SANnet II 200 SCSI Array Technical Product Description Guide January 2003 83-00002947 Revision A 1
Copyright Copyright 2001-2003 – Dot Hill Systems Corp. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, translated, transcribed, or transmitted, in any form or by any means – manual, electric, electronic, electromechanical, chemical, optical, or otherwise – without prior explicit written permission of Dot Hill Systems Corp., 6305 El Camino Real, P.O. Box 9000, Carlsbad, CA 92009-1606. Trademarks The Dot Hill Systems Corp.
Contents TABLE OF CONTENTS ............................................................................................................................. 3 INTRODUCTION ........................................................................................................................................ 4 KEY FEATURES ........................................................................................................................................... 4 ARCHITECTURE FUNDAMENTALS ........................
Introduction The SANnet II 200 SCSI array is a robust Ultra160 SCSI storage system and worthy successor to the well-established SANnet 3300 series arrays. Fast performance, high availability and simple manageability are condensed into a flexible, versatile, rugged and cost-effective package. Each SANnet II 200 SCSI array holds twelve 1-inch high disks in a 2U package and is available with dual redundant RAID controllers, a single RAID controller, or without controllers for use as an expansion unit.
· Unified graphical user interface (GUI) provides intuitive configuration, management and reporting for every Dot Hill SANnet II series array in your environment. · Support for a broad range of RAID levels: 0, 1, 0+1, 3, 5, 3+0 and 5+0 assures flexible performance and protection. · Dynamically expand capacity and performance by adding disks without interrupting storage operations or attached servers.
FIGURE 3: Typical Sequence of Installation Steps Reliability, Availability and Serviceability Comprehensive reliability, availability and serviceability (RAS) are assured within SANnet II 200 SCSI arrays. Proven low voltage differential (LVD) signaling across all disk and host SCSI buses keeps data transfers reliable. Flexible support for RAID levels 0, 1, 0+1, 3, 5, 3+0 and 5+0, plus global and dedicated spare disks provide powerful choices for protecting valuable information.
Performance SANnet II 200 SCSI arrays provide very high levels of performance within a compact package. The use of Ultra160 SCSI technology throughout for disks and servers results in performance approaching that of storage area networks (SAN) from a direct attached storage (DAS) solution. Cache memories of up to 1GB per array with support for 15,000RPM disks and flexible choices of hardware-based RAID level protection ensure optimum performance.
Architecture SANnet II 200 SCSI arrays use modular design principles. Essentially, each array is a collection of independent modules that operate cooperatively, providing all the necessary array functions. The foundation of every array is the enclosure module. Enclosure Modules The primary building block of any SANnet II 200 SCSI array is its enclosure. It consists of a compact metal chassis with internal controller mid-plane and disk mid-plane.
With redundant controllers, the system can use the controllers in active/passive or active/active (recommended) mode. When a controller is configured as active/passive, only a single controller is used to process I/O, while the other controller passively stands by in case of failure. When configured as active/active, both controllers actively process I/O. Each time a host writes information to one controller, it is automatically copied to the other for protection.
mode. The 146GB disk is available in speeds of 10,000RPM only at this time, but the other capacities are available in 10,000RPM or 15,000RPM speeds. Disks FRUs are hot-swappable or hot-serviceable depending their logical configuration – i.e., depending on whether the disks are configured to provide redundant RAID protection or not. The disks are not slot dependent.
The cooling fans within each FRU operate as matched pairs with tachometer-controlled spindles and position-synchronized blades for maximum efficiency. Every cooling fan operates from a common power bus so the failure of a power supply will not interrupt the operation of any fans, even those in the same FRU as the failed power supply. To remove a power/cooling module, turn the thumbscrew collar at the upper right corner to the left until loose then lower the handle down and pull the FRU out of the enclosure.
