JPH0614312B2 - Control method for rotary storage device - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a control system for a rotary storage device such as a magnetic disk device, and particularly to a rotation suitable for effectively utilizing a wasteful time generated during a sequential operation. The present invention relates to a method for controlling a storage device.

[Background of the Invention]

In recent years, with the fixed and large-capacity magnetic disk devices, it takes a lot of time to drop the backup file onto a magnetic tape or the like (dump processing), and it is desired to shorten the time. Since this dump process cannot predict the format and number of records on a track, the dump program uses the method of reading all the records on one track with one command. It has been improved by processing and batch processing of multiple tracks. However, even when there are only a few records on the track, the track portion is combined with the channel unless it needs to be processed, so that the busy time of the magnetic disk controller and the channel increases, and multi-processing is not possible. It has a bad influence.

Conventionally, various control methods have been devised in order to validate the path of the magnetic disk control device. For example, a set sector method in which a target record position is designated from the upper side, and a channel and a disk device are released until the head is positioned at this record start point, or Japanese Patent Publication Sho 56-1.
As described in No. 7692, there is a method of releasing the channel and the disk device until the erase operation of the disk device is completed. However, the former method is limited to the time when the record start time is known in advance on the upper side, and the latter method mainly focuses on causing the disk device to execute the erase function and release the path. No mention is made of forecasting or repositioning until the start of the record.

[Object of the Invention]

An object of the present invention is to improve the system throughput by effectively utilizing the wasted time generated in the sequential operation of a certain storage device in a system in which a plurality of rotary storage devices exist under the processing device.

[Outline of Invention]

According to the present invention, when a certain storage device sequentially performs data processing in track units, the time between the record end point and the next record start point is calculated in the control device, and if this time is a certain time or more, , The path route is released and it is possible to service other storage devices.

Example of Invention

An embodiment of the present invention will be described in detail below with reference to the drawings.

FIG. 2 is a system configuration diagram of an embodiment of the present invention. In FIG. 2, the processing unit (CPU) 101 is a channel 1
02, and the disk controller (A) 103 via the channel interface 107. The disk controller (A) 103 is a control interface (A)
The disk connection device (A) 105 is connected via a device interface (A) 109 to the magnetic disk device group 10 via a device interface (A) 109.
6,113 are connected. Disk controller (A)
103 is further connected to the disk control device (B) 104, and the disk control device (B) 104 is connected to the disk connection device (B) 111 via the control interface (B) 110, and the disk connection device (B) 11
1 is connected to another magnetic disk device group 114 via a device interface (B) 112.

Now, the processing device (CPU) 101, the channel 102, the disk connection device (A) 103, the disk connection device (A)
105, a magnetic disk device (O) 106 is a channel interface 107, a control interface (A)
108 and the device interface (A) 109, let us consider a dump process for reading data from a disk device 106 over a plurality of tracks.

Normally, when dumping data from a disk device to a magnetic tape, it is not known in what format the data is written in the track, so a method of reading all records on the track with one command is used. In this case, the command chain is MT as shown in FIG.
SEARCH ID (search operation) 201, TIC 20
2, READ MCKD (read operation) 203 is repeated. A conventional operation flowchart for the command portion of READ MCKD is shown in FIG. That is, the records are read in the order of records 1, 2, ..., After the processing of all records is completed, the next command is received, and the command is MT SEARCH ID204.
If so, it waits until the index marker, switches to the next track, and continues processing. Therefore, when only a small number of records are present on the track, as in the case of track 1 in FIG. 3, the path routes 107, 108, and 109 are set even after the end of the final record processing and during the time when data transfer to the index marker is not performed. It is a privately owned company.

In view of this point, the present invention predicts the time from the record end point to the next record start point on a track when processing a plurality of tracks collectively in a batch, and if this time is a certain time or more, , The corresponding path route is released, the re-connection is requested from the storage device side at the next record start point, and the service to another storage device is enabled within this release time. As a method for grasping the position of the last record, when creating a record format, for example, as shown in FIG.
The additional information of the segment number (SN) is recorded in A) and the count section (C0 to CN) as position information indicating the distance from the index marker. This segment number can be realized by adding the gap length, the key length, the data length, etc. to the segment number of one record before.

