User manual IBM IBM I ON A POWER BLADE READ-ME FIRST 7-14-2010

[. . . ] IBM iTM on a Power Blade Read-me First Vess Natchev Cloud | Virtualization | Power Systems IBM Rochester, MN vess@us. ibm. com July 14th, 2010 1 This "read-me first" document provides detailed instructions on using IBM i on a Power blade. It covers prerequisites, supported configurations, preparation for install, hardware and software install, firmware updates and post-install tasks such as backups. The document also contains links to many additional information sources. Table of Contents 1. Overview of I/O concepts for IBM i on blade 1. 2. [. . . ] When the tape drive is assigned to an IBM i partition in IVM, IVM will automatically create a new Virtual SCSI server adapter in VIOS and a Virtual SCSI client adapter in that IBM i partition. IVM will do so for each tape device assigned to IBM i. It is possible to manually add a new pair of Virtual SCSI server and client adapters for tape virtualization and map the tape drive to IBM i, as described in sections 4. 1. 1 and 4. 1. 2. It is also possible to map a tape drive to an already existing Virtual SCSI server adapter in VIOS used for 34 disk or optical virtualization. However, the recommended approach to making tape drives available to IBM i is to use IVM. 4. 11. Fibre Channel over Converged Enhanced Ethernet (FCoCEE) In October 2009, IBM announced the availability of FCoCEE for IBM i, AIX and Linux workloads on Power blades. FCoCEE is a new industry standard that allows a single converged adapter in a server to send and receive both Fibre Channel (FC) and Ethernet traffic over a 10-Gb Ethernet (10GbE) fabric. The benefits of this technology are: · · · Fewer adapters per server Fewer switches and therefore lower power consumption in the datacenter A single converged fabric for both FC and Ethernet traffic, enabling lower infrastructure costs and simplified management FCoCEE works by encapsulating a FC frame in a new type of Ethernet frame. The standard also allows for quality of service (QoS) prioritization for both FC and Ethernet traffic. For a complete general overview of FCoCEE, consult the Fibre Channel Industry Association's Website at http://www. fibrechannel. org. All Power blades support FCoCEE through a Converged Network Adapter (CNA) ­ the QLogic 2port 10 Gb Converged Network Adapter (CFFh) for IBM BladeCenter (P/N 42C1830). For IBM i the CNA is owned by VIOS, as with other blade server expansion adapters. VIOS then provides disk and network virtualization to any IBM i, AIX and Linux client LPARs through the CNA. Note that the converged adapter does not have both FC and Ethernet ports. Instead, the adapter sends and receives both 8Gb FC and 10Gb Ethernet traffic through its two 10GbE ports using FCoCEE frames. Traffic from the CNA is routed to a 10GbE passthrough module in I/O bays 7 and/or 9, and from there to a separate FCoCEE-capable top-of-the-rack (TOR) switch outside the BladeCenter H. The TOR switch is capable of routing both types of traffic ­ 10GbE and 8Gb FC ­ to separate networks connected to it. The following diagram illustrates the FCoCEE configuration for Power blades in BladeCenter H: 35 Note that Power blades support only Virtual SCSI (VSCSI) disk virtualization through FCoCEE. N_Port ID Virtualization (NPIV) via FCoCEE is not currently supported. 10GbE network connectivity for Power blades is supported. FCoCEE is not supported in BladeCenter S for any blade servers, including Power blades. Requirements for FCoCEE The following minimum hardware and software requirements must be met in order to use FCoCEE for IBM i on Power blades: · · · · · · · IBM BladeCenter JS12, JS22, JS23, JS43, PS700, PS701 or PS702 One QLogic 2-port 10 Gb Converged Network Adapter (CFFh) for IBM BladeCenter per blade One 10Gb Ethernet Passthrough Module for IBM BladeCenter (P/N 46M6181) One supported TOR FCoCEE switch. [. . . ] IBM hardware products are manufactured from new parts, or new and serviceable used parts. Any performance data contained in this document was determined in a controlled environment. Actual results may vary significantly and are dependent on many factors including system hardware configuration and software design and configuration. Some measurements quoted in this document may have been made on development-level systems. [. . . ]