JPH0156464B2 - - Google Patents

Description

[Detailed Description of the Invention] (a) Technical Field of the Invention The present invention relates to a mass storage system, and particularly to data writing in a method in which a host computer directly accesses a magnetic tape of a data cartridge without going through a magnetic disk. Regarding the time retry method.

(b) Background of the Technology A data cartridge used in a mass storage system stores a magnetic tape having a format as shown in FIG. 1 is a magnetic tape, 2 is an ID track on which a stripe number unique to each stripe is prerecorded, 3 and 3' are servo tracks that control the rotating head to come to the center of the stripe, and 4 is a stripe for recording data. . For example, there are 13,544 stripes, and each stripe records 4,096 bytes of data.

Conventional mass storage systems have introduced the concept of virtual magnetic disks, in which the data on the magnetic tape is temporarily transferred to the magnetic disk, and the host computer accesses the magnetic disk instead of directly accessing the magnetic tape. Giving and receiving. In a typical example, 67 magnetic tape stripes are grouped together into one block, which corresponds to the cylinder of a magnetic disk. Therefore, the magnetic tape is divided into blocks and data stripes 61 as shown in Figure 2.
There are four stripes for alternating the book, defective stripes, and two stripes for the system, and one stripe for the system represents the division of each block. In FIG. 2, 5 indicates one block, which includes 67 stripes. 6 indicates one stripe for the system. Data is exchanged between this block and the cylinder of the magnetic disk.

When writing data to stripe 4 of magnetic tape 1, read the stripe number of ID track 2 to confirm the stripe on which data is to be recorded, then read servo track 3 to confirm the on-track, and then write data to stripe 4. At this time, if the magnetic tape 1 moves or if the magnetic tape 1 is not properly positioned by the servo signal, there is a risk that data will be written to the adjacent stripe and the adjacent stripe on which data has already been recorded will be destroyed. Therefore, if the servo track 3' cannot be read, it is determined that data writing has not been completed normally and an error has occurred that may destroy adjacent stripes. When such an error occurs, in conventional large-capacity storage systems, the rotating head is returned to the beginning of each block to rewrite the data, but the host computer accesses the magnetic tape of the data cartridge directly without going through the magnetic disk. In the method of transmitting and receiving data, if an error occurs in a stripe of the magnetic tape that may destroy an adjacent stripe while data is being written, rewriting using the conventional method as described above cannot be performed. That is, in a system in which a host computer directly accesses a magnetic tape, it is necessary to effectively utilize the stripes of the magnetic tape and to sequentially record data sets of various lengths in order to increase processing speed. Data has various lengths, and if it is allocated to each block, unused stripes will occur, reducing the utilization rate of the magnetic tape. Therefore, in order to prevent the occurrence of unused stripes, the data set is sequentially packed and recorded as shown in Figure 3. In FIG. 3, 1 is a magnetic tape, 5 is a block, and a1, a2, a3, a4, and a5 are data sets. For example, if short data such as data set a3 is allocated to one block, there will be many unused stripes and it will be wasteful, so the data will be recorded compactly as shown in the figure, but in this case, if an error that may destroy adjacent stripes occurs, the data will be re-recorded. The problem lies in the trial method, and countermeasures are desired.

(c) Purpose of the Invention The purpose of the present invention is based on the above-mentioned request, by providing a temporary storage means so that when an error that may destroy an adjacent stripe occurs, the data can be rewritten using the data left in the storage means. An object of the present invention is to provide a retry method for a magnetic recording/reproducing apparatus which has a high utilization rate of magnetic tape in a system in which a host computer directly accesses a data cartridge and has a fast data writing processing speed.

(d) Structure of the Invention The structure of the present invention includes a plurality of storage means for storing write data, a means for reading and writing data from one of the storage means, and a method that may destroy adjacent stripes during writing of the data. means for detecting the occurrence of an error, and the maximum amount of data written in one activation is the same as the maximum capacity of one of the storage means, and even if data writing to the first storage means is completed, the data is is retained without being erased, and if an error that may destroy adjacent stripes occurs while writing data to the second storage means, the area where the data was written in the first storage means and the error If the occurrence stripes are in the same physical block, rewriting is performed using the data left in the first storage means from the beginning of the area where the data in the first storage means is recorded.

