JP2006114071A - Defect management of hdd by variable index architecture - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a sector level defect management technique which performs a local track spare processing to a headerless format AV HDD which uses a variable index write technique. <P>SOLUTION: A data stream inputted by a HDD200 is organized into at least one cluster which has a data block and a spare data block (502). With the use of the variable index write technique, each data block in a cluster is written in a corresponding sector of a track to which the head encounters when the sector does not have a defect (506). When the sector has a defect, the sector of the track is skipped in writing (507). The number of the spare data blocks in the cluster part corresponding to one track is reduced by the number of the sectors skipped. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a hard disk drive (HDD) system. In particular, the present invention relates to a system and method for managing defects in a streaming file HDD system optimized for writing and reading audio and / or visual (audio visual or AV) data.

  A hard disk drive (HDD) generally has a plurality of disks. Each disk has a plurality of concentric data tracks. These data tracks have a common index representing the start of each track and a plurality of physical sectors. Each physical sector has a sector number or address. An HDD having data that is formatted and stored in a physical sector of a fixed length (ie, a fixed number of bytes) on the disk of the HDD is called a fixed block architecture (FBA) HDD.

  Sector level defect management is performed by the HDD if there are large defects in the sector that significantly affect the readability of the disk. Data affected by such defects is remapped to new sectors in the global spare sector pool in the conventional manner. The headerless recording format used by conventional HDDs provides logical block address (LBA) access technology. Logical block address (LBA) access technology allows any bad physical sector to be mapped to another sector that is part of the global spare sector pool.

  The HDD data processing workload in a computer data processing environment is characterized by a random access pattern of small data blocks. In contrast, HDD data processing workloads in audio and / or visual (AV) processing environments are characterized by interleaving and interleaving with small reads and writes that are frequently written but contain only a small amount of data. The large reads and writes of multiple streams that can be dominant. For example, interleaving allows a digital video recorder (DVR) to record and play back a television broadcast in real time or to record two television programs simultaneously. Using a global spare pool for sector-level defect management tends to adversely affect the performance of streaming data applications due to the extra time required to access the global spare sector pool There is.

  Another feature of AV disk drives is that the AV disk drive spends considerable time waiting for the disk to rotate before the index is detected by the read / write head before the next write can be made. is there. This state is called rotation waiting time. US patent application Ser. No. 10 / 227,494 by the same applicant (named “Method of Writing Streaming Audio / Visual Data to Disk Device”, filed on August 23, 2002 and incorporated herein by reference) The variable index recording technique such as that disclosed in US Pat. No. 5,849, can eliminate the adverse effects of rotational latency on the performance of streaming data applications. The variable index recording technique allows a write operation to be performed as soon as the recording head is placed on a predetermined track. Data is written to the first sector encountered in a given track by creating a virtual index at the location of the first encountered sector without waiting for the index (ie, start) of the physical track.

  FIG. 1 illustrates a typical variable index recording technique for writing data clusters of fixed length logical data blocks B0-B7 to physical sectors S0-S7 of a typical data track 100 of the HDD. As used herein, a “data cluster” is the basic unit of AV stream data and is an integer multiple of a full data track on the disk. Therefore, when the integer multiple is 1, one full track corresponds to one AV data cluster. Similarly, when the integer multiple is 4, four full tracks correspond to one AV data cluster. As described above, writing of all data is full track writing. The AV data cluster is further organized into integer fixed length data blocks, similar to the conventional FBA format. As a result, the number of data blocks in the AV data cluster is equal to the number of physical sectors in all the tracks that form the data cluster because the cluster is an integer multiple of the full track. Become. It should be understood that at least one of the data blocks in the cluster can include error correction information calculated from a group of AV data blocks.