FIG. 1 shows an operation flowchart of the operation according to the present invention.

Upon completion of processing the last record on the track of the disk device 106 (steps 11 and 12), the disk control device 103 starts the record on the next track from the end point of this record by the microprogram from the segment number on the track. The time α to the point is calculated (step 13). That is, MT SEARCH in this case
The ID command is a request to process the index marker of the next track and subsequent ones, and the current segment number is subtracted from the number of segments (corresponding to the time for one rotation) of the entire track, and it is a micro program Can be easily realized by If this time α is less than the reconnection time between the channel and the disk device, the current processing is performed (steps 14 and 20), and if it is more than that, a set target for positioning the index marker is performed (steps 14 and 15). , The channel path route is released by command retry (step 16). Note that the set target is that by specifying a certain sector on a certain track from the disk connection device to the disk connection device, the disk device reconnects to the disk connection device and channel when the head comes to that position. It is the function that you request.

When the head is positioned near the index, the disk device 106 requests the disk controller 103 to reconnect (step 17), and if the channel 102 and the disk controller 103 are free at that time, the path is reconnected. I (steps 18 and 19), MT SEARCH I
The D command is issued again. At this time, the head of the disk device 106 is located near the index marker, and the SEARCH operation can be performed by switching the head to the next track without wasting time. While the path is open, the channel 102 is the disk controller 1
If a service other than the disk device 106 under the control of 03 is servicing any device under the control of the disk control device 104 and the service is not completed when the reconnection request is issued from the disk device 106, The reconnection request is not accepted and is executed after waiting for one rotation (step 21).

Next, the effect of the reconnection positioning method will be described with reference to FIG. This is the same as the rotation wait (positioning method 601) in the normal set sector command. In this case, the head positioning completion timing 603 at the time of reconnection becomes A, and the reconnection request from the disk device disappears after passing the index marker 604, so there is no reconnection completion timing after the B timing and record 0 is detected. Therefore, useless search operation is not repeated. For example, if the completion method 602, which has conventionally used the reconnection request for the completion of erasure, is required to complete the reconnection on the condition that both the disk controller and the channel are free. There are cases in which it becomes possible. At this time, since the corresponding record 0 has already passed, the channel is occupied for one rotation time. Therefore, it is apparent that the positioning method 601 is also desirable for reconnection after erasing is completed.

〔The invention's effect〕

According to the present invention, with respect to data processing in track units by a sequential operation in a rotary storage device, a path route can be released for the time between the last echo and the start point of the next record, and other service requests can be satisfied during that time. ,
There is an excellent effect that the effective use is improved in a multiple operation environment.

In particular, when dumping the data in the disk device, there are many tasks in which the entire physical image is read without being aware of the file attributes. At this time, it is often the case that there is no data or no data on a certain track, and it ends with a small record, and therefore, the wasteful occupation time is considerable, but according to the present invention, this is improved. It

[Brief description of drawings]

FIG. 1 is an operation flowchart of one embodiment of the present invention, and FIG.
FIG. 3 is a system configuration diagram of an embodiment of the present invention, FIG. 3 is a diagram showing a relationship between a track format and a command chain,
FIG. 4 is a conventional operation flowchart, FIG. 5 is a diagram showing a track format for explaining a segment number, and FIG. 6 is a timing diagram for explaining the effect of the reconnection positioning system according to the present invention. 101 ... Processing device, 102 ... Channel, 103, 104 ... Disk control device, 105, 111 ... Disk connecting device, 106, 113, 114 ... Magnetic disk device.

Claims (1)

[Claims] 1. In an information processing system in which a plurality of rotary storage devices are connected to a processing device via a channel and a control device, the rotary storage device sequentially and collectively processes data of a plurality of tracks. In the control device, when the rotary storage device finishes processing the last record of one track, the control device predicts the time to the record start point of the next track, and It is determined whether or not the time is equal to or longer than a predetermined time, and if the time is longer than or equal to the predetermined time, the connection between the channel and the control device and between the control device and the rotary storage device is temporarily released, and the next connection from the rotary storage device is performed. A method of controlling a rotary storage device, characterized in that a reconnection request is received and connection is reestablished at the start of recording a track.