deleting users Deleting users is done using the same panel as we used for adding users. Note that you cannot delete the user that you are logged in with. To remove any user, select the victim in the list of already defined and saved users. Click Delete and confirm this action as described in Figure 3-20.
Once you have confirmed this deletion, that particular user is deleted in the IN-VSN database of the IN-VSN server. It is also removed immediately from the user list of the IN-VSN client as shown in Figure 3-21. Figure 3-21 Deleting an IN-VSN user: Updated user list Changing user definitions Once you have added users, you can change their attributes. All attributes, excepting the name, can be changed. These are: Password User rights Description Chapter 3.
To change a user, first select it in the Users list. Once you have selected a user, you are automatically in change mode. Change the desired settings and click Accept as shown in Figure 3-22. Figure 3-22 Change attributes of an existing IN-VSN user We have changed the rights of user itso_2 from admin to operator. After accepting the changes, this modified user will appear with updated status. You can modify multiple users before actually making these changes effective by clicking Save.
3.2.8 Security considerations As a SAN may store a great deal of sensitive and often confidential corporate information, it might be interesting to consider some security issues, such as: User management Consequences of zoning methods Location access We will describe these issues as they relate to the usage of the INRANGE FC/9000 director. User management As described in 3.2.7, “Defining users” on page 197 INRANGE provides different types of users.
Zoning methods INRANGE provides both these types of zoning: Hard zoning Name server zoning (often called soft zoning) Hard zoning makes it impossible to have I/O between different hard zones, whereas with soft zoning, one port could easily be assigned to any other port in the fabric. However, both kinds of zoning are port related. It is the director’s port that is referred to in both methods. At this time, a WWPN based zoning referring to the attached server or storage ports is not available.
When leaving the fabric without any name server zone, the only effective zone is the all-inclusive hard zone. That means all ports are enabled to communicate with every other port. Be aware of this any-to-any connectivity when leaving the fabric unzoned. Once you have added one name server zone, then the nodes, if they are part of any zone, have limited access to the name server table. Name server zone members can communicate with only those nodes that are part of the same name server zone. 3.
We will first perform a manual backup. Therefore, we use the IN-VSN server window to open the backup menu, as shown in Figure 3-23. Figure 3-23 Manual backup of IN-VSN server: Step 1 Then we specify the name of our IN-VSN database file. We could leave the default name had we wanted to. However, we choose to have our own specific file named itso_db2.mdb, as shown in Figure 3-24.
The database file created by this manual backup could be used for IN-VSN database restore. However, in real life it may be difficult to perform manual backups periodically, since you need access to the IN-VSN server interface. Therefore, it is possible to set up an automatic IN-VSN database backup. We recommend that you utilize this function, as it saves time, increases our recoverability, and does not rely on human memory to start the process.
Figure 3-27 AutoBackUp settings: Specifying Filename and backup interval The IN-VSN server main window then displays “Auto Backup Enabled” status. Setting up page home/call home In some environments it may be useful to have INRANGE or IBM automatically contacted via telephone or pager by the IN-VSN server. This is called Page Home or Inventory Update Call. However these parameters should be set up by IBM or INRANGE Customer Engineers only.
3.3.2 Connecting to an INRANGE fabric A fabric is made up of one or multiple INRANGE directors and or switches which are linked via Inter Switch Links (ISLs). Directors that are not linked via ISL are considered as different fabrics. Initially, the IN-VSN software is not aware of any fabric components at all, as shown in Figure 3-29 Figure 3-29 IN-VSN fabric tool without any fabric known to it To make a specific fabric known to the IN-VSN software, click Add.
Figure 3-30 IN-VSN: Connecting to a new fabric It is not necessary that the directors specified with their IP addresses be actually online or connected at the time of the fabric creation. The actual IP connectivity to these directors is obtained every time you click the specific Fabric tab in the IN-VSN client interface. Furthermore, at the time of fabric creation, it is not checked if the specified directors are actually connected via ISL or not.
Figure 3-31 IN-VSN: Initial fabric view The default name of this director is FC64. If you had multiple directors in one fabric, they would appear as multiple symbols in the fabric view. 3.3.3 Assigning names and aliases Our fabric consists of one director with 64 ports. We want to attach four Hewlett Packard (HP) servers and one Enterprise Storage Server (ESS) to this director. Using the IN-VSN tool, we are able to assign names to physical director ports and to the director itself.
To simplify further usage of this environment, we decided to use the names as shown in Table 3-2.
Assigning the directors name To assign the directors name we enter the director view by clicking the specific director in the Navigation explorer of the IN-VSN tool as shown in Figure 3-32. Figure 3-32 IN-VSN: Director view We enter the new name osvl_fc64 in the Name field in the director view. Click Apply and confirm this action to make the change effective as shown in Figure 3-33. Figure 3-33 IN-VSN: Changing the directors name Chapter 3.
Assigning port names To assign names to individual FC ports of our director, we click the specific director in the IN-VSN navigation tree to have the director view displayed. Then we click the Ports tab to have a list of all 64 ports. This is shown in Figure 3-34.
Figure 3-35 IN-VSN: Changing port names We start with the two ports used for ESS attachment. Therefore, we change the name of port 40 to ospl5b1p4 and port 48 to ospl5b4p4. Click Apply and confirm this action to make the new names effective. Note: Take into account that all the port names you have assigned refer to the physical port of an INRANGE director. These names are technically independent of any attached external ports.
We have now physically attached all server ports and the two ESS ports. That is why we now have Device information (WWNN) displayed in the port list. Figure 3-36 shows a port list with all of our desired name changes active and with the WWNN information displayed for the attached ports. Figure 3-36 IN-VSN: Director’s updated port view Once the new names are effective they will be used throughout the whole IN-VSN fabric management, including zoning, FIO-blade monitoring and name services. 3.3.
To set the director clock, just click the specific director in the navigation tree and then choose Set Director Clock from the Director menu, as illustrated in Figure 3-37. Figure 3-37 IN-VSN: Selecting the director clock menu Chapter 3.
Enter your desired time settings and apply this by clicking OK as shown in Figure 3-38. Figure 3-38 IN-VSN: Setting the director clock 3.3.5 Attaching loop ports Today, the storage and server industry is moving rapidly towards switched fabrics. However, there are still a lot of systems that use Loop protocol. For instance, most tape devices use FC-AL, as well as lots of legacy FC host adapters. Of course, some of them could be upgraded or replaced by newer devices to support native fabric login.
Note: To read the following topics, it is useful to understand the differences between terms like director port, loop port, loop devices, initiators, targets: A director port is an actual physical port of the FC/9000 director. Loop ports are the external ports attached to a director port. Loop ports use loop protocols like private loop or public loop. They are sometimes referred as loop devices or loop nodes. Initiators are ports that actually control the IO flow and give commands.
Figure 3-39 IN-VSN: Port View to enable port specific loop attachment Once the Auto Sense field is activated in the port view, click Apply and confirm this action to make this change effective, as shown in Figure 3-40.
Figure 3-40 IN-VSN: Confirming the enabling of loop attachment Now we are able to use loop attachments on this port. In our case, we have a Hewlett Packard Tachyon Lite Adapter set to Public Loop. The actual adapter settings are displayed in Figure 3-41. Figure 3-41 IN-VSN: FC adapter set to public loop Chapter 3.
This adapter is set to Public Loop as shown in the Topology row. By selecting the Loop Devices tab in the IN-VSN port view, we can ensure that the director actually recognizes this loop attachment as shown in Figure 3-42. Figure 3-42 IN-VSN: Port Loop Devices View Enabling loop attachment for the entire director Instead of individually enabling loop attachment for lots of specific ports, you can enable this for the entire director as well.
To enable all ports to attach loop devices, select the specific director in the navigation tree of IN-VSN and click Auto Sense Arbitrated Loop Enable in the Director menu, as shown in Figure 3-43. Figure 3-43 IN-VSN: Enabling Auto Sense Arbitrated Loop for the entire director Be aware that this change could interrupt in-progress I/O on ports. Therefore, use this procedure only if you are sure about the consequences of the warning given in the confirmation window, as illustrated in Figure 3-44.
