US007191286B2 (12) United States Patent (10) Patent No.: US 7,191,286 B2 Forrer, Jr. et al. (45) Date of Patent: Mar. 13, 2007 (54) DATA REDUNDANCY IN INDIVIDUAL HARD (56) References Cited DRIVES U.S. PATENT DOCUMENTS (75) Inventors: Thomas R. Forrer, Jr., Round Rock, 6,028,725 A * 2/2000 Blumenau ...................... 360/8 TX (US); Jason Eric Moore, Austin, 2004/0179386 A1* 9/2004 Jun ............................ 365,145 TX (US); Abel Enrique Zuzuarregui, Paige, TX (US) * cited by examiner (73) Assignee: International Business Machines Primary Examiner Reginald Bragdon Corporation, Armonk, NY (US) Assistant Examiner Shawn Gu (74) Attorney, Agent, or Firm—Duke W. Yee; Diana R. (*) Notice: Subject to any disclaimer, the term of this Gerhardt; James O. Skarsten patent is extended or adjusted under 35 U.S.C. 154(b) by 366 days. (57) ABSTRACT (21) Appl. No.: 10/809,589 A method and system are disclosed for recovering lost data with redundancy in an individual hard drive. A mirroring (22) Filed: Mar. 25, 2004 type process is used in a single hard drive to maintain a backup copy of all data stored on that hard drive. The hard (65) Prior Publication Data drive maintains two copies of the stored data on different US 2005/0216657 A1 Sep. 29, 2005 storage media segments or partitions defined by separate read/write heads. In the event that a hard drive media defect (51) Int. C. or error occurs with respect to any of the data stored on a F6F 12/00 (2006.01) segment defined by one head, that data can be restored from (52) U.S. Cl. ........................ 711/114; 711/112: 711/150 the segment defined by the other head(s). (58) Field of Classification Search ..................... None See application file for complete search history. 19 Claims, 2 Drawing Sheets INTIAL POWER ON-DRIVE OPERATES IN NON-RAID MODE OF OPERATION 300 MODE SELECT SENT TO DRIVE FOR RAID MODE OF OPERATION DRIVE READES PHYSICAL DISK SURFACE FOR RAID OPERATION DRIVE IS IN RAID MODE OF OPERATION POWER CYCLE MODE SELECT SENT TO DRIVE FOR 310 NON-RAID MODE OF OPERATION DRIVE OPERATES IN NON-RAID MODE OF OPERATION 316 U.S. Patent Mar. 13, 2007 Sheet 1 of 2 US 7,191,286 B2 FIG. I. 102 FIG. 24 201 SYSTEM BUS 200 MEMORY A1 208 N CONTROLLER/ I/O BRIDGE 210 CACHE 214 216 PC BUS 209 LOCAL PCBUS MEMORY BRIDGEas Heeze NETWORK 212-1E, ADAPTER - 222 218 220 PCBUS PCBUS BRIDGE —I 226 SCSI HOST BUS ADAPTORY-231 GRAPHICS ADAPTER 230 HARD DISK 232 PCBUS PC BUS BRIDGE HD E 228 224 U.S. Patent Mar. 13, 2007 Sheet 2 of 2 US 7,191,286 B2 250 252 258 254 266 PROCESSORKOcACEBRDGEHOST/PC <-SeyMAIN ADAPTERAUDIO FIG. 2B BUS SCS 256 EXPANSION AUDIO/ LAN GRAPHICS HOST BUS ADAPTER BUS ADAPTER VIDEO ADAPTER INTERFACE ADAPTER 262 260 264 268 269 DISK 276 KEYBOARD AND TAPE MoSEADAPTE MODEM | MEMORY 278 C CD-ROM 28O 270 272 274 DVD 282 FIG. 3 302 INTIAL POWER ON-DRIVE OPERATES IN NON-RAID MODE OF OPERATION 300 MODE SELECT SENT TO DRIVE FOR RAID MODE OF OPERATION DRIVE READES PHYSICAL DISK SURFACE FOR RAD OPERATION DRIVE IS IN RAID MODE OF OPERATION POWER CYCLE MODE SELECT SENT TO DRIVE FOR 310 NON-RAID MODE OF OPERATION DRIVE OPERATES IN NON-RAID MODE OF OPERATION 316 POWER CYCLE US 7,191,286 B2 1. 2 DATA REDUNDANCY IN INDIVIDUAL HARD drive. The hard drive maintains two copies of the stored data DRIVES on different storage media segments or partitions defined by separate read/write heads. In the event that a hard drive BACKGROUND OF THE INVENTION media defect or error occurs with respect to any of the data stored on a segment defined by one head, that data can be 1. Technical Field restored from the segment defined by the other head(s). The present invention relates generally to data recovery in computer systems. More specifically, the present invention BRIEF DESCRIPTION OF THE DRAWINGS provides an improved method and system with redundancy for recovering lost data in individual hard drives. 10 The novel features believed characteristic of the invention 2. Description of Related Art are set forth in the appended claims. The invention itself, Hard disk drives (“hard drives') are the main data storage however, as well as a preferred mode of use, further objec devices for most computers or processors. Typical hard tives and advantages thereof, will best be understood by drives contain a number of hard disk platters (“disks”) reference to the following detailed description of an illus coated with magnetic material for storing data in magnetized 15 trative embodiment when read in conjunction with the form. The disks are affixed to a spindle that spins them in accompanying drawings, wherein: unison at a constant rate. An actuator arm moves one or more FIG. 1 is a pictorial representation depicting a data read/write heads radially across the disks to retrieve or store processing system in which the present