  Returning to FIG. 1, each logical block B0-B7 is the size of one physical data sector. Blocks B0-B7 are written to data track 100 using a variable index recording technique such as disclosed in US patent application Ser. No. 10 / 227,494. The most efficient implementation of variable index recording technology is to use a “full track” data format. In this data format, read / write data is organized by a hard disk controller (HDC) to occupy the entire data track. Another advantage of the full track data format is that the rotational latency can be eliminated, reducing seek operations and improving overall storage throughput. Thus, as shown in the example of FIG. 1, since physical sector S6 is the first sector encountered after the head is placed on track 100, block B0 is written to physical sector S6. Next, block B1 is written to data sector S7. Thereafter, block B2 is written to data sector S0. The subsequent steps are as shown in the figure.

  Variable index recording techniques typically use additional track level error correction codes (ECC) such as parity sectors to handle large defects on the spot. However, it is desirable to have the ability to handle large soft errors in track level ECC and to allow hard errors to be handled in spare processing.

  What is needed is a sector level defect management technique that performs local track spare processing for headerless format AV HDDs that use variable index writing techniques.

  The advantages of the present invention can be obtained by a technique for writing data to a disk device. The disk device includes at least one disk and a head for writing data to each disk. Each disk has at least one data track. The head is positioned on the selected data track of the selected disk of the disk device. The selected data track has a first predetermined number of physical sectors. If the head encounters each physical sector and that physical sector is not defective, the second predetermined number of data blocks of the data stream is selected using a variable index write technique. Written to the corresponding physical sector of the track. At this time, the size of the data block is equal to the size of the physical sector. If the physical sector is defective, the physical sector of the selected data track is skipped. When all data blocks of the data stream are written to the selected data track, the third predetermined number of physical sectors of the selected data track is skipped. The third predetermined number of skipped physical sectors of the selected data track has a second predetermined number of data sectors from the first predetermined number of physical sectors. Equal to the block minus the sum of the number of skipped defective physical sectors in the data track.

  The size of each data cluster is equal to an integer multiple of the first predetermined number of physical sectors, and a portion of each data cluster corresponds to a single data track; The data stream can be organized into at least one data cluster to include a second predetermined number of data blocks. If the physical sector is free of defects when the head encounters each physical sector, the second predetermined number of data blocks in the selected portion of the selected data cluster may use variable index writing techniques. Use to write to the corresponding physical sector of the selected data track. When all the data blocks of the selected portion of the selected data cluster have been written to the selected data track, the third predetermined number of physical sectors are skipped. Each portion of the data cluster corresponding to a single data track may include at least one error correction information data block.

  If the size of the selected data cluster is greater than the size of a single data track, the second predetermined number of data blocks in the selected portion of the selected data cluster is a variable index A write technique is used to write to the corresponding physical sector of the selected data track. However, the condition is that the physical sector is not defective when the head encounters each physical sector of the selected data track. The head is then positioned on a second selected data track having a first predetermined number of physical sectors. If the physical sector is free of defects when the head encounters each physical sector, the second predetermined number of data blocks of the second selected portion of the selected cluster is a variable index writing technique. Is written to the corresponding physical sector of the second selected data track. If the physical sector is defective, the physical sector of the second selected data track is skipped. When all data blocks of the second selected portion of the selected data cluster have been written to the second selected data track, the fourth of the second selected data track A predetermined number of physical sectors are skipped. The fourth predetermined number of skipped physical sectors of the second selected data track has a second predetermined number in number from the first predetermined number of physical sectors. Equal to the sum of the data block plus the number of skipped defective physical sectors in the second data track.

  The present invention uses metadata associated with each selected data track to determine whether a physical sector is defective. The metadata associated with each selected data track includes the number of physical sectors on the selected track, the number of parity sectors on the selected track, and the number of spare sectors on the selected track. Contains information on at least one. In addition, the metadata associated with each selected data track includes the stream handle of the data stream, whether the selected track is free or occupied, and the physical start of the selected track Information about at least one of the starting sector of the selected data track in relation to and the location of all defective sectors on the selected track.