We show a representation of this in Figure 3-45. Figure 3-45 IN-VSN: loop port in name server table as NL_Port Port hewlett_td0 was set to enable AutoSense AL and we actually attached this port physically. After logging in, this port is displayed as an NL_Port, since this is a public loop port. Bypassing loop devices In cases where you have multiple loop devices attached to one director port, you can specify which devices should be actually used in the fabric.
In our example, we have four loop devices attached to one director port. This is shown in Figure 3-46. Actually attached Loop Devices device no: 1 Not in Bypass device no: 2 Not in Bypass device no: 3 Bypassed device no: 4 Not in Bypass Loop enabled director port Loop Devices made available to the fabric Figure 3-46 IN-VSN: Bypassing loop devices We have decided that device 3 should not be seen in the fabric. To achieve this kind of filtering, device 3 must be disabled and thereby set to Bypassed.
To enable a particular device, use the same menu and click Enable as shown in Figure 3-47. Figure 3-47 IN-VSN: Enabling and Disabling Loop devices You should have at least one device per port left that is Not in Bypass. By default, all attached loop devices are set to Not in Bypass. Note: Not all loop devices support Bypass. In such cases, even after clicking the Disable button, they will remain in Not in Bypass mode. That means such devices are always enabled.
Private loop Private loop is the most basic form of Fibre Channel. Private loops are generally built using hubs, and communication is handled by passing messages around the loop from one device to the next in order to gain control of the loop and to communicate on it. Because each device is dependent upon the device next to it for communication, private loops tend to be unstable. For example, adding, removing, or even rebooting a device causes other devices on the loop to pause.
Usage of public loop and private loop Since these two loop modes inherit different protocol methods, there are differences in the actual usage of them with the INRANGE IN-VSN tool. Usage of public loop You can attach public loop initiators or targets without any additional settings, provided that the director port is set to enable AL. However, attaching a public loop port does not automatically mean that this port can talk to any other ports in the fabric.
Usage of private loop ports Normally, private loop ports are not able to talk with other ports outside of their own loop. Therefore, their addressing is different from normal ports. However, with the Translative Loop mode of specific director ports, it is possible to let them talk with other ports in the fabric. To enable this you have to set the Admin Type of this particular director port to TL as shown in Figure 3-49.
The creation and maintenance of this list is different between private targets and private initiators. Therefore, you have to specify whether a particular TL_Port has private targets or a private initiator attached to it. Click the TL-CFG button to open the TL Config window as shown in Figure 3-50.
Note: The selection of Private Target or Private Initiator for a TL_Port does not necessarily mean that you have such types attached to it. The private target option is only useful if a private initiator should access this particular target port. If a public initiator should get access to this private target you should actually set this private target to Private Initiator. Then you should proceed to specify the WWN of the Public initiator that should have explicit access to this private target.
Figure 3-51 IN-VSN: adding possible targets If you want to add device addresses that are currently not attached to the fabric, then you can add them by typing in their WWN. Click the Add button and enter the WWN as shown in Figure 3-52.
Each TL_Port has its own translation entries list, as shown in Figure 3-53.
Note: Even if the translation entries table for private target ports is updated automatically, you will not see its content. The auto learning feature is always used in the background. You cannot deactivate it or activate it. TL_Ports and zoning Once the translation entries list is created, these TL_Ports can actually talk to other ports in the fabric.
However, since INRANGE is using a Translation Entries List, these LIPs will not be propagated to other fabric members. This is true even if multiple private loop ports are zoned together. Therefore, all LIP impact is limited only to the external physical loop (for example, an FC_AL hub). This feature is a major benefit of using TL_Ports. 3.4 INRANGE zoning One of the basic purposes of SAN fabric products is to enable or disable communication between the different ports (devices) attached to them.
By no means is it possible to have communication over the boundaries of hard zones. This is also true if malfunctioning fabric initiators try to get around the name server tables. Hard zones take precedence over all other kinds of zoning (for example, over broadcast and name server zoning). If no hard zone is enabled at all then all ports are considered as being part of one large default hard zone. Name server zones and broadcast zones can be implemented within hard zones.
FIO 1 FIO 2 FIO 3 FIO 4 FIO 5 FIO 6 FIO 7 FIO 8 group of 4 director ports Inrange FC/9000 Director Figure 3-54 INRANGE hard zoning: fixed location of port groups The two ports of one FIO blade being part of such a group are called adjacent ports. The entire director is always automatically segmented into such groups. Consequently, with a fully equipped 64 port director, 16 port groups are automatically defined. Chapter 3.
Building hard zones using port groups Now we know that a hard zone must consist of port groups. In our example, we have created two hard zones, as illustrated in Figure 3-55. Figure 3-55 INRANGE hard zoning: basic example with 2 zones This picture was taken by using the IN-VSN tool. Hard Zone A consists of only one port group. Hard Zone B consists of the remaining 15 port groups. Remember the “all-or-nothing” rule: When implementing hard zoning, all ports must be zoned.
When creating hard zones, all port groups in a particular hard zone must be adjoining. In Figure 3-56 we show an example of an incorrect implementation. Figure 3-56 Violating the adjoining rule The upper port group of Hard Zone A has no adjoining contact to the remaining two port groups of Hard Zone A. This is a violation of the hard zoning rules, and therefore it is not possible to create a hard zone layout such as this. 3.4.
3.4.3 Understanding INRANGE name server zoning In contrast to hard zoning, there is no actual physical segmentation of fabrics with name server zoning. Instead, a name server table is used to implement that type of zoning. Name server tables store information about nodes that have logged into the fabric. Name server zones restrict the access of affected ports to this information. All unzoned nodes have full access to the name server table.
3.4.4 Hard zones and name server zones together Hard zoning can strictly separate port groups, and this can be seen as an effective security feature. Name server zoning allows us to further define the communication control on a per port granularity. When name server zones and hard zones are used in conjunction, we need to consider the following principles: If you have no hard zone(s) created at all, then all director ports are considered as being part of one big default hard zone.
The following rules are enforced with INRANGE directors: Both TL-CFG lists as well as name server zones cannot span hard zone boundaries. You can use name server zoning to further limit access between nodes TL-CFG lists are for TL_Ports only. They overrule the name server principles. So, TL_Port attached nodes can only communicate with ports specified in the TL-CFG lists. This hierarchy is shown in Figure 3-57.
The scenario illustrated in Figure 3-58 describes a possible layout. All director ports beeing part of any hard zone Hard Zone A Hard Zone B Name Server Zone A_3 Name Server Zone A_1 Name Server Zone A_2 Figure 3-58 INRANGE fabric scenario with hard zoning and name server zoning In this layout: All three defined name server zones are within Hard Zone A. They cannot span over to Hard Zone B. Name server zone A_1 and A_2 are overlapping. Both have an ESS node as a member.
3.4.5 Defining hard zoning In our example shown in Figure 3-59, we want to separate eight director ports for usage by the finance department only. In no way should it be possible for anyone else to gain access to their ports. That is why these eight director ports will be put into one dedicated hard zone (the Finance hard zone). Finance dept. IT systems to be attached to the director hardzone of 8 ports that are dedicated for the Finance dept.
To start our hard zoning setup, we first select the particular director in the navigation tree and then click the Hard Zoning tab as shown in Figure 3-60. Figure 3-60 IN-VSN: Selecting hard zoning in the director view This window is used for all hard zoning configurations related to the selected director.
Figure 3-61 IN-VSN: Specifying a name for a hard zone We have chosen to name the first zone Finance. However, later on, we could change the name again. The remaining ports will be placed in a second zone, which we will call Other. Note: The used color or number for a dedicated zone does not have any effect on the actual behavior or performance. These numbers and colors are solely used to simplify the organization and usage of hard zoning.
We have selected the two upper leftmost port groups to be part of hard zone Finance as illustrated in Figure 3-62. Figure 3-62 IN-VSN: two ports assigned to a hard zone To test the integrity of our zone layout, we click the Test button. As you can see, we got a Test Failed feedback. This is because we violated a hard zoning rule: Either all ports of a fabric belong to hard zones or none of them are hard zoned.