invention may be the magnetized data. The hard drive’s components (disks, implemented in accordance with a preferred embodiment of spindle, actuator arm, read/write heads, associated motors the present invention; and electronics, etc.) are enclosed within a head disk assem FIG. 2A is an example block diagram illustrating hierar bly (HDA). chically organized internal components of a server-type data The HDA protects the hard drive's components from processing system that may be used to implement the contamination by dust, condensation, and other sources. present invention; Contamination of a hard drive's disk or read/write head can 25 FIG. 2B is an example block diagram illustrating hierar result in disk failure, head crash and/or an unrecoverable chically organized internal components of a client-type data loss of data. Also, data loss can be caused by other hard drive processing system that may be used to implement the problems, such as electronic malfunctions, physical shocks, present invention; and worn out components, improperly manufactured disks, etc. FIG. 3 is an example flowchart illustrating a process for Notably, unrecoverable data error rates for the existing 30 recovering lost data with redundancy in an individual hard generation of hard drives have remained essentially drive, which can be implemented in accordance with a unchanged for the past 10 years. However, the storage preferred embodiment of the present invention. capacity of hard drives has increased at a Substantial rate (e.g., doubling on average every 18 months). In fact, the DETAILED DESCRIPTION OF THE typical storage capacity of the existing generation of hard 35 PREFERRED EMBODIMENT drives ranges from about 8 GB to 300 GB (up from about 20 MB just a few years ago). Consequently, the probability that With reference now to the figures, FIG. 1 depicts a unrecoverable data errors (“hard errors') can occur during pictorial representation of a data processing system in which hard drive read/write operations has increased significantly the present invention may be implemented, in accordance along with the Substantial increase in storage capacity. 40 with a preferred embodiment of the present invention. For One hardware approach to solving the problem of this exemplary embodiment, data processing system ("com increased hard errors in hard drives is Redundant Arrays of puter) 100 includes system unit 110, video display terminal Independent Disks (RAID). The RAID approach connects a 102, keyboard 104, mouse 106, and storage devices 108, plurality of hard drives together and treats them as a single, which may include one or more floppy drives, a hard drive, logical unit. Consequently, a computer sees the plurality of 45 a CD-ROM drive, and/or other types of permanent and drives as one, large hard drive that can be formatted and removable storage media. Additional input devices may be partitioned similar to that of a smaller, individual drive. The included with computer 100. Such as, for example, a joy redundancy thus provided by the multiple hard drives of the Stick, touch pad, touch screen, trackball, microphone, and RAID approach enhances the recovery of lost data due to the like. hard errors, and also increases the performance of the 50 Computer 100 can be implemented using any suitable computer system involved. As such, the multiple hard drive computer, such as, for example, an IBM RISC/System 6000 RAID approach provides an acceptable solution for large, computer system or IntelliStation computer, which are prod desk-side computers and server applications. However, the ucts of International Business Machines Corporation in use of multiple hard drives for redundancy is unsuitable for Armonk, N.Y. Although the depicted representation shows a those relatively small computer systems or products where 55 computer, other embodiments of the present invention may the primary design considerations are to minimize space, be implemented in other types of data processing systems, weight and cost. Such as a network computer, and the like. In any event, FIG. Therefore, it would be advantageous to have an improved 1 is intended as an example and not as an architectural method and system for recovering lost data with redundancy limitation for the present invention. in an individual hard drive. 60 With reference now to FIG. 2A, a block diagram is shown that depicts hierarchically organized internal components of SUMMARY OF THE INVENTION a server-type data processing system, which may be used to implement the present invention. For this exemplary The present invention provides a method and system for embodiment, data processing system 200 may be a sym recovering lost data with redundancy in an individual hard 65 metric multiprocessor (SMP) system including a plurality of drive.