  According to the present invention, there is provided a sector level defect management technique for performing local track spare processing for a headerless format AV HDD using a variable index writing technique.

  The present invention is illustrated by way of example and not limitation in the accompanying drawings. In the accompanying drawings, the same reference numerals denote the same parts.

  According to the present invention, there is provided a headerless format HDD which performs sector level defect management using variable index writing technology and track local spare processing technology (hereinafter referred to as floating spare area spare processing technology). Is done. Unlike conventional sector level defect management techniques, bad sectors have a fixed relationship with the physical index on the track, and spare sectors have a fixed relationship with the variable index location.

  The functional block diagram of FIG. 2 illustrates a typical HDD 200 that utilizes floating spare area spare processing techniques for sector level defect management according to the present invention. HDD 200 includes disk drive actuator arm 201, read / write (R / W) head 202, disk 203, head read / write processor 240, servo controller 241, control logic 242, buffer memory 243, and interface A control device 244 is provided. Disk 203 rotates clockwise as indicated by arrow 250. The disk drive actuator arm 201 positions the read / write head 202 on a predetermined data track of the disk 203. This is done by moving the read / write head 202 to a predetermined data track, generally across the disk 203, as indicated by arrows 251 and 252 in the radial direction. On the disk 203, representative selected data tracks including physical sectors S0-S7 are shown. Here, it should be understood that disk 203 includes a plurality of concentric data tracks, and only representative selected data tracks are shown for clarity. It should also be understood that the HDD 200 can comprise a plurality of disks, of which only the disk 203 is shown in FIG. 2 for clarity of explanation.

  In a typical preferred embodiment, each physical sector number can be determined from the track number and servo positioning information. The present invention also uses an “indexless” format that does not require a physical index mark that identifies the start of a track. That is, even if the first physical sector number S0 exists on each track, the purpose of this physical sector S0 is not to specify the start of the data cluster recorded on that track. When an AV data cluster that includes logical blocks B0-B7 is written to an empty track, the first data block B0 is written to the first physical sector that the read / write head 202 encounters. This minimizes the rotational latency that occurs when the read / write head waits for a track index or a particular physical sector. A suitable and typical indexless format is disclosed in US patent application Ser. No. 10 / 227,494.

  When a write operation begins, the read / write head 202 is moved to a data track previously identified as a track with free space for the control logic 242. “Seek” to a predetermined track is a signal sent to the servo controller 241 to move the head 202 and the actuator arm 201 that read pre-recorded servo positioning information from the disk and a predetermined distance in a predetermined direction Performed in a well-known manner by the communicating head read / write processor 240.

  During a seek operation, a data stream from an AV data source 245 such as a digital TV or DVR is transferred through the interface controller 244. The data stream from the host is typically transmitted as a series of data blocks having a starting address and length. In some streaming systems, such as an object or file-based stream, where the stream command consists of an object identifier, data set, offset address, and length, the start address is the base address of the identified object And the offset address. When there are multiple streams, the AV data source 245 identifies data belonging to each stream, such as by using an object identifier.

  Control logic 242 organizes the received data stream into clusters having sequential data blocks. The organized cluster is then loaded into the buffer memory 243. The buffer 243 temporarily stores the write data input from the AV data source 245. In general, the size of the buffer 243 is larger than one cluster. A particular series of data tracks that form a cluster on disk 203 is selected from a series of clusters on disk 203 that are recognized as being free and is preferably started on cluster boundaries. In general, a free area table includes a series of free clusters.

  The write operation can begin before the buffer 243 has entered all the data for the cluster. Thus, when the read / write head 202 is positioned on a given track and fits in the write position, the first block B0 of the cluster uses a variable index recording technique to encounter the first physical sector encountered. Is written to. For example, physical sector S6 is shown in both FIG. 1 and FIG. 2 as the first physical sector encountered by read / write head 202, and is therefore the first sector of the cluster. Thereafter, the remaining cluster data blocks B1-B7 are written sequentially to physical sectors S7 and S0-S5 until all of the clusters are written to the entire track. FIG. 2 shows a state where block B7 has not yet been written because the physical sector S5 has not yet passed under the read / write head 202. FIG.