Again we test this layout by clicking the Test button. This time everything looks fine, as illustrated in Figure 3-63. Figure 3-63 IN-VSN: Having two hard zones defined We now have two zones defined: Finance and Other. The Finance zone consists of two port groups giving a total port count of eight. The Other zone has the remaining 14 port groups assigned to it having a total port count of 56. However, this hard zone setup is not yet active since we have not clicked the Apply button at this stage.
As illustrated in Figure 3-64, when clicking the Test button now we get a violation message, as expected. Figure 3-64 Violation of ports To successfully finish our hard zone setup as planned, we have put the bottom left port group back into the Other zone. Chapter 3.
To activate our hard zone layout, we click Apply as illustrated in Figure 3-65. Figure 3-65 IN-VSN: Applying a hard zone setup As explained earlier, a director reconfiguration task such as this will affect in-progress I/O. A warning message is issued to indicate this. To activate our setup, we click Yes. After this hard zone setup is activated, there is no communication possible between the Finance zone and the Other zone. This is the most secure way to separate nodes in a fabric environment. 3.4.
All primary FC ports of these servers will be zoned to ESS Bay 1, Port 4, and all secondary FC ports of these servers will be zoned to ESS Bay 4, Port 4. We use pvlinks as multipathing software. Another possibility is to use the IBM Subsystem Device Driver (SDD). In Figure 3-66 we show an overview of what we will achieve. Hewlett Packard servers running HP-UX 11.
By implementing two zones, we limit the access of a particular host FC port to only one ESS FC port. We do this because we like to have only two paths to a ESS logical volume. These two paths are created by giving the two host FC ports of one server access to the particular ESS logical volume using ESS LUN masking. In Figure 3-67 we show how these ports are physically attached to the INRANGE director.
We shall now explain how our zoning will look logically. As illustrated in Figure 3-68, we plan to have two zones.
By implementing such a layout, each server is attached to two zones. Since each zone has a different ESS port as a member, the server will get exactly two paths to the ESS. Use of pvlinks and ESS based LUN masking will enable sufficient multipathing. To actually implement name server zoning, click the specific fabric icon, in our case osvl_fc64, in the navigation tree, and then select the Zoning tab as shown in Figure 3-69.
Leave the Name Server Zone field selected. Click Add to create a new zone. Enter zone number, zone name, and a description as illustrated in Figure 3-70. Figure 3-70 IN-VSN: Entering number and name for a new zone Chapter 3.
From the Ports scroll list, select the ones you want to be a member of that zone as shown in Figure 3-71. Figure 3-71 IN-VSN: Selecting the members of a new zone We added these ports to the zone ospl5_b1_zone: osplsun2_tdo osplhp1_a ospl5b1p4 hewlett_tdo The effect is that all these server ports will have access only to the ESS port Bay1Port4.
After putting all needed ports in the Ports in Zone list, we continue with clicking Accept. This is shown in Figure 3-72. Figure 3-72 IN-VSN: Accepting settings for first new zone Chapter 3.
However, now we are getting an error message, since our previously chosen zone number of 0 is not valid. This is shown in Figure 3-73. Figure 3-73 IN-VSN: Zone number 0 is not allowed to use So we just change this zone number to 1, and click Accept again. Now this action is accepted. This zone is now listed in the zone list with a status of Added.
To make this zone effective, click Save as shown in Figure 3-74. Figure 3-74 IN-VSN: Saving a newly added zone Using the same procedure, we now add a second zone called ospl5b4_zone which has these ports as members: ospl5b4p1 osplsun2_td1 hewlett_td1 geode_td1 osplhp1_b Chapter 3.
After saving this second zone, the zone list contains both zones with a status of Saved as illustrated in Figure 3-75. Figure 3-75 IN-VSN: Zone list with both zones added and saved Another way to verify the right zone settings is to have a look at the Ports View of the director.
To do this, select the specific director in the navigation tree and then click the Ports tab. After scrolling down we see the used ports with our given names, the WWN of the attached ports, and then the zone these ports belong to as shown in Figure 3-76. Figure 3-76 IN-VSN: Director Port View showing the port to zone relationship We now have two zones defined in a fabric consisting of one director.
The motivation to use cascaded directors is not limited to increased port counts. Cascaded directors can be used to create a SAN that spans large distances. Unique to INRANGE Directors, ports used for Inter Switch Links are called T_Ports (Trunk ports). Every director port can be used for T_Port operation. No manual setting is needed since ISL connections are automatically detected and the affected ports are set to T_Port. ISLs can be run on either ShortWave GBICs or LongWave GBICs.
– Each director or switch in a fabric must have a unique switch id. – When having no zoning implemented at all then all nodes have access to any other node fabric-wide. When actually using Inter Switch Links for I/O, there will be a delay called latency. The actual usage of a ISL is commonly called a hop. Typically up to three hops are supported for a particular I/O. T_Ports cannot be part of name server zones.
3.6 Monitoring and maintenance of an INRANGE SAN This topic will briefly cover these topics: Communication protocols to manage an INRANGE SAN Microcode load information Monitoring user activities Using the INRANGE event log 3.6.
Call home functionality Call home enables the IN-VSN software to propagate event information and configurations to INRANGE or IBM systems. This allows better response times in case of component failures. Call home is disabled by default but can be enabled by Customer Engineers. A modem connection for the Management PC is required. Changing the IP address The IP settings of INRANGE directors should be done by IBM or INRANGE Customer Engineers.
All activities are logged and categorized into different types. Following are some examples of these operation types: User login IP address of user login Fabric definition User definitions Name server zoning Switch name changed To look at the audit trail, click AuditTrail in the navigation tree as illustrated in Figure 3-79.
3.6.4 Using the IN-VSN event log The event log contains all important events that have occurred. This includes events triggered by users and events caused by other external or internal influences, such as FRU failures or losing power. To read the event log, click EventLog in the navigation tree of IN-VSN, as shown in Figure 3-80. Figure 3-80 IN-VSN: Accessing the INRANGE event log This event log is valid for all fabrics managed by this IN-VSN server.
3.7 ESS configuration guidelines In this chapter we have addressed some of more commonly encountered features of the INRANGE director. The 2042-001 is designed to provide maximum bandwidth switching from any port in the system to any other port in the system. There is no mandatory requirement for attachment of server or device ports to the 2042-001.
4 Chapter 4. Implementing the SAN Data Gateway In this chapter, we describe the steps involved in planning and implementing the IBM Storage Area Network Data Gateway, 2108-G07. The SAN Data Gateway is a hardware solution to allow connection of Fibre Channel ready host systems to attach to SCSI storage systems. © Copyright IBM Corp.
4.1 SAN Data Gateway The IBM Storage Area Network Data Gateway is an essential component of the SAN infrastructure. It provides several benefits to bridge the legacy gap as storage products migrate from SCSI based attachments to Fibre Channel. A diagram to show a SAN Data Gateway configuration using a single host is shown in Figure 4-1.
Expand connectivity to storage devices with use of IBM SAN hubs, switches, and directors Perform channel zoning and LUN masking capability to allow access at a volume level Overcome the distance limitations of SCSI based host systems using longwave ports that support distances up to 10 km Utilize the StorWatch SAN Data Gateway Specialist which is an easy to use interface for managing and controlling access of host systems to storage devices The SAN Data Gateway is available as a rack-mount unit or a
If a PC with terminal emulation is used, a 9-pin female to 9-pin female, null modem cable is required and is provided with the unit. Once connected, power on the SAN Data Gateway and the start up messages will appear and scroll across the window. When the power on sequence has completed, a prompt Gateway> appears on the window, as shown in Figure 4-3. Figure 4-3 IBM Storage Area Network Data Gateway startup If you type in help and then press Enter, a list of available commands is provided.
4.2.1 Setting the Ethernet address Once restarted, the Ethernet port must be configured and attached using network information provided by the network administrator. To set the IP address, use the ethAddrSet command. The address must contain the double quotes (“): Gateway > ethAddrSet “9.111.24.66” Network not Enabled Write complete Host Address set to 9.111.24.
