Error detection/correction and fault detection/recovery – Data processing system error or fault handling – Reliability and availability
Reexamination Certificate
1998-03-02
2001-11-27
Le, Dieu-Minh (Department: 2184)
Error detection/correction and fault detection/recovery
Data processing system error or fault handling
Reliability and availability
C714S710000
Reexamination Certificate
active
06324655
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to an I/O controller that controls an I/O unit based on a command from a host unit, and more particularly to an I/O controller that supplies, to a host unit, preventive maintenance information about a spare I/O unit, which is used in place of an I/O unit in which an abnormality occurs, when any abnormality occurs in an I/O unit.
Recently, the demand for high reliability in a computer system has accompanied the demand for high reliability in an I/O subsystem. This has led to the practice of providing an I/O subsystem with a spare I/O unit. In such an I/O subsystem, the spare I/O unit is not used during normal operation. It is used only when any abnormality including a read error or a hardware failure occurs elsewhere in a current I/O unit.
Consequently, it is required to provide an I/O subsystem with a means of monitoring normality /abnormality of a spare I/O unit and of supplying such preventive maintenance information as regarding a hard error or a read error, to a host unit, in order for a spare I/O unit to carry out the performance expected of a normal spare I/O unit.
BACKGROUND ART
An array-type magnetic disk controller has been in use as an I/O controller for supplying preventive maintenance information about a spare I/O unit, to be used in place of an I/O unit where an abnormality is found, to a host unit. An array-type magnetic disk controller comprises a host unit interface controller, a subordinate unit interface controller, a processor, and various registers. A plurality of data disk drives, one parity disk drive, and one spare disk drive are connected as subordinate units and are under synchronous rotation control.
An array-type magnetic disk controller checks for data abnormality by reading data and parity, parallel from each of a plurality of data disk drives and the parity disk drive. When no abnormality is found and a normal operation is confirmed, data is transferred to a host unit unmodified.
When any data abnormality is detected, an array-type magnetic disk drive restores correct data on the basis of data and parity which are read from a normal data disk drive, other than the one disk drive the defective data is read from, and a parity disk drive, and transfer it to the host unit.
If, for some reason, any abnormality including a read error or a hardware failure occurs more frequently than the specified times in one of the data disk drives acting as a current I/O unit, or in a parity disk drive, all the data written in the abnormal data disk drive or the abnormal parity disk drive is automatically and instantly recovered from the other normal data disk drives and a normal parity disk drive and restored in the spare disk drive. By using the spare disk drive as a current I/O unit, in place of the abnormal disk drive, continued operation is ensured and system down time is prevented.
However, it could happen, that the presence of a hard error or a read error in a spare disk drive itself cannot be readily detected by a host unit, and the spare disk drive itself is found to contain a hard error or a read error the first time the spare disk drive is used as an I/O unit in place of the abnormal disk drive, and cannot perform the function of a spare disk drive.
That is why a method has been in use, in an equipment comprising a main system consisting of a plurality of units, and a spare unit for backing up the units of the main system, wherein a spare unit is monitored by providing a spare unit with the same input signal as the one given to any unit of the main system, and comparing both signals output from the spare unit and the aforementioned unit of the main system (Japanese Laid-Open Patent Application 56-72359). However, this method of monitoring necessitates always operating the spare unit concurrently with the current unit in the main system. It is equivalent to having one more I/O unit to be controlled by an I/O controller, and results in increased load on the host unit.
Another method has been in use wherein abnormality of an I/O controller is regularly monitored according to a time table (Japanese Laid-Open Patent Application 62-212856). This does not, however, either monitor abnormality of a spare I/O unit or utilize a time interval not in use for an operation. Still other methods have been in use, including a method of monitoring, wherein an equipment, provided with dual central processing units (CPU), namely the current and spare CPUs, allows a test to be carried out under a condition similar to a normal operation by running a software in the spare CPU concurrently (Japanese Laid-Open Patent Application 2-93953), and a method of monitoring wherein a test is carried out at a regular interval on the spare CPU (Japanese Laid-Open Patent Application 62-90068). However, these methods of equipment monitoring having dual CPUs are not the ones for monitoring a spare CPU utilizing a time interval not in use for an operation. The present invention was developed in view of these considerations.
SUMMARY OF THE INVENTION
It is an object of the present invention is to provide an I/O controller designed for monitoring the status of a spare I/O unit and/or a current I/O unit.
It is another object of the present invention is to provide an I/O controller designed for monitoring the status of a spare I/O unit and/or a current I/O unit, at a time interval not in use for an operation when no command from a host unit is being processed.
In order to achieve the aforementioned objects, the I/O controller of the present invention consists of a subordinate unit interface controller connected to a plurality of I/O units and one spare I/O unit, a host unit interface controller which receives action commands issued from a host unit to the aforementioned I/O units or the spare I/O unit, a means of checking normality of the data in the I/O units by reading data and an error-detecting code concurrently from the I/O units, a controlling unit, a decision unit, and a preventive maintenance information storage register.
When the data in all the I/O units is found to be normal by the aforementioned checking unit, the aforementioned controlling unit selects the main I/O units only. When any of the I/O units is found to contain a data abnormality with the checking unit, the aforementioned controlling unit selects the other normal units and the aforementioned spare I/O unit, and controls the aforementioned I/O units or spare I/O unit based on the aforementioned command input via the host unit interface controller.
The aforementioned decision unit selects the aforementioned I/O unit or spare I/O unit via the aforementioned subordinate unit interface controller at a time interval not in use for an operation when the controlling unit is not processing the command input from the aforementioned host unit via the aforementioned host unit interface controller, and decides whether the selected I/O units or spare I/O unit operates normally. A preventive maintenance information storage register temporarily stores information about the abnormality when the decision unit gives a decision of abnormality.
The present invention makes it possible for a host unit to monitor a hard error and a read error in an I/O unit or a spare I/O unit as part of a normal operation. This allows repairing or exchanging any abnormal I/O unit during a normal operation, and putting a spare I/O unit into a normal operation without delay when any abnormality including a read error or a hardware failure occurs in a current I/O unit so that the spare I/O unit is used for the first time by the host unit.
The present invention uses a plurality of data disk drives and one parity disk drive as the aforementioned I/O units, uses one spare disk drive as the aforementioned spare I/O unit, comprises synchronous rotation controller that puts these disk drives under synchronous rotation controller, and comprises the aforementioned controlling unit, checking unit, decision unit, a subordinate unit interface controller, and a channel interface controller. By using the present invention, it is p
Matoba Tatsuo
Yorimitsu Keiichi
Fujitsu Limited
Le Dieu-Minh
Staas & Halsey , LLP
LandOfFree
Input/output controller providing preventive maintenance... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Input/output controller providing preventive maintenance..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Input/output controller providing preventive maintenance... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2586993