  According to the floating spare area spare processing technique of the present invention, the spare sector is preferably located at the end of the user data on the track. Thus, the spare sector has a fixed relationship with the variable index location, while the bad sector has a fixed relationship with the physical index on the track. Consider, for example, a data track with 400 physical sectors starting at physical sector S0 and a data cluster with a length of 397 data blocks starting with data block B0 and having 3 spare sectors SP1-SP3. To do. In general, spare sectors SP1-SP3 have a fixed relationship with variable index locations. In contrast, every bad sector has a fixed relationship with the index of the track. During a typical data cluster data write operation to the data track, the first user data block B0 is written to the physical sector S301 of the data track, for example. In this case, since the spare sectors SP1 to SP3 are the physical sectors S298, S299, and S300, writing is performed in the 297 physical sectors after the physical sector S301. The starting sector location of the adjacent data track may be physical sector S100. In this case, the physical sectors S97, S98, and S99 are the spare sectors.

  FIG. 3 shows a generalized typical conventional track topology for a data track 300 having three spare sectors. Data track 300 starts at physical index I. The physical index I is generally part of the servo field. The physical index I is followed by data sectors S1, S2,. Here, n is the number of physical sectors of the track not including the spare sector and the parity sector. The data sector Sn is followed by three spare sectors SP1, SP2, SP3. Finally, a spare sector SP1-SP3 is followed by a parity sector P1. The virtual index is in physical sector S1.

  In a state where the physical sector S3 is defective, the conventional HDC reassigns the contents of the physical sector S3 to, for example, a physical sector corresponding to the spare sector SP1, and records the physical sector S3 as a BAD. FIG. 4 shows a case where the contents of the bad physical sector S3 of the track 300 are reassigned to the physical sector corresponding to the spare sector SP1 by the conventional method.

  According to the present invention, data is written using a variable index writing technique during a write operation, starting from the first data sector in which the write head is located. When a defective physical sector is encountered during a write operation, the defective sector is skipped. After skipping the defective physical sector, the write sequence is resumed immediately.

  FIG. 5 is a flowchart illustrating an exemplary basic floating spare area spare processing technique 500 for writing data to a disk according to the present invention. In step 501, a data stream is received from a host. In step 502, the data stream is organized into at least one data cluster having a size equal to an integer multiple of the number of physical sectors in the data track. Each portion of the data cluster corresponding to a single data track includes a data block, a spare data block, and an error correction information data block. The sum of the total number of data blocks, the total number of spare data blocks, and the total number of data blocks containing error correction information in each part of the data cluster is equal to the total number of physical sectors in the data track. . In the description of FIG. 5, a typical data cluster is selected to have as many blocks as there are physical sectors in two data tracks. That is, a typical data cluster has two parts, each corresponding to a single data track.

  In step 503, the head is positioned on the selected data track of the disk. In step 504, after the head is positioned on the selected data track, the first physical sector of the data track is encountered. In step 505, it is determined whether the first physical sector encountered on the data track is defective. If there are no defects, proceed to step 506 where the first data block of the selected portion of the data cluster is written to the first physical sector. At the start of this example, the selected portion of the data cluster is the first portion of the data cluster. Thereafter, the process proceeds to step 508. If it is determined in step 505 that the physical sector is defective, the process goes to step 507 where the physical sector is skipped. Note that each time a physical sector is skipped, one of the spare data sectors at the end of the current portion of the data cluster being written is reallocated to store the data block. . Thereafter, the process proceeds to step 508.

  In step 508, it is determined whether all the data blocks that form the selected portion of the data cluster have been written to the data track. If, in step 508, it is determined that all the data blocks forming the selected portion of the data cluster have not yet been written, go to step 509 where the next physical sector is encountered. Otherwise, go to step 505. There it repeats the process of determining if a physical sector is defective and writing the data block to the sector if the sector is not defective or skipping the physical sector if the sector is defective .