You cannot telnet to the Gateway and use the service port at the same time. When you telnet to the Gateway, the service port on the rear of the unit will stop its communications. After you end the telnet session, then the service port will become available again. 4.2.3 Startup sequence You must start up the SAN Data Gateway and the attached host and target devices in a specific order. When you add or remove SCSI devices or update firmware, you must restart.
Currently, up to eight different hosts can be attached to each Fibre Channel port. If all six ports are installed, then 48 different hosts can attach to the SAN Data Gateway. 4.3 StorWatch SAN Data Gateway Specialist The StorWatch SAN Data Gateway Specialist software provides remote capability for all management, configuration, and event notification.
4.3.1 Installing StorWatch Specialist The Specialist software is not bundled with the SAN Data Gateway. The Specialist software is downloaded using a Web browser by going to the IBM Storage Area Network Data Gateway Web site: http://www.storage.ibm.com/hardsoft/products/sangateway/support/form1.htm This will take you to a registration window. Enter the required information and select Submit Information. A license agreement window is shown, and once reviewed, select I agree.
Figure 4-5 StorWatch SAN Data Gateway Specialist server The client software can now be launched. If the server and client are not on the same PC, then a dialog box will appear to allow you to enter in the IP address of the computer that has the server software loaded. If the server and client are on the same computer you will be automatically connected to this server. After connection to the server is complete, a dialog box will appear, in which you can enter in a user name and password.
4.3.2 Using the StorWatch SAN Data Gateway Specialist Once you are logged in to the Specialist, you must now connect to the SAN Data Gateway. At this point, a dialog box should appear, requesting the IP address of the SAN Data Gateway. As it connects, it will download the information from the SAN Data Gateway and be presented on your window. If a dialog box does not appear automatically, select Tools-> Connect SAN Data Gateway or SAN Data Gateway Router from the toolbar.
As we can connect to several SAN Data Gateway systems from one client session, select the particular Gateway you want and it will be highlighted in blue, as shown in Figure 4-7. Figure 4-7 Selecting from multiple SAN Data Gateways On the left-hand side of the highlighted Gateway, there is a small key, and by selecting this, it expands the view to show you all SCSI ports and installed Fibre Channel ports. For example, Figure 4-8 shows a Gateway with four SCSI ports and two Fibre Channel ports.
SCSI channels 1, 2, and 3 and Fibre Channel ports 1 and 4 also have a key on the left-hand side to depict that there are devices attached. By selecting a key, you will now expand the tree, as seen in Figure 4-9, and view the different disk devices attached.
You can also select and highlight each SCSI Channel. You will notice that as you do this, the information window on the right side will provide data that is unique to that SCSI channel, as shown in Figure 4-10. Figure 4-10 SCSI channel data Chapter 4.
Information pertaining to a particular disk device is shown in Figure 4-11.
You can perform the same drill-down on the Fibre Channel host as we did with the SCSI channel and disk devices. Select one of the Fibre Channel port connections, as shown in Figure 4-12, and its data will be shown on the right-hand view pane. Figure 4-12 Fibre Channel port data By selecting the key to the left, you can expand the tree and select the host system attached to that port. Figure 4-13 shows the detail on the specific host. Chapter 4.
Figure 4-13 Fibre Channel host data As you select and highlight the different ports or devices, there are different options available from the top toolbar. If an SCSI channel is highlighted, select Controls from the toolbar. You will notice that all options are grayed out except for SCSI Channel. Once selected, a dialog box will appear, as shown in Figure 4-14, and display the settings for the SCSI channel.
Figure 4-15 Advanced SCSI parameters Fibre Channel parameters are displayed in a similar fashion. Highlight a Fibre Channel port and select Controls from the toolbar, you will notice that now all options are grayed out except the Fibre Channel option. By selecting this, a dialog box will display the parameters that can be changed for the Fibre Channel port selected. If any of the settings, as shown in Figure 4-16, are changed, the SAN Data Gateway must be restarted.
A detailed description of the SCSI and Fibre Channel settings can be found in the IBM Storage Area Network Data Gateway Installation and User’s Guide, SC26-7304. 4.3.3 Upgrading the firmware New versions of the SAN Data Gateway firmware can be downloaded from http://www.storage.ibm.com/hardsoft/products/sangateway/gatewayspec.htm This site has a link to a Downloads section from where the latest versions of the SAN Data Gateway can be downloaded.
From this screen, we selected the Windows NT version, and we will save this into a directory for downloading into the SAN Data Gateway at a later stage. From the main screen of the SAN Data Gateway Specialist, you can check the current Firmware Revision level, as shown in Figure 4-18. Figure 4-18 SAN Data Gateway Firmware Revision Level New versions of the SAN Data Gateway can be downloaded, as shown in Figure 4-19, by selecting the Controls option and then selecting the Update Firmware option.
The SAN Data Gateway Specialist will then prompt for the location of the new firmware as shown in Figure 4-20. This is the file that was downloaded from the SAN Data Gateway Web site previously. Figure 4-20 Specifying location of the firmware Downloading the firmware into the SAN Data Gateway is a disruptive process, so the Specialist displays a warning message as shown in Figure 4-21.
Once the firmware process has been started, the SAN Data Gateway Specialist displays a message as shown in Figure 4-22. Figure 4-22 Download in progress This indicates that the firmware is being downloaded to the SAN Data Gateway. The Specialist will then give the option to restart the SAN Data Gateway as shown in Figure 4-23. Figure 4-23 Message prior to restarting the SAN Data Gateway Chapter 4.
As restarting the SAN Data Gateway is a disruptive process, the Specialist issues a warning as shown below in Figure 4-24. Figure 4-24 Warning message prior to restarting the SAN Data Gateway The Specialist will now return to the main screen, and a message is displayed at the bottom of the screen as shown in Figure 4-25 indicating that the Gateway is in the process of restarting.
Once the restart process is completed, the Specialist displays the prompt shown in Figure 4-26 before refreshing the display. Figure 4-26 Restart completed message Once the Specialist has finished refreshing, we can see that the firmware revision level has been updated successfully, as shown in Figure 4-27. Figure 4-27 New firmware revision level Chapter 4.
4.4 SCSI devices The four SCSI ports on the SAN Data Gateway support Differential Ultra Wide SCSI devices. It will automatically negotiate speed for wide or narrow bus width devices as well as standard, fast and, ultra speeds. The SAN Data Gateway provides a termination on each of its SCSI buses. The termination can be separately disabled if so desired from the StorWatch SAN Data Gateway Specialist window. 4.4.1 LUN support The SAN Data Gateway provides support for up to 256 LUNs.
The SAN Data Gateway must add this extra layer of addressing as the host is no longer directly attached to the disk devices but will see a single target ID, the SAN Data Gateway. In a regular SCSI environment with the host attached directly to the SCSI device, the host is able to recognize the target and LUN ID of that device. Since we add the SAN Data Gateway in between the host and the device, the host is no longer directly attached to the SCSI device to recognize its target or LUN ID. 4.4.
When this command is issued, you will also be given an option to clear the Virtual Private SAN (VPS) access settings. Because this allows host access to specific LUNs and by issuing this command, we may change the assigned LUN ID; it is recommended that you always say Yes to this option. If not, a host may access a volume that you do not want it to access and be restricted from a volume that it had access to previously.
A check mark will allow access, and as shown in Figure 4-28, you can see that SAN connection 1 has access to SCSI channels 3 and 4, but not to SCSI channels 1 and 2. SAN connection 4 has access to SCSI channels 1 and 2, but not to SCSI channels 3 and 4. To change the settings, click any box and the check mark will toggle on and off. All combinations are possible. Once the desired settings are selected, click OK. For the new zone settings to take effect, the SAN Data Gateway must be restarted. 4.5.
A dialog box appears requesting the entry of a license key number. For units with serial numbers lower than 1300600 you are required to get a license key that is matched to the serial number of the SAN Data Gateway. For units with serial numbers higher than 1300600, the VPS feature is bundled into the product. Type the word enable and the feature will now be enabled. Host registration With VPS enabled, the host registration service is also available. This service is provided to simplify VPS configurations.