To remove an EMU, loosen the small thumbscrews at each end and then remove the EMU by pulling it out from the enclosure. To install one, slide the EMU into the enclosure until it stops and tighten the thumbscrew until secure. To maintain optimum cooling, avoid operating a SANnet II 200 SCSI array for more than a few minutes with an EMU removed. SCSI Module The SCSI connections to hosts and between enclosures are made at the SCSI modules (i.e., I/O modules).
Termination Module The SANnet II 200 SCSI array terminator board is a field-serviceable FRU that supplies the SCSI termination throughout a base enclosure. It is hot-serviceable but not hot-swappable, so replacing one will require planned downtime. The termination module is essentially hidden between controllers, so it requires the removal of controllers before it can be replaced.
Features Controller Optimization SANnet II 200 SCSI array controller operations are highly automated to minimize management workloads and are self-tuning to optimize performance. Powerful advanced features are supported, such as multi-threading, segmentation, read-ahead, write-through, write-back, queuing and sorting. Array users can adjust controllers to more closely match their particular environment by selecting between optimization for random or sequential access patterns.
The surviing controller of a fail-over process always becomes the primary controller. Replacing the faulty controller will result in an automatic fail-back process, where the array restores a redundant controller configuration. The new controller will become the secondary controller until the next reset or power cycle of the array, at which time the original primary/secondary arrangement will be restored. A particular logical drive must be assigned to either the primary or secondary controller.
amount of read cache memory is filled or sequential requests are no longer detected. If available read cache memory fills and sequential read requests continue, the controller will flush write cache content and allocate additional memory to read operations automatically. Since read cache contents are identical to information already on disks, read cache memory is never synchronized between redundant controllers.
Spare Disks A SANnet II 200 SCSI array disk can be assigned as a data disk or a spare disk. When configured as a spare, it can serve as a local spare to one specified logical drive or as a global spare available to all logical drives. More than one spare disk can exist for a particular logical drive and even combinations of local and global disks. It is a good practice to ensure every logical drive is protected by at least one spare disk, even if many logical drives are sharing spare disks.
Management Configuration Dot Hill’s SANscape software is a Java technology-based software program that combines storage configuration, maintenance and monitoring tools into a single, easy-to-use package. SANscape software provides centralized administration of SANnet II 200 SCSI storage systems across existing local and wide area networks (WANs). It greatly simplifies storage management and reduces its administration costs.
Monitoring The main SANscape window provides status at a glance for all networked array devices. It indicates the status of the devices and logical drives connected to servers monitored by the SANscape Console. The devices include host adapters, array controllers, disk storage enclosures, physical disk drives and other SCSI devices. The window’s tree structure offers detail for all devices connected to each server.
When the console receives any event, it logs it into SANscape’s Event Log file, EVENTLOG.TXT, and displays it in the Event Log window. Also, if the event occurs on a server, the notification of the event is sent to that server’s operating environment/system event log. On an NT server, it would go to the NT event log.
In-Band & Out-of-Band Management SANscape’s out-of-band storage management capability enables you to monitor and manage Dot Hill arrays over the network using TCP/IP. Unlike in-band storage management (the standard method of storage management for storage), which requires the SANscape Agent to be running on the server that is physically attached to the storage, out-of-band storage management does not require the SANscape Agent to be running on the server that is physically attached to the storage.
Event Reporting You can use Dot Hill’s SANscape to run as a background service to report events to specified email addresses.
Firmware Updates Fast and easy download of new versions of controller and disk firmware is possible using the SANnet II 200 SCSI array management suite. The Dot Hill CLI and SANscape both support controller updates over in-band SCSI connections and out-of-band Ethernet connections. SANscape provides updates for disks over in-band SCSI connections. When firmware updating is performed on a redundant controller array, firmware is loaded onto both controllers without interrupting operations.
Detailed Specifications RAID RAID levels 0, 1, 0+1, 3, 5, 1+0, 3+0 and 5+0. Enhanced RAID Levels supported with logical volume implementation. Maximum number of logical drives 8 RAID level dependency to each logical drive Independent.
