Abstract:
Apparatus and method for managing a media cache of a data storage device. In some embodiments, a media cache master table is maintained in a memory as a data structure having a plurality of entries that describe data sets stored in a non-volatile media cache memory. A first timecode stamp value is written to respective first and second locations in the table at the commencement of a data transfer operation to transfer data associated with the plurality of entries in the table. The first location is updated with a new, second timecode stamp value responsive to detection of an error condition that interrupts the data transfer operation. An error recovery operation is subsequently performed responsive to a detected mismatch between the timecode stamp values in the first and second locations.
Abstract:
Implementations disclosed herein provide a method comprising comparing high-latency data sectors of a storage band, the high-latency data sectors having latency above a predetermined threshold, with target sectors for storing new data to determine one or more of the high-latency data sectors that may be skipped during retrieval of at-rest data from the storage band.
Abstract:
The disclosed technology provides a method that improves CCT in SMR device systems. In one implementation, the method comprises writing data to a shingled magnetic recording (SMR) band in a storage device, determining whether an off-track write has occurred, identifying unsafe written data in response to determining that an off-track write has occurred, determining whether caching space is available upon identifying unsafe written data, continue writing data to the SMR band without a write retry upon determining that caching space is available, and writing unsafe written data to the available caching space. In another implementation, the method comprises receiving a request to repair an encroached track in an SMR band, recovering encroached data to a dynamic random-access memory, determining whether caching space is available, writing the recovered data to the available caching space upon determining that caching space is available, and merging other cached data in the SMR band.
Abstract:
Methods and systems that reduce off-track write retry operations in shingled magnetic recording systems. In one implementation, the method includes writing data to an initial track, determining which side of the initial track is a shingled side, calculating a percentage of position error signal (PES) at a shingled side end of the initial track (PES1) when an off-track write operation occurs, determining whether the PES1 meets a first pre-determined threshold, continue writing data to a second track responsive to determining the PES1 is below a first pre-determined threshold, calculating a percentage of PES at a shingled side end of the second track (PES2), determining whether a combined value of PES1 and PES2 is above a second predetermined threshold to determine a probability value of the initial track being erased, and continue writing to a third track if the combined value is below the second predetermined threshold.
Abstract:
The disclosed technology provides a method that improves CCT in SMR device systems. In one implementation, the method comprises writing data to a shingled magnetic recording (SMR) band in a storage device, determining whether an off-track write has occurred, identifying unsafe written data in response to determining that an off-track write has occurred, determining whether caching space is available upon identifying unsafe written data, continue writing data to the SMR band without a write retry upon determining that caching space is available, and writing unsafe written data to the available caching space. In another implementation, the method comprises receiving a request to repair an encroached track in an SMR band, recovering encroached data to a dynamic random-access memory, determining whether caching space is available, writing the recovered data to the available caching space upon determining that caching space is available, and merging other cached data in the SMR band.
Abstract:
In one implementation, this disclosure provides a method for executing a partial band rewrite operation comprising identifying a first track of a shingled data band to receive data of a write received at a shingled media storage device and writing a first subset of the received data to a media cache, where the first subset corresponds to the first track of the shingled data band to receive data of the write command. The method also includes writing a remaining subset of the received data to target locations within the shingled data band without updating data on the first track within the shingled data band corresponding to the first subset of the received data.
Abstract:
The disclosed technology provides a method that improves CCT in SMR device systems. In one implementation, the method comprises writing data to a shingled magnetic recording (SMR) band in a storage device, determining whether an off-track write has occurred, identifying unsafe written data in response to determining that an off-track write has occurred, determining whether caching space is available upon identifying unsafe written data, continue writing data to the SMR band without a write retry upon determining that caching space is available, and writing unsafe written data to the available caching space. In another implementation, the method comprises receiving a request to repair an encroached track in an SMR band, recovering encroached data to a dynamic random-access memory, determining whether caching space is available, writing the recovered data to the available caching space upon determining that caching space is available, and merging other cached data in the SMR band.
Abstract:
The disclosed technology provides a method that improves SMR throughput in vibration in storage systems. In one implementation, the method comprises receiving a write command to write data on a first track in a band of a storage medium, performing a vibration detection scheme to identify vibration events, determining if a number of vibration events is above a predetermined threshold, skipping the first track responsive to determining the number of vibration events is above a predetermined threshold, seeking to a second track adjacent to the first track, increasing an OCLIM on the second track adjacent to the first track from a default OCLIM to an increased OCLIM, and shifting the writer center on the second track adjacent to the first track.
Abstract:
A data sector is read using two or more passes of a read head over a recording medium. Each of the passes corresponds to a different cross-track offset, each of the data sectors being divided into two or more blocks. For each of the blocks, data is selected from one of the passes that read the block with a higher quality than other passes that read the block. The selected data from each of the blocks to is combined to form recovered data of the data sector.
Abstract:
Systems and methods are disclosed for performing band rewrites on shingled storage media based on error scan counts. In one embodiment, a device may comprise a data storage medium including a first band and a processor. The first band may include a plurality of tracks storing data in a shingled manner where a first track at least partially overlaps a second track. The processor may be configured to receive a write command directed to the first band, increment a first rewrite count of the first band based on the write command, and when the first rewrite count exceeds a first rewrite threshold, perform a read-modify-write (RMW) operation for all of the first band. In some embodiments, a device may maintain write counts for sections of a band, and perform RMW operations on less than all of the band.