  If it is determined in step 508 that all the data blocks forming the selected portion of the data cluster have been written, the process proceeds to step 510 and skips by the number of remaining unused spare data sectors. . Next, the process proceeds to step 511. There, data blocks containing error correction information are written successively to the physical sectors encountered. Thereafter, the process proceeds to step 512 to determine whether all parts of the data cluster have been written. If all have not been written, return to step 503 and repeat the process for the current portion of the data cluster. If it is determined in step 512 that all parts of the data cluster have been written, the process ends in step 513.

  FIG. 6 is a diagram illustrating a typical track topology for a data track 600 having a spare sector followed by a parity sector when using the floating spare area spare processing technique according to the present invention. As shown in FIG. 6, the size of the data cluster is equal to the number of physical sectors in the data track 600. In FIG. 6, the write head is placed between the second half of the track 600, and the data sectors S1-Sn-3 are written before the physical index I. After physical index I, data sectors Sn-2, Sn-1 are written and then the BAD physical sector is skipped. Data sector Sn is written immediately after the BAD sector. The data sector Sn is followed by spare sectors SP2, SP3 and a parity sector P1.

  FIG. 7 is a diagram representing another example track topology for a data track 700 having a parity sector followed by a spare sector when using the floating spare area spare processing technique according to the present invention. As shown in FIG. 7, the size of the data cluster is equal to the number of physical sectors in the data track 700. Data track 700 does not include bad sectors, and this typical track topology includes two spare sectors SP2, SP3 and one parity sector P1. In FIG. 7, the write head is placed between the second half of the track 700 and data sectors S1-Sn are written. The parity sector P1 is written immediately after the sector Sn is written. The parity sectors P1 are followed by spare sectors SP2 and SP3. The topology of track 700 improves system performance, but requires parity data to be calculated before completing the writing of the last data sector Sn. The typical data track topology 700 writing technique is the typical basic floating spare shown in FIG. 5, except that steps 510 and 511 are interchanged to create a typical data track topology 700. Similar to Area Spare Processing Technology 500.

  Variable index recording typically uses metadata to store index location and occupancy information per track. In the present invention, the metadata is expanded to include all the information necessary to locate the defective spare sector in each track. For example, the defective physical sector information for each track is stored as a physical sector number from the index I. Information identifying the location of the spare sector need not be stored directly. This is because the spare sector can be calculated based on the number of spares allowed on the track and the number of defective sectors on the track. Thus, if the user sector can know the physical sector number and the virtual index location together with the sector defect information, it can be mapped to the physical sector (and vice versa).

  The sequential nature of streaming data maintains the performance of streaming data applications. On the other hand, the amount of RAM required to hold the metadata is limited. For each read stream, a buffer is used to hold metadata for the next N tracks in the stream. For example, if N = 20, streaming throughput is only affected at the 5% level because access to metadata is required every 20 data tracks. For the write stream, a similar buffer holds metadata for the next N free tracks. Thus, for the write stream as well, the impact of metadata on streaming data application performance is similarly negligible for N = 20 tracks.

  The disk device maintains and manages metadata associated with the track. Metadata common to a series of tracks includes the number of physical sectors per track (constant for each zone), the number of parity sectors on the track, and the number of spare sectors on the track. Other information maintained by the disk unit and specific to the individual track includes the associated stream handle, track free / occupied indicator (may be a reserved stream handle value), physical start of the track There is the starting sector of the data track, the location of all defective sectors on the track (eg, as a list or bitmap), etc. The metadata is modified in response to stream allocation events, such as stream creation, stream deletion, and stream modification, for example. Furthermore, if dynamic allocation is used during stream writing, the metadata is modified accordingly.