Once installed, the service runs automatically and does not require further administrator interaction. Communicating with the Gateway Once the host is attached to the Gateway and restarted, the registration service will communicate to the Gateway. The data shown on the Gateway will have to be refreshed by selecting View-> Refresh SAN Data Gateway. This will cause the updated data to be shown in the Specialist window.
Figure 4-32 Specialist after VPS enabled and host registration If no registration software is loaded on the host, or is not available for a specific operating system, for example a Fibre Channel switch, only the WWN of the attached system will register to the VPS database, all other fields will have unknown. This is shown in Figure 4-33. Figure 4-33 Host system with no host registration software A diagram to depict the setup that was described, is shown in Figure 4-34.
Servers 'unknown' 'plymouth' Fibre Channel 1 4 Gateway 1 2 3 SCSI ESS Figure 4-34 SAN Data Gateway with two hosts Adding host and connection information The host name and host type and connection information can be added manually by selecting the VPS feature and modifying the data. To do this, select the desired SAN Data Gateway so that it is highlighted. Select Controls -> Access Options -> Virtual Private SAN, and you will enter into the VPS settings window.
Select the host to be modified, and the information is transferred to the bottom part of the window where the modifications can take place, as shown in Figure 4-35. Figure 4-35 VPS host settings The Host information can now be entered in the left-hand column labeled New Host Data. The Original Host Data column displays the values before any changes are made. Once completed, select Apply Changes and then Close.
Setting up a Virtual Private SAN Remember that the SAN Data Gateway assigns its own LUN numbers, and it does so in a sequential order. An attached disk device may have a SCSI target of 1 with a LUN ID of 4, but when recognized by the SAN Data Gateway, its LUN number assigned will be something completely different. An example is shown in Figure 4-36. Figure 4-36 SCSI LUN assignment In this example, the disk device attached has a SCSI target of 6 and LUN ID of 1.
SCSI to LUN map Prior to setting the LUN masking, it makes sense to prepare a list that has each LUN and shows which host is allocated to that LUN. To assist in this process the Gateway has a map that provides a cross reference of the actual SCSI target and ID to the Gateway’s assigned LUN number. To access this map you must access the service terminal from the serial port as described earlier in 4.2, “Installation” on page 273. Once connected, type in the command mapShowDatabase.
Setting host access to LUNs To view and set host access to particular LUNs, access the Virtual Private SAN Access Settings by selecting Controls -> Access Options -> Virtual Private SAN. This window will show all the hosts that have registered to the Gateway. To allow a host to access a particular LUN, place a check mark in the row that corresponds to the host. To disable access, the square must be clear, without a check mark. The check mark is toggled on and off by clicking in each square.
Once completed, select Apply Changes and then Close. The host system may now have to be restarted or some other method used to rescan the bus for the host to detect that it now has access to new devices. Any combination is allowed, so if the same LUN is to be shared by two different hosts, a check mark for that LUN must be set for both. If this is the case, the host systems must have a device sharing software installed to control access to the disk device for data integrity.
SAN connection 1 has access to the LUNs on SCSI channels 2 and 4. VPS will control access to LUNs on SCSI channels 2 and 4 for SAN connection 1. Since SAN 1 is not zoned for SCSI channel 1 and 3, it will not reach any LUNs on these channels. Even, if a LUN on SCSI 3 has been enabled for access to a host on SAN 1 in the VPS settings window, the host will not see that LUN because of the zoned settings. The same is true for SAN connection 4.
Figure 4-40 Fibre Channel port setting for switch attachment By default, the setting in the Connection Options box will have Loop enabled. For switch or fabric connection, select Point to Point, and then select OK. The SAN Data Gateway must be restarted for the change to take effect. Note: Connection Options box does not appear, the Fibre Channel module installed will support loop connection only and will not support fabric connection.
Switch registration With VPS enabled, the switch will register with the database the WWPN of the port on the IBM SAN Fibre Channel Switch. Figure 4-41 shows a switch connected to port 4 of the Gateway, and which has registered to the database. Since we cannot load any host registration software onto the switch, all other fields are left unknown.
Figure 4-42 Switch port login You can also check the port on the switch, by accessing the switch configuration from a Web browser, to ensure that the port has registered and is communicating properly. Figure 4-43 shows that, in this example, port 7 of the switch was used to connect to the Gateway. Note that the WWPN is the same as in the Specialist window and that the port type is F_port.
Changing the switch information From the VPS Access Settings window, it is possible to change the unknown information of the switch. Select Controls -> Virtual Private SAN, and select the entry that has the WWN of the switch. You can now change the information to further describe the switch, or other pertinent information if desired. This is shown in Figure 4-44. Figure 4-44 Changing switch information Chapter 4.
The information on the host plymouth, that was attached using port 4, is still kept in the VPS database, but it is now shown in blue to indicate that it is offline. Figure 4-45 is a diagram showing the configuration with the switch.
Once changed, as shown in Figure 4-46, the information shown on the main window will reflect the change, and the icon on the left-hand side of the WWN changes to depict a switch. Figure 4-46 Switch port information Attaching hosts to the switch Any hosts that will attach to the switch should have the host registration software loaded and installed. Refer to , “Host registration” on page 298 for details. Plug in the Fibre Channel cable from their respective adapters and power on, or restart the host.
By selecting and highlighting the new NT host, we can see its information, which was automatically sent by the host registration process. The configuration with two different hosts connected to the switch, as described previously, is shown in Figure 4-48.
Device access To have the new NT host access some of the devices, you need to set up the VPS Access parameters by selecting Controls -> Access Options -> Virtual Private SAN. Figure 4-49 shows the switch and the two hosts in the VPS database. As plymouth was previously connected direct to the Gateway, its settings have been maintained, but now that it is reconnected, it is back online. The NT host does not have any LUN access yet. Figure 4-49 VPS Access window with switch and two hosts Chapter 4.
A check mark in the box allows a host access to a LUN, or clear the box if you want to restrict access as described in , “Setting up a Virtual Private SAN” on page 303. Figure 4-50 shows the host plymouth with access to assigned LUNS 17 to 24, and the NT host FIBRE1 is now set with access to LUNs 25 to 30. Figure 4-50 Setting LUN access for the host FIBRE1 Once you select Apply Changes and then Close, the new settings will be in effect.
ESS considerations As you configure and allow access from host systems to volumes that are in the ESS through the Gateway, you must consider how the volumes were created within the ESS. Volumes in the ESS are assigned to a particular SCSI port in the ESS. As they are assigned, you also specify the host type that will access these volumes.
Software, similar to SDD, is required in the host for it to recognize that it may have two or more paths to the same volume. If this was not loaded on the host, the host would recognize the volumes as being different, and there will be contention problems, and data corruption. As SDD is currently only supported on SCSI adapters and not with Fibre Channel adapters, it is also not supported when using the SAN Data Gateway to connect to an ESS. 4.7.
Another option here is to utilize the zoning and LUN masking capabilities of the SAN Data Gateway. This would ensure that certain volumes can only be accessed on one Fibre Channel Gateway port and by a particular host. Also available is to add zoning within the switch. The switch Fibre Channel ports can be zoned so that the host only has one path to the SAN Data Gateway. This would be used in combination with the zoning and LUN masking features of the SAN Data Gateway. Chapter 4.
320 Implementing an Open IBM SAN
5 Chapter 5. Implementing the Vicom Fibre Channel SLIC Router The Vicom Fibre Channel SLIC Router, 7139-111, enables all IBM 7133, 7131, and 3527 SSA Serial Disk Systems to attach to host systems using Fibre Channel host adapters and drivers. This allows you to protect your investment in SSA disk, while being able to create and build a SAN infrastructure.
5.1 SLIC Router features The Instant Copy function can create a separately addressable copy of mirrored data that can be used for tape backup. After the backup has completed, data is resynchronized with the primary copy. The Vicom Fibre Channel SLIC Router also can create composite drives by concatenating up to 16 physical disks. Using these functions, physical drives become members of larger or more complex logical drives. A diagram to depict a single host to Router configuration is shown in Figure 5-1.
