Browse Prior Art Database

Disconnected Write/Read Diagnostic with Real Time Error Collection Disclosure Number: IPCOM000196378D
Publication Date: 2010-Jun-02
Document File: 6 page(s) / 42K

Publishing Venue

The Prior Art Database


Disclosed is a method for testing magnetic tape devices, which both increases test efficiency and also decreases the cost over typical testing methodologies. The disclosed method is not limited to tape devices and could easily be applied to other types of computer storage devices. This testing methodology employs the use of an extended write / read sequence, where the tape device controls the majority of the operation and the host computer is only required to initiate the test and then monitor its progress. With this testing methodology, two primary benefits can be realized - 1) reliability testing can be performed without requiring continuous connection to a computer server for data transfers 2) detailed error information and counting statistics can be collected and saved while the test is in progress.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 36% of the total text.

Page 1 of 6

Disconnected Write/Read Diagnostic with Real Time Error Collection

The notion of having magnetic tape devices (or other computer storage devices) run outboard diagnostics and self-tests is well known and widely used. This type of testing has some limitations though, such as having a small number of testing variations or a short test duration. A self-test diagnostic can often be looped, but this typically involves

just running the same short test over and over.

Therefore, it is normal for a tape device development organization to use a different methodology to perform the reliability and functional verification testing required to certify new hardware or storage media. This kind of testing typically requires multiple computer servers, along with a large suite of test software. Because tape storage devices are being tested, the servers are kept continually busy, transferring large amounts of data to or from the tape drives. And the test software is typically kept continually busy checking every command response, trying to vary the size and content of the data records being transferred, and then also verifying that every bit in every byte is correctly stored and returned by the tape drive.

In order to conduct a large scale reliability test of this type, the development and test organization may need to deploy a large number of high speed servers, along with associated host bus adapters and switches. The result is a very high cost for test equipment. This problem increases over time, since tape devices continue to provide improvements in storage capacity and data transfer speeds.

One possible method to help relieve this problem could be to do more outboard diagnostics and self-tests, as a trade-off for doing fewer tests while connected to the computer servers. But this suffers from the limitations listed above. In addition, this type of outboard testing may allow for the tape device to collect some errors statistics while running self-tests, but this is generally just counts of different events and it does not provide the type of detailed information that can be collected when testing is done directly by the computer server.

This article now presents a testing methodology for running an extended write/read sequence, where the tape device controls the majority of the operation and the host computer is only required to initiate the test and then monitor its progress. Thus, the need for continual interaction between the tape device and the server is reduced. The method makes use of currently existing SCSI commands, and in particular, extensions to the Send Diagnostic command. Additionally, a key component of the method makes use of an existing notion to execute the command in "Immediate Mode". The purpose of Immediate Mode is to allow the tape device to perform initial parameter checking on the command, and then if conditions are correct, the tape device ends connection with the host server and starts execution of the extended write/read...