  When writing a data stream, the disk device allocates a series of tracks to hold the write data. Any allocation method may be used, such as a file allocation table (FAT) that can provide a series of empty tracks. Once a track is assigned, data is written sequentially to the assigned track. Before writing to a track, the disk device determines a series of defect-free sectors from the metadata associated with the track. When writing to a track, the disk device records the start data sector number of the first defect-free sector that can be used for writing data. If there are no defects in subsequent sectors on the track, writing to those sectors is performed. If there is a defect, the defective sector is skipped. After all data has been written to that track, the parity sector is written to the subsequent data sector. The contents of the parity sector are calculated before writing. Although the parity sector is referred to herein as a parity sector, any suitable error or erasure correction code may be used to calculate its contents. For example, a single parity sector may be calculated using simple parity. A pair of parity sectors may be calculated using the EVENODD code. Furthermore, a larger number of parity sectors may be calculated using a Reed-Solomon code.

  Once written to the track, all remaining non-defective sectors are assigned as spare sectors. Their location and number are determined from the number of sectors on the track and the number of defective sectors on the track. The metadata is then updated with the saved starting sector location, stream handle, and track free status. As already described, the metadata update process from a series of tracks may be combined into a single update operation. As a result, the impact on performance can be minimized.

  When a disk device reads a stream, it identifies a series of tracks associated with the stream. Once a track is identified, data is read sequentially from the track. Before reading from a track, the disk device identifies a series of defect-free sectors from the metadata associated with that track and the starting data sector number. When reading from a track, the disk device starts reading at the position of the first non-defective sector available for reading data. Subsequent sectors on the track are read if they are not defective and are not spare sectors. Otherwise, skip defective sectors and spare sectors. Data sectors are each stored in buffer locations according to their stream sector number (ie, the number from the starting data sector on the track). If all data sectors are read normally, the disk device has read normally from the track and can transfer data to the host. If none of the data sectors can be read, their contents are reconstructed from the remaining data and parity sectors. Once the operation is complete, it has been successfully read from the track and data can be transferred to the host.

  Compared to conventional methods, the present invention can save a considerable amount of metadata buffer size. In conventional headerless defect mapping, the RAM requirement is linearly proportional to the number of defects in the disk device and almost linearly proportional to the total number of sectors in the disk device. For the present invention, the metadata buffer size requirement is proportional to the number of defects allowed on a track and the number of tracks specified by the metadata.

  Although the foregoing invention has been described in some detail for purposes of clarity of understanding, it will be apparent that certain changes and modifications may be made without departing from the scope of the claims. Accordingly, this example should be considered as illustrative rather than limiting of the present invention. Also, the invention is not limited to the details described herein, but modifications may be made within the scope of the claims and the equivalents.

FIG. 3 is a diagram showing a typical variable index recording technique for writing a data cluster of fixed-length logical data blocks into a physical sector of a typical data track of the HDD. 1 is a functional block diagram illustrating a representative HDD that utilizes floating spare area technology for sector level defect management in accordance with the present invention. FIG. FIG. 2 shows a generalized typical conventional track topology for a data track having three spare sectors. It is a figure which shows the conventional reallocation which reallocates the content of the bad data sector of a track | truck to a spare sector. 4 is a flowchart showing a floating spare area spare processing technique for writing data to a disk according to the present invention. FIG. 7 shows an exemplary track topology for a data track having a spare sector followed by a parity sector when using the floating spare area spare processing technique according to the present invention. FIG. 7 is a diagram showing another example track topology for a data track having a parity sector followed by a spare sector when using the floating spare area spare processing technique according to the present invention.

Explanation of symbols

200: Hard disk drive (HDD),
202 ... read / write head,
203: Disc.