5.2 Installing the SLIC Router To install and establish a SLIC storage system, the Router and all the disks to be used must be set up in a proper sequence. During this sequence, only a single SLIC Router must be used to do the configuration. Once configured, other SLIC Routers can be added to the loop. This sequence is described below. Note: For information and a description to understand the LED codes that will be discussed, please refer to the SLIC Router Installation and User’s Guide, 310-605759. 1.
LED on the SSA drive will flash and then it will move to the next drive. This test should continue until all drives have been tested. The test runs in a continuous cycle, so once all drives have been tested at least once, the Router is powered off. If a drive fails the test, the testing will stop, and the Router’s Status LED will flash a diagnostic code. A code map with a description of the errors can be found in the SLIC Router Installation and User’s Guide, 310-605759. 5. Assign Fibre Channel target.
7. Create general spares. The mode on SW2 is changed to mode 12 to set all drives to be general spares. Mode 12 is represented by setting switch 2 and 3 down and the rest turned up. Power on the Router again, the Status LED will flash. After approximately one minute, the LED will flash code 100 to indicate it has completed. The Router is again powered off. 8. Format the drives. The Router is set to mode 14, switch 1, 2, and 3 down on SW2, to format all disk drives.
5.3 SLIC Manager software Rather than using the dip switches to configure the features, another option is to use the SLIC Manager software. The SLIC Manager also provides configuration, monitoring and management capabilities of the SLIC Router and the SSA drive loop. The SLIC Manager can be setup to allow remote access if desired. The Manager software consists of server and client portions. The server includes a daemon service and a user interface. The client has the user interface only.
The SLIC Manager has a graphical user interface (GUI) and a command line interface (CLI) available for Windows NT systems. UNIX systems will only have the command line interface available. The following installation and configuration examples will detail using the GUI from a Windows NT platform. To review the commands that are available for UNIX platforms and installation instructions on other operating systems, please refer to the SLIC Manager Installation and User Guide, 310-605807. 5.3.
Editing the configuration file When the software is loaded, a sample configuration file called 7190.cfg is added in the C:\ibm7190\sdus directory. This is a text file that can be viewed and edited by simple text editors, such as Windows Wordpad. Open up the 7190.cfg file and it will contain a sample of how the file should look. Also note that on the left hand side, the # sign is entered in every line to mark it out as a comment. This is shown in Figure 5-4.
This file can now be edited to be used as the configuration file for your SLIC system. Begin by deleting the # sign on the lines that contain the sample configuration. The rest of the information can be entered in, as shown in Figure 5-5. A description of each entry field is also provided. Figure 5-5 Edited configuration file Chapter 5.
Configuration file information The SLIC_name can be any name that you would like to use to identify the Router. Creating a SLIC Zone The path refers to the SLIC Zone, file or partition, used for the Manager to communicate to the Router. To edit this option, it requires that a drive on the SSA loop has been recognized by the host and that the drive has been formatted. In the example above, a Windows NT host was used.
Installing the SLIC Manager daemon With the configuration file edited and a SLIC Zone created, the daemon service can be installed and run. To install the service in Windows NT, open a DOS prompt and go to C:\ibm7190\sdus. Type in slicd -install, and the daemon will be installed. Starting the SLIC Manager daemon To start the daemon service, select Start -> Settings -> Control panel from Windows NT. Double-click the Services icon. Scroll down until you see Vicom SLIC Manager; select and highlight it.
5.3.3 Starting the SLIC Manager To start the Vicom SLIC Manager software, select Start -> Programs -> Vicom -> Vicom SLIC Manager. The software will load, and a dialog box will appear. In the box with the heading Hostname, enter in the name or IP address of the host the daemon service is running. Enter in the SLIC name you entered in when editing the 7190.cfg file. An example is shown in Figure 5-7. Figure 5-7 SLIC connection window Click OK and the software will begin to communicate to the Router.
5.4 Using SLIC Manager You can now look to see that all communications are working properly by going to the toolbar and selecting Tools -> Control Center. A dialog box will appear, as shown in Figure 5-9. Figure 5-9 Control Center window In the Physical Drive box, the drives that are on the SSA loop can be seen. This window will be useful as you start to create mirrors and composite drives, because it provides a summary of all drives. Chapter 5.
5.4.1 Drive properties You can get detailed information on each drive. Select the drive so that it is highlighted and then select Properties. A dialog box will appear with the drive’s information, as shown in Figure 5-10. Figure 5-10 Disk drive properties Here you can see its SSA attributes, its Fibre Channel attributes and its model type and serial number. By clicking the Identify button, the LED on the selected drive will begin to flash. 5.4.
Figure 5-11 SLIC Router properties 5.4.3 Setting Router to master As you move through the toolbar, you may notice that most selections have been grayed out. This is due to the fact that the Router is currently in a subordinate role and does not have access to create mirrors or composite drives. This function is done by a Master Router. There can be only one master in a SLIC loop. This is used as more Routers and more disks can be added to the loop.
Click in the Always Master box so that a check mark appears. Once you click OK, the Router will then be set as the master. You will notice that now all options in the toolbar are available and can start to use the features of the SLIC Router. By placing a check mark in the Auto Start box, the SLIC Manager will automatically connect to the Router defined in the Connection window, as seen in Figure 5-7 on page 332. 5.4.
5.5 Composite drive A composite drive is a large drive that consists of two or more smaller drives. The capacity of the composite drive is an aggregate of the capacities of all the smaller drives that are used to comprise this one large drive. 5.5.1 Creating a composite drive To create a composite drive from the SLIC Manager, select Tools -> Composite Drive Setup Wizard. A dialog box, Composite Drive List, will appear. Currently, the list will be blank, because there are no composite drives created.
From the Available Drive Pool list, click a desired drive and then click the Add>> button. The drive name will be added to the Member window. An asterisk will appear on the left hand side of the drive that was selected in the Available Drive window, to denote that the drive has been selected. Each drive is added one at a time. To remove a drive from the Member window, select the desired drive and click the Remove<< button. Below each window there is a Drive Capacity box.
When all the desired drives are added, click Next>. The Assigning Properties window opens, as shown in Figure 5-16. Figure 5-16 Assigning Composite Drive Properties window The properties that can be changed are the FC LUN and the Drive Name. There will be a suggested LUN number in this field that can be accepted. If not, simply type in the desired LUN number. The name can also be defined to the composite drive for easier identification, with a limit of up to eight characters.
Click the Next> button and a dialog box will appear, as shown in Figure 5-17, to allow you to create another composite drive. Click Yes If you would like to create another composite drive, and the Composite Drive List window opens and the steps described above can be repeated. Figure 5-17 Completing the Composite Drive setup Click Finish when you have created all the desired composite drives. Up to this point, the configuration has been kept within the SLIC Manager software.
5.5.2 Composite drive properties If you view the Control Center again, by selecting Tools -> Control Center, as shown in Figure 5-18, the newly created composite drive is listed in the Composite Drive box. Figure 5-18 Control Center with composite drive Chapter 5.
Select the composite drive and then click the Properties button. The Composite Drive Properties dialog box opens, as shown in Figure 5-19. Figure 5-19 Composite Drive Properties Here you can find information on the Composite Drive about the member drives that make up the composite, as well as the Composite Drive characteristics. Clicking on the Identify Composite button will cause the LED on the actual SSA drives, that belong to the Composite, to flash.
5.6.1 Creating a mirror drive To create a mirror using physical drives, from the toolbar, select Tools -> Mirror Drive Setup Wizard. You will see a dialog box, Mirror Drive List, that will be blank. If there were mirror drives created, then it would display the names of the drives. Click the Next>> button and the Mirror Drive Members Selection window opens. The window on the left named Available Drive Pool contains a list off all drives that are candidates to participate in a mirror drive.
Below each window there is a Capacity window that will display the size of the available drive, or of the mirror drive. Each drive that participates in a mirror should be of equal capacity. If you select a drive that has a greater capacity and add it to the mirror, the mirror capacity will still be the smaller of the two, and the rest of the capacity of the larger drive will be unused.
The Assigning Mirror Drive Properties window is shown in Figure 5-22. Figure 5-22 Mirror drive properties Click the Next> button and a dialog box appears to allow you to create another mirror drive. Click Yes, if you would like to create another mirror drive, and the Mirror Drive List window opens, and the steps described above can be repeated. Click Finish when you have created all the desired mirror drives. Up to this point, the configuration has been kept within the SLIC Manager software.
