Method for determining a source of failure during a file system access

A method for determining a source of data corruption in a computer system, comprising:

• generating a sequence of file references from a first location in the computer system;
• maintaining a record of the generated sequence of file references in a form that allows a file reference in the generated sequence of file references to be retried if necessary;
• examining results returned in response to the sequence of file references to detect a failed file reference;
• and
• if a failed file reference is detected,
• taking an action to correct the cause of the failed file reference, and
• retrying the failed file reference to determine the source of failure by transferring the failed file reference from the record of the generated sequence of file references to a second location in the computer system, and retrying only the failed file reference and no other references from the generated sequence of file references from the second location in the computer system.

A method for determining a source of data corruption in a computer system, comprising:

• generating a sequence of file references from a first location in the computer system using a plurality threads that are able to execute concurrently, so that different storage devices can be tested at the same time;
• maintaining a record of parameters specifying the generated sequence of file references in a form that allows a file reference in the generated sequence of file references to be retried if necessary, including updating a data structure containing the generated sequence of file references;
• examining results returned in response to the sequence of file references to detect a failed file reference;
• and
• if a failed file reference is detected,
• taking an action to correct the cause of the failed file reference, and
• retrying the failed file reference to determine the source of failure by transferring the failed file reference from the record of the generated sequence of file references to a second location in the computer system, and retrying only the failed file reference and no other references from the generated sequence of file references from the second location.

An apparatus for determining a source of data corruption in a computer system, comprising:

• a reference generation mechanism coupled to the apparatus, for generating a sequence of file references from a first location in the computer system;
• a recording mechanism coupled to the apparatus, for maintaining a record of the generated sequence of file references in a form that allows a file reference in the generated sequence of file references to be retried if necessary;
• an examining mechanism coupled to the apparatus, for examining results returned in response to the sequence of file references to detect a failed file reference;
• and
• a failure retrying mechanism for retrying the failed file reference to determine the source of failure, the failure retrying mechanism being configured to take an action to correct the cause of the failed file reference, and to use the record of the generated sequence of file references to retry only the failed file reference and no other reference from the generated sequence of file references in order to determine the source of data corruption.