Claims (26)

A method of writing data to a disk drive having at least one disk and a head for writing data to each disk, wherein each disk has at least one data track,
Positioning the head over a data track having a selected first predetermined number of physical sectors of a selected disk of the disk drive;
When the head encounters each physical sector and the physical sector is not defective, the data block equal to the size of the second predetermined number of physical sectors of the data stream is used using variable index writing techniques. Writing to the corresponding physical sector of the selected data track;
If the physical sector is defective, skipping the physical sector of the selected data track;
When all data blocks of a data stream have been written to the selected data track, the second predetermined number of physical sectors from the first predetermined number of physical sectors of the selected data track. Skipping a third predetermined number of physical sectors equal to a predetermined number of data blocks minus the sum of the number of skipped defective physical sectors of the data track. A method of writing data to a disk drive.   Skipping the third predetermined number of physical sectors includes writing at least one error correction information data block as part of the second predetermined number of data blocks. The method according to claim 1.   3. The method of claim 2, wherein the at least one error correction information data block is continuously present before the first spare data block.   The method of claim 2, wherein the at least one error correction data block is continuously present after the last spare data block. At least one disk of the disk drive has a plurality of data tracks, each of the plurality of data tracks having the first predetermined number of data sectors;
The method further includes organizing the data stream into at least one data cluster, wherein the size of each data cluster is an integer multiple of the first predetermined number of physical sectors. And a portion of each data cluster includes a second predetermined number of data blocks corresponding to a single data track and
Writing the second predetermined number of data blocks comprises the step of selecting the selected portion of the selected data cluster when the head encounters each physical sector and the physical sector is not defective. Writing a second predetermined number of data blocks to a corresponding physical sector of the selected data track using a variable index writing technique;
The step of skipping the third predetermined number of physical sectors is that all data blocks of the selected portion of the selected data cluster have been written to the selected data track. 2. The method of claim 1, wherein the third predetermined number of physical sectors is skipped. Each portion of the data cluster corresponding to a single data track further includes at least one error correction information data block;
The step of skipping the third predetermined number of physical sectors includes the step of writing each error correction information data block of the data cluster to a physical sector assigned as a spare sector. The method of claim 5. The size of the selected data cluster is larger than the size of a single data track,
Writing the second predetermined number of data blocks includes selecting the selected data when the head encounters each physical sector of the selected data track and the physical sector is not defective. Writing the second predetermined number of data blocks of a selected portion of the cluster to a corresponding physical sector of the selected data track using a variable index writing technique;
The method further includes
Positioning the head on a second selected data track having a first predetermined number of physical sectors;
When the head encounters each physical sector and the physical sector is not defective, the second predetermined number of data blocks of the second selected portion of the selected cluster is changed to a variable index. Writing to a corresponding physical sector of the second selected data track using a writing technique;
If the physical sector is defective, skipping the physical sector of the second selected data track;
The second selected data track when all data blocks of the second selected portion of the selected data cluster are written to the second selected data track; Subtracting the sum of the second predetermined number of data blocks plus the number of skipped defective physical sectors of the second data track from the first predetermined number of physical sectors. Skipping a fourth predetermined number of physical sectors equal to
The method of claim 5 comprising: Each portion of the data cluster corresponding to a single data track further includes a fourth predetermined number of error correction information data blocks;
6. The step of skipping the third predetermined number of physical sectors includes writing at least one error correction information data block to a physical sector corresponding to a spare sector. The method described.   9. The method of claim 8, wherein the at least one error correction information data block corresponds to the first spare data block.   9. The method of claim 8, wherein at least one error correction data block corresponds to the last spare data block.   The method of claim 1, further comprising determining whether a physical sector is defective based on information contained in metadata associated with the selected data track.   The metadata associated with the selected data track includes the number of physical sectors of the selected track, the number of parity sectors on the selected track, and a spare sector on the selected track. 12. The method of claim 11, comprising information regarding at least one of the numbers.   The metadata associated with each selected data track includes the stream handle of the data stream, whether the selected track is free or occupied, and the physical of the selected track. The method of claim 11, comprising information regarding at least one of a starting sector of the selected data track related to starting and a location of all defective sectors on the selected track. . At least one data track, each data track having a first predetermined number of physical sectors; and