5.6.2 Mirror drive properties If you go to the Control Center window by selecting Tools -> Control Center, you will see that the mirror drive is now displayed in the Mirror Drive window. This is shown in Figure 5-23.
If you select and highlight the mirror drive and then click the Properties button, the Mirror Properties window opens, and you can see the information in the mirror drive. Figure 5-24 shows an example of the properties of the mirror drive. Figure 5-24 Mirror Drive Properties Clicking on the Identify Mirror button will cause the LED on the actual SSA drives that belong to the mirror, to flash.
5.7.1 Creating an Instant Copy drive To create an Instant Copy drive, select Tools -> Instant Copy Drive Setup Wizard. You will see a dialog box, Instant Copy Drive List, that will be blank. If there were copy drives created, it would display the names of the drives. Click the Next>> button and the Instant Copy Drive Members Selection window is displayed. The window on the left named Available Drive Pool contains a list off all drives that are candidates to become a copy drive.
Figure 5-26 Instant Copy Drive Properties The properties that can be changed are the FC LUN and the Drive Name. There will be a suggested LUN number in this field that can be accepted. If not, simply type in the desired LUN number. A name can also be defined to the copy drive for easier identification, with a limit of up to eight characters. Click the Next> button and a dialog box appears to allow you to create another copy drive.
Figure 5-27 Control Center with Instant Copy Drive Notice that in the information provided for the copy drive, there is an IC included to distinguish between mirror drives and copy drives within this window. 5.7.3 Adding an Instant Copy Drive to a mirror To add or detach the copy drive from a mirror, you select and highlight the mirror drive, and then click the Properties button. The Mirror Drive Properties window opens, as shown in Figure 5-24 on page 347.
Figure 5-28 Add Mirror Member display Select and highlight the copy drive from the Available Drive Pool window, click the Add Member button, and the name of the copy drive will appear in the New Mirror Member window. This is shown in Figure 5-29. Figure 5-29 Adding drive members to a mirror Chapter 5.
Click the OK button, and the Mirror Drive Properties will now reflect the change, as shown in Figure 5-30. Figure 5-30 Mirror drive properties with copy drive attached Click OK to complete the process. 5.7.4 Detach Instant Copy Drive from a mirror To detach, or split off the copy drive from the mirror, the procedure is similar except at the Mirror Drive Properties window, select Delete Member. A window will appear that displays all current members of the Mirror.
5.8.1 Creating a second composite drive To provide an example of a mirror using only composite drives, another composite drive is required. The example shown in Figure 5-31 shows that drive 6 and 7 were used to create another composite drive. Figure 5-31 Creating composite drive to be used in a mirror Chapter 5.
Follow the steps to create a composite drive as described in 5.5, “Composite drive” on page 337. Once created, you can view the Control Center window by selecting Tools -> Control Center from the toolbar. Figure 5-32 shows that there are now two composite drives.
5.8.2 Creating the mirror The mirror can now be created by selecting Tools -> Mirror Drive Setup Wizard. When the Member Selection window appears, select the composite drives as members of a mirror. Figure 5-33 shows where composite drives ‘ITSO1’ and ‘ITSO3’ are selected as members of a mirror. Figure 5-33 Creating mirror drive from two composite drives Continue through the Mirror Drive Setup Wizard to complete the process as described in 5.6, “Mirror drive” on page 342. Chapter 5.
5.8.3 Viewing mirror drive using composite drives With the Mirror Drive Setup Wizard completed, you can now view the Control Center window once again, as shown in Figure 5-34.
In the Mirror Drive window, you can see the new mirror drive that was created above and named CompMir. In the Composite Drive window you can see that there are still the two composite drives, but instead of having Fibre Channel LUN numbers assigned to them, they are shown as belonging to a mirror with the name CompMir. You can highlight the CompMir drive and click the Properties button. All the same functions that were described in 5.6.2, “Mirror drive properties” on page 346 are available. 5.
Figure 5-35 Removing a logical drive The logical drive that was created as an Instant Copy Drive in 5.7.1, “Creating an Instant Copy drive” on page 348, has been removed and is now a general spare.
5.9.2 Mapping a general spare You will notice in Figure 5-35 that the general spare does not have a LUN number assigned to it. To get a new LUN number for this drive, you select the drive and click the Properties button. The Drive Properties window appears; select the Change FC button. A dialog box opens, as shown in Figure 5-36. Figure 5-36 Mapping a general spare The Original Mapping box will indicate that no LUN was assigned previously.
However, since each composite drive had its attributes changed as it became a member of the mirror, it will no longer be mapped. The composite drives will show up as UnMapped in the Control Center window. This is shown in Figure 5-37. The mirror created in 5.8.2, “Creating the mirror” on page 355 was removed. Figure 5-37 UnMapped composite drives The existing composite drives ‘ITSO1’ and ‘ITSO3’ cannot be mapped or given a new LUN number at this point. Each logical composite drive must be removed as well.
5.10 Expanding the SLIC system The SLIC storage system can be expanded to add more SSA disks or more SLIC Routers. Each SLIC storage system can support up to 64 SSA disks and have 16 Routers. 5.10.1 Adding disk To add disk to an existing SLIC system is very easy, because they are SSA disks and the rules for SSA disks apply here as well. If there is a dummy drive in an existing SSA drawer, then it can be replaced by a real drive.
On the SSA side, there are a few options available. Each Router can have its own SSA loop so that each one can support 64 SSA disks. In this way, storage capacity is scalable, because it can be increased by adding more Routers. This is shown in Figure 5-38. Server Fibre Channel Vicom SLIC Router Vicom SLIC Router A1 A2 A1 SSA Disks A2 SSA Disks Figure 5-38 Increasing storage capacity The other option is to have each additional Router added to the same SSA loop.
Installing additional Routers With the existing Router and storage system powered on: 1. The node map on the new Router must be cleared first. 2. Connect the Router to the existing system with the power off. 3. Set SW2 to mode 3 and set SW1 to an unique Fibre Channel target ID. 4. Power on the new Router. 5. When the Status LED on the new Router is on (solid lit), the install is complete. When the new Router is powered on, communication between the two Routers will occur to query and update the new Router.
5.10.3 Adding hosts The SLIC storage system can be expanded to include more hosts whether they are homogeneous or heterogeneous. It is recommended that as hosts are added, each host is connected to its own and separate SLIC Router. If more than one host was connected to a single Router, there will be arbitration and performance issues. Also, it would have a single point of failure with the possibility of losing data access to many systems.
Related publications The publications listed in this section are considered particularly suitable for a more detailed discussion of the topics covered in this redbook. IBM Redbooks For information on ordering these publications, see “How to get IBM Redbooks” on page 368.
IBM Enterprise Storage Server Quick Configuration Guide, SC26-7354 IBM Enterprise Storage Server Introduction and Planning Guide, 2105 Models E10, E20, F10 and F20, GC26-7294 IBM Enterprise Storage Server User’s Guide, 2105 Models E10, E20, F10 and F20, SC26-7295 IBM Enterprise Storage Server Host Systems Attachment Guide, 2105 Models E10, E20, F10 and F20, SC26-7296 IBM Enterprise Storage Server SCSI Command Reference, 2105 Models E10, E20, F10 and F20, SC26-7297 IBM Enterprise Storage Server S
Referenced Web sites These Web sites are also relevant as further information sources: www.storage.ibm.com/ibmsan/index.htm IBM Enterprise SAN www.pc.ibm.com/ww/netfinity/san IBM Storage Area Networks: Nefinity Servers www.storage.ibm.com/hardsoft/products/sangateway/supserver.htm IBM SAN Data Gateway www.storage.ibm.com/hardsoft/products/tape/ro3superserver.htm IBM SAN Data Gateway Router www.storage.ibm.com/hardsoft/products/fcss/fcss.htm IBM Fibre Channel RAID Storage Server www.storage.ibm.
www.vicom.com Vicom Systems www.vixel.com Vixel www.scsita.org SCSI Trade Association www.futureio.org InfiniBand (SM) Trade Association www.nsic.org National Storage Industry Consortium www.ietf.org Internet Engineering Task Force www.ansi.org American National Standards Institute www.standards.ieee.org Institute of Electrical and Electronics Engineers www.pc.ibm.
