Corruption checking during backup
AssigneeNetApp, Inc.
InventorsSridhar Chellappa, E. Rudolph Nedved
Life of this patent
Estimating the term…
Abstract
In one embodiment, a data corruption checker is integrated into the data backup operation. In one embodiment, the data corruption checker can skip over minor corruptions and back up the non-corrupt files, or can abort the backup and locate a major corruption. In one embodiment, the present invention includes receiving a backup command at a storage server, the backup command requesting that the storage server perform a backup operation on a storage component, scanning a set of inodes associated with the storage component to determine a subset of inodes corresponding to files to be backed up, and, as part of the scanning, identifying one or more corrupted files.
Claims26
A method comprising:
- receiving a backup command at a storage server, the backup command requesting that the storage server perform a backup operation on a storage component;
- and
- performing corruption checking over the storage component as part of the backup operation in response to the backup command,
- wherein the corruption checking comprises, in the storage server,
- initial scanning of the storage component for corruption,
- identifying one or more corrupted files of a file system of the storage server on the storage component by utilizing a plurality of corruption tests, each of which tests for a different type of error, and
- determining whether corruption in the one or more corrupted files is fatal or non-fatal, and
- the backup operation comprises
- copying uncorrupted files from the storage component to a backup facility and skipping the one or more corrupted files with non-fatal corruption.
A storage server comprising:
- a storage adapter through which to communicate with a storage subsystem that includes a plurality of storage components;
- a network adapter through which to receive data access requests from a storage client;
- and
- a processor configured to control the storage server so as to cause the storage server to perform corruption checking over a target storage component as part of a backup operation performed in response to the storage server receiving a backup command, wherein the corruption checking comprises
- identifying one or more corrupted files of a file system of the storage server on the target storage component by utilizing a plurality of corruption tests, each of which tests for a different type of error,
- determining whether corruption in the one or more corrupted files is fatal or non-fatal, and
- aborting the backup operation when the corruption checking detects a fatal error, but when the corruption checking detects only one or more non-fatal errors, then copying uncorrupted files from the storage component to a backup facility and skipping any corrupted files with non-fatal corruption.
A method comprising:
- receiving a backup command at a storage server, the backup command requesting that the storage server perform a backup operation on a storage component;
- scanning a set of inodes associated with the storage component to determine a subset of inodes corresponding to files to be backed up;
- identifying one or more corrupted files of a file system of the storage server during the scanning of the set of inodes, wherein a plurality of corruption tests are utilized for identifying the one or more corrupted files, wherein each of the plurality of corruption tests is for detecting a different type of error;
- and
- determining whether corruption in the one or more corrupted files is fatal or non-fatal.
A method comprising:
- receiving a backup command at a storage server, the backup command requesting that the storage server perform an NDMP based backup operation on a storage component;
- and
- in response to the backup command,
- scanning a set of inodes associated with the storage component to determine a subset of inodes corresponding to files to be backed up;
- identifying one or more corrupted files during the scanning of the set of inodes;
- and
- determining whether the one or more corrupted files warrant aborting the backup operation.
Dependencies are inferred from the claim text.