When each physical sector is encountered and the physical sector is free of defects, a variable index writing technique is used to generate a data block equal to the size of the second predetermined number of physical sectors of the data stream. If the corresponding physical sector of the selected data track is written to and the physical sector is defective, the physical sector of the selected data track is skipped, and the data stream of the selected data track is further skipped. When all the data blocks have been written, the data data from the first predetermined number of physical sectors of the selected data track to the second predetermined number of data blocks. A third value equal to the sum of the number of skipped defective physical sectors in the track plus the total. And head to skip the beforehand determined number of physical sectors,
A disk drive characterized by comprising:   When the head skips the third predetermined number of physical sectors, the head selects at least one error correction information data block from one of the second predetermined number of data blocks. 15. The disk drive of claim 14, further writing as a part.   The disk drive of claim 15, wherein the at least one error correction information data block is continuously present before the first spare data block.   The disk drive of claim 15, wherein the at least one error correction data block is continuously present after the last spare data block. At least one disk of the disk drive has a plurality of data tracks, each of the plurality of data tracks having the first predetermined number of data sectors;
The data stream is organized into at least one data cluster, wherein the size of each data cluster is equal to an integer multiple of the first predetermined number of physical sectors, and each data cluster A portion of which corresponds to a single data track and includes the second predetermined number of data blocks;
The head encounters each physical sector and writes the second predetermined number of data blocks of the selected portion of the selected data cluster to a variable index when the physical sector is free of defects. Using technology to write to the corresponding physical sector of the selected data track, and all data blocks of the selected portion of the selected data cluster are stored in the selected data track. The disk drive of claim 14, wherein when written, the head skips the third predetermined number of physical sectors. Each portion of the data cluster corresponding to a single data track further includes at least one error correction information data block;
19. The disk drive according to claim 18, wherein the head writes each error correction information data block of the data cluster to a physical sector assigned as a spare sector. The size of the selected data cluster is larger than the size of a single data track,
When the head encounters each physical sector of the selected data track and the physical sector is free of defects, the head is the second pre-selected portion of the selected portion of the selected data cluster. Writing a determined number of data blocks to the corresponding physical sector of the selected data track using a variable index writing technique;
The head is positioned on a second selected data track having a first predetermined number of physical sectors and when the head encounters each physical sector and the physical sector is not defective; The head uses the variable index write technique to convert the second predetermined number of data blocks of a second selected portion of the selected cluster into the second selected data. If writing to the corresponding physical sector of the track and the physical sector is defective, the head skips the physical sector of the second selected data track and the first of the selected data cluster; When all the data blocks of two selected portions have been written to the second selected data track, the head Skipped defective physical sectors in the second data track from the first predetermined number of physical sectors of the two selected data tracks to the second predetermined number of data blocks 19. The disk drive of claim 18, skipping a fourth predetermined number of physical sectors equal to the sum of the number plus. Each portion of the data cluster corresponding to a single data track further includes a fourth predetermined number of error correction information data blocks;
19. The disk drive of claim 18, wherein the head writes at least one error correction information data block to a physical sector corresponding to a spare sector.   The disk drive of claim 21, wherein the at least one error correction information data block corresponds to the first spare data block.   The disk drive of claim 21, wherein at least one error correction data block corresponds to a last spare data block.   The disk drive of claim 14, wherein information regarding whether a physical sector is defective is included in metadata associated with the selected data track.   The metadata associated with the selected data track includes the number of physical sectors of the selected track, the number of parity sectors on the selected track, and a spare sector on the selected track. 25. The disk drive of claim 24, wherein information about at least one of the numbers is included.   The metadata associated with each selected data track includes the stream handle of the data stream, whether the selected track is free or occupied, and the physical of the selected track. 25. The method of claim 24, comprising information regarding at least one of a start sector of the selected data track related to start and a location of all defective sectors on the selected track. Disk drive.

Images