Special notices References in this publication to IBM products, programs or services do not imply that IBM intends to make these available in all countries in which IBM operates. Any reference to an IBM product, program, or service is not intended to state or imply that only IBM's product, program, or service may be used. Any functionally equivalent program that does not infringe any of IBM's intellectual property rights may be used instead of the IBM product, program or service.
The following terms are trademarks of other companies: Tivoli, Manage. Anything. Anywhere.,The Power To Manage., Anything. Anywhere.,TME, NetView, Cross-Site, Tivoli Ready, Tivoli Certified, Planet Tivoli, and Tivoli Enterprise are trademarks or registered trademarks of Tivoli Systems Inc., an IBM company, in the United States, other countries, or both. In Denmark, Tivoli is a trademark licensed from Kjøbenhavns Sommer - Tivoli A/S. C-bus is a trademark of Corollary, Inc.
Glossary 8B/10B A data encoding scheme developed by IBM, translating byte-wide data to an encoded 10-bit format. Fibre Channel's FC-1 level defines this as the method to be used to encode and decode data transmissions over the Fibre channel. Adapter A hardware unit that aggregates other I/O units, devices or communications links to a system bus. ADSM ADSTAR Distributed Storage Manager.
Class-1 A class of service providing dedicated connection between two ports with confirmed delivery or notification of non-deliverability. Class-2 A class of service providing a frame switching service between two ports with confirmed delivery or notification of non-deliverability. Class-3 A class of service providing frame switching datagram service between two ports or a multicast service between a multicast originator and one or more multicast recipients.
Enterprise Network A geographically dispersed network under the auspices of one organization. Entity In general, a real or existing thing from the Latin ens, or being, which makes the distinction between a thing's existence and it qualities. In programming, engineering and probably many other contexts, the word is used to identify units, whether concrete things or abstract ideas, that have no ready name or label.
X.271-1996) which defines how the ESCON command set protocol is transported using the fibre channel. FSP Fibre Channel Service Protocol - The common FC-4 level protocol for all services, transparent to the fabric type or topology.
data between CPUs and from a CPU to disk arrays and other peripherals. HMMP HyperMedia Management Protocol HMMS HyperMedia Management Schema - the definition of an implementation-independent, extensible, common data description/schema allowing data from a variety of sources to be described and accessed in real time regardless of the source of the data.
Loop Circuit A temporary point-to-point like path that allows bi-directional communications between loop-capable ports. Loop Topology An interconnection structure in which each point has physical links to two neighbors resulting in a closed circuit. In a loop topology, the available bandwidth is shared. LVD Low Voltage Differential Management Agent A process that exchanges a managed node's information with a management station.
N_Port Login N_Port Login (PLOGI) allows two N_Ports to establish a session and exchange identities and service parameters. It is performed following completion of the fabric login process and prior to the FC-4 level operations with the destination port. N_Port Login may be either explicit or implicit. Name Server Provides translation from a given node name to one or more associated N_Port identifiers.
Port Bypass Circuit A circuit used in hubs and disk enclosures to automatically open or close the loop to add or remove nodes on the loop. Private NL_Port An NL_Port which does not attempt login with the fabric and only communicates with other NL Ports on the same loop. Protocol A data transmission convention encompassing timing, control, formatting and data representation. Public NL_Port An NL_Port that attempts login with the fabric and can observe the rules of either public or private loop behavior.
Fast Wide SCSI -2 Ultra SCSI Ultra SCSI -2 Ultra 2 LVD SCSI 10 16 20 15 6 SMART Self Monitoring and Reporting Technology SM Single Mode - See Single-Mode Fiber 20 8 20 7 1.5 20 16 40 7 12 40 16 80 15 12 SCSI-3 SCSI-3 consists of a set of primary commands and additional specialized command sets to meet the needs of specific device types.
higher value functions including -- reporting of capacity, performance, etc. over time (trends), configuration of multiple devices based on policies, monitoring of capacity and performance, automated responses to events or conditions, and storage related data mining. StorWatch Specialist A StorWatch interface for managing an individual fibre Channel device or a limited number of like devices (that can be viewed as a single group).
WAN Wide Area Network - A network which encompasses inter-connectivity between devices over a wide geographic area. A wide area network may be privately owned or rented, but the term usually connotes the inclusion of public (shared) networks. WDM Wave Division Multiplexing - A technology that puts data from different sources together on an optical fiber, with each signal carried on its own separate light wavelength.
382 Implementing an Open IBM SAN
Index Numerics 1-way mirror 342 2042-001 270 2766 adapter 10 2-way mirror 342 3527 321 3-way mirror 342 7131 321 7133 321 A access levels 94 access profile 163 Access_Any 163 Access_Restricted 163 activate hard zone layout 252 adapter types 2 Add Volumes 126 Adding an Instant Copy Drive to a Mirror 350 Adding Disk 361 Adding Hosts 364 Adding Routers 361 Adding SCSI devices 295 administrator 279 Advanced Options 286 AdvFS 22 Agent 277 AIX host system 15 All or Nothing 240 anonymous host 163 Arbitrated Loop
E Emulex 46 Emulex LP8000 32, 42, 47, 53, 56, 58, 65, 67 enabling the Ethernet port 275 Enabling VPS 297 ESCON 5, 104 ESS 317–318 ESS configuration update 110 ESS configuring 89 ESS Copy Server 90 ESS default user 81 ESS disk group section 100 ESS FC port 98 ESS FC port access 123 ESS FC port definitions 116 ESS FC ports config 158 ESS FC ports topology 159 ESS host interfaces 97 ESS interface adapters 160 ESS introduction panel 78 ESS logical volume 130 ESS logical volume assignments 142 ESS logical volume
2108-G07 271 Initialization Method 339 initializeBox 274 INRANGE defining hard zoning 246 INRANGE FC/9000 175 INRANGE hard zoning 237 INRANGE hard zoning rules 238 INRANGE Name Server zoning 242 INRANGE zoning methods 237 Installing additional Routers 363 Installing StorWatch Specialist 278 Installing the SLIC Manager 327 Installing the SLIC Router 323 Instant Copy 322, 347 Instant Copy Drive 342 Instant Copy Drive Properties 349 Inter Switch Link 263 interconnected directors 7 IN-VSN event log 269 IOC-0210
nicknames 122 non-RAID 109 notification options 85 Novell NetWare FC attachment 26 NUMA-Q 15 O Open System Storage 104 orphan zone 245, 256 OS/400 operating system 11 P parallel interface 6 path group 5 Point to Point 308 port granularity 243 port groups 238 PPRC 77 Private Attributes 364 Private Initiator 232 Private loop 229 private loop 229 private loop initiators 232 Private Target 232 private target 233 Problem Log 84 Problem Notification 85 pSeries attachment requirements 12 pSeries FC host 11 pseud
Sun FC attachment 30 switch cascading 8 Switch registration 309 T T_Port 264 T_Ports 264 Target Hosts 144 Technology Pack 17 Telnet 275 TFTP 266 TimeOutValue 53, 64 Tivoli SANergy 364 TL_Port 235, 243 TL_Ports 235 TL_Ports zoning 236 TL-CFG 244 TL-Cfg 235 topology 167 Track Format 109 track format 108 Translation Entries List 237 translation entries list 236 translative device addresses 236 trivial file transfer protocol 266 trunk ports 264 Windows NT 4.
388 Implementing an Open IBM SAN
Implementing an Open IBM SAN Featuring the INRANGE Portfolio (0.5” spine) 0.475”<->0.
Back cover ® Implementing an Open IBM SAN Featuring the INRANGE Portfolio Discover the latest additions to the IBM SAN family Enhance your skills while using an easy-to-follow format Grow with the new technology "Do everything that is necessary and absolutely nothing that is not."' In this IBM Redbook, we have tried to consolidate as much of the critical information as possible while covering procedures and tasks that are likely to be encountered on a daily basis.
