Described are techniques used in detection of a data corruption in a computer system. A host issues a write request that includes a checksum value determined in accordance with data associated with the write request. The write request is received by a data storage system that performs data validation using the checksum. If the data validation succeeds, the write operation proceeds. Otherwise, it is determined that the data is corrupt and a checksum error is returned to the issuing host. The host issues a vendor-defined write request operation that includes the checksum as a data field in the request packet sent to the data storage system. Filter drivers are used in obtaining the checksum and modifying the write request packet to specify a vendor-defined write operation if checksumming is enabled for the write operation.
A method, system, and computer instructions for using MD4 checksum to link transactions across machines. Transaction information is received, wherein transaction information comprises a checksum and topology information, and wherein the checksum serves as a unique identifier of the current transaction. A first data structure is populated with the checksum and attribute information for the current transaction. A second data structure is then populated with the checksum and topology information for the current transaction, wherein populating the second table preserves a root/parent/current transaction relationship. Preserving the root/parent/current transaction relationship allows for automatically linking a transaction to its related transactions when the transaction is uploaded from a client to a transaction database.
Techniques for verifying input/output (I/O) command data are provided. Information about the contents of the data are specified in the I/O command. After an application issues the I/O command, a subsequent component, such as a controller, uses the information to verify the contents of the data before the I/O command is performed.
A protection domain for a set of errors is defined using an association between data and first integrity metadata to protect data traversing an input/output datapath. The datapath has a storage device as one endpoint and a first generation integrity point for a host as an opposite endpoint. A first sub-domain is defined within the protection domain using an association between the data and second integrity metadata to further protect a portion of the datapath having a second generation integrity point as one endpoint. In another aspect, a second sub-domain is defined within the protection domain using an association between the data and third integrity metadata further protect data traversing a portion of the datapath having a third generation integrity point as one endpoint. The first and second sub-domains are nested within the protection boundary and may be in a hierarchical relationship.