Supported memory type SDRAM memory for enhanced performance. Fast page memory with parity for enhanced data security. ECC-protected. Read-ahead operation Intelligent dynamic read-ahead operation for sequential data accessing Multi-threaded operation Yes Scatter/gather supported Supported I/O sorting Supported. Optimized I/O sorting for enhanced performance Variable stripe size RAID5: Optimization for random I/O (32k), optimization for sequential I/O (128k), user selectable.
Head: 64,127,255 or variable Cylinder: <1024, <32784,<65536 or variable Drive motor spin-up Supported The controller will send spin-up (start unit) command to each drive at 4 sec. intervals.
No single-point-of-failure Supported Automatic engagement of replacement controller Supported Dynamic cache memory allocation Supported Cache memory is dynamically allocated, not fixed. Environment management Supported SAF-TE, S.E.S., ISEMS (I2C interface), and onboard controller voltage/temp monitor are all supported in both single and redundant controller mode. In the event of controller failure, services can be taken over by the existing controller.
Drive S.M.A.R.T. support Supported Clone failing drive Supported Users may choose to clone data from a failing drive to a backup drive manually User Interface RS-232C terminal Supported terminal modes: ANSI, VT-100, ANSI Color. Provides menudriven user-friendly text-based interface.
Performance Parameters Parameter Range RAID level JBOD, N-RAID, RAID 0, RAID 1, RAID 1+0, RAID 3, RAID 3+0, RAID 5, RAID 5+0 Rebuild priority Low, Normal, Improved, High Verification on write Enabled or disabled Max tag count 1-128 or disabled Max queued I/O count 32-1024 or auto LUNs per SCSI ID Up to 32 Number of host-LUN connections 1-64 Tag per host-LUN connection 1-256 Optimization for random/sequential RAID 3 Random - 4K Block Size RAID 3 Sequential - 16K Block Size RAID 5 Random -
Glossary active termination, regulated Terminates the SCSI bus with a series of resistors tied to +5 volts. The terminator is labeled Regulated but is often referred to as an Active Terminator. active-active controllers A pair of components, such as storage controllers in a failure-tolerant RAID array that share a task or set of tasks when both are functioning normally. When one component of the pair fails, the other takes the entire load.
fast write Allows disk write commands to be safely acknowledged to the host before the data is actually written to the disk media. This can be enabled/disabled through the storage management software. fast/wide SCSI Data transfer rate of 20 MB/sec. Wide devices can be connected to a standard SCSI interface but the extra data lines need to be terminated.
hot spare or hot sparing A drive in a RAID 1 or RAID 5 configuration that is held in reserve to replace any other drive that fails. After a reconstruction, the hot-spare drive is returned to the standby status. hot-serviceable The ability to remove, replace or add a device while power is still applied but all I/O processes are suspended. hot-swappable The ability of a field-replaceable unit (FRU) to be removed and replaced while the array remains powered on and operational.
commission to complete a repair process. N port A Fibre Channel port in a point-to-point or Fabric connection. NVRAM Non-Volatile Random Access Memory. A memory unit equipped with a battery so that the data stays intact even after the main power had been switched off. out-of-band Refers to the connections and devices that are not in the data path. parity Additional information stored along with the data that allows the controller to reconstruct lost data on RAID 3 or 5 LUNs if a single drive fails.
devices and servers providing accelerated data access. SCA Single connector attachment. A SCSI disk connector technology coinvented by Sun Microsystems. The SCA provides all SCSI, power, and control signals in a single connector, and enables easy servicing and highly reliable, pluggable disk drives. SCSI Small Computer Systems Interface. An industry standard for connecting disk and tape devices to a workstation.
independent disk spindles. The amount of data written on each disk before moving to the next drive is the stripe width. terminator A part used to end a SCSI bus. Terminators prevent energy from reflecting back into a cable plant by absorbing the radio frequency signals. throughput A measure of sequential I/O performance, quoted in MB/sec. See IOPS. Ultra160 SCSI LVD Ultra 3 SCSI command set plus a raw data rate of 160 MB/sec.