Proactive driver response to an operating system if a drive is removed from a RAID configuration
AssigneeLSI Corporation
InventorJose K. Manoj
Life of this patent
Estimating the term…
Abstract
A method for responding to a particular drive being removed from a drive array, comprising the steps of (A) determining a maximum drive response time of the particular drive being removed from the drive array; (B) determining a duration of each of one or more commands needing completion; (C) if a particular one of the commands takes longer than the maximum drive response time, aborting the particular command and checking if the drive is physically present; and (D) if the command takes less than the maximum drive response time, completing the command.
Claims17
A method for responding to a particular drive being removed from a drive array, comprising the steps:
- (A) determining a maximum drive response time of said particular drive being removed from the drive array;
- (B) determining a duration of each of one or more commands needing completion;
- (C) if a particular one of the commands takes longer than said maximum drive response time, aborting the particular command and checking if said drive is physically present;
- and
- (D) if said command takes less than said maximum drive response time, completing said command, wherein said method detects and responds to said particular drive being removed before an operating system IO timeout has occurred.
A method for detecting and responding to the removal of one or more individual drives from a disc array, said method comprising:
- (A) storing one or more individual drives in a disc array;
- (B) establishing a maximum drive response time for said individual drives being removed from said disc array;
- (C) issuing a command to each of said one or more individual drives with a driver stored on a computer readable medium at a register start time;
- (D) determining if said command was executed by said one or more individual drives within said maximum drive response time with a command monitoring thread;
- (E) if said command is executed by said one or more individual drives at a time less than or equal to said maximum response time, repeating steps (C)-(D);
- and
- (F) if said command is not executed or executed by said one or more individual drives at a time greater than said maximum response time, checking whether said one or more drives are physically present, wherein said method detects and responds to said particular drive being removed before an operating system IO timeout has occurred.
An apparatus comprising:
- a disc array configured to send and receive data from a controller, said disc array comprising one or more drives;
- a driver stored on a computer readable medium for issuing a command to each of said drives;
- and
- a command monitoring thread stored in said drive and configured to (i) monitor said commands sent to each of said drives and (ii) determine whether said commands sent to each said drives are executed within a maximum drive response time, wherein said apparatus detects and responds to said particular drive being removed before an operating system IO timeout has occurred.
A method for detecting and responding to the removal of one or more drives from a disc array, said method comprising:
- (A) storing one or more drives in a disc array;
- (B) establishing a maximum drive response time for said drives being removed from said disc array;
- (C) issuing a command to each of said drives with a driver at a register start time;
- (D) determining if said command was executed by said drives within said maximum drive response time with a command monitoring thread;
- (E) if said command is executed by said drives at a time less than or equal to said maximum response time, repeating steps (C)-(D);
- and
- (F) if said command is not executed, or executed by said drives at a time greater than said maximum response time, checking whether said one or more drives are physically present, wherein said method detects and responds to said particular drive being removed before an operating system IO timeout has occurred.
Dependencies are inferred from the claim text.