If the data in the second storage means is recorded from the first physical block to an adjacent second physical block, and the error occurs in a stripe in the second physical block, the second physical block is recorded. Rewriting starts from the first stripe.

If the data in the first storage means is recorded across a second physical block adjacent to the first physical block, the data is left in the first storage means from the first stripe of the second physical block. The data is rewritten using the data. Furthermore, when writing of the data in the second storage means is completed, the data in the first storage means is erased and new data is stored and written on the magnetic tape, but the data in the second storage means is not erased. waits for completion of data writing in the first storage means, or
If an error that may destroy an adjacent stripe occurs during writing to the storage means, the same processing as described above is performed.

(e) Embodiment of the Invention FIG. 4 is a block diagram of a circuit showing an embodiment of the invention. Data sent from the host computer enters the interface 10 and is stored in the buffer memory 11. When the necessary data is stored in the buffer memory 11, the data is sent to the buffer memory 12, and under the control of the control circuit 13, the multiplexer 14 transfers the data to the buffer memory 11. The data is sent to the write circuit 16 via the interface 15, and the data is written onto the magnetic tape stored in the data cartridge 19 under the control of the control circuit 17. For example, the magnetic tape 1 shown in FIG.
Assume that the data is written in the area indicated by 24. Control unit 13
The data in the buffer memory 11 is held without being erased, and the data in the buffer memory 12 is written in the area 25 of the magnetic tape 1 in FIG. 5 in the same manner as described above.
When the writing of data in the buffer memory 12 is completed, the control section 13 erases the data in the buffer memory 11, causes the host computer to write the next new data into the buffer memory 11, and causes the data in the buffer memory 12 to be held without being erased. buffer memory 1
When the error detection circuit 18 detects an error that may destroy an adjacent stripe while writing new data of No. 1 to the area shown in FIG. . When the control circuit 13 confirms based on the count values of the physical block counter 20 and the stripe counter 21 that the data write area of the buffer memory 12 and the data write area of the buffer memory 11 at the time of the error occurrence are within the same physical block 22 area. , the magnetic tape 1 is rewound to the beginning of the area 25 as shown by the arrow in FIG. If the capacity of the buffer memories 11 and 12 is n stripes, and the physical block has m stripes, the data write area 25,
It is possible to determine whether or not the magnetic tape 26 is in the same physical block, and it is also possible to rewind the magnetic tape 1 by counting the number of stripes written until an error occurs. The control circuit 13 sends the remaining unerased data in the buffer memory 12 to the write circuit 16 via the multiplexer 14 and interface 15, and rewrites the data from the beginning of the area 25 of the magnetic tape 1 in FIG.

The new data in the buffer memory 11 is transferred to the sixth buffer memory 11.
Error detection circuit 18 while writing to the area shown in FIG.
When the control circuit 13 detects an error that may destroy adjacent stripes, the control circuit 13, which has received the error report in the same manner as described above, uses the count values of the physical block counter 20 and the stripe counter 21 to block the data writing area of the buffer memory 12 from becoming a physical block. When the data is written to the area 27 in the buffer memory 11 across the physical blocks 22 and 23 and it is recognized that the error has occurred in the physical block 23, the interface 15 The magnetic tape 1 is rewound as shown by the arrow in FIG. 6 to the first stripe position of the physical block 23 by controlling the control circuit 17 via the physical block boundary.
Control circuit 1 based on the count value of stripe counter 21
3 determines the position of the rewriting start data in the buffer memory 11 and performs rewriting in the same manner as described above.

The new data in the buffer memory 11 is transferred to the seventh buffer memory 11.
Error detection circuit 18 while writing to the area shown in FIG.
When the control circuit 13 detects an error that may destroy an adjacent stripe, the control circuit 13 receives the error report and changes the data in the buffer memory 12 as shown in the area 27 based on the count values of the physical block counter 20 and the stripe counter 21. When it is recognized that the magnetic tape is recorded across physical blocks 22 and 23 and that the error has occurred within the area of physical block 23, it controls the control circuit 17 to move the magnetic tape 1 to the position of the first stripe of physical block 23. Rewind as shown by the arrow in Figure 7. Based on the count value of the stripe counter 21, the control circuit 13 determines the position of the rewriting start data of the remaining data in the buffer memory 12, and rewrites the remaining data in the buffer memory 12 in order from the data in the buffer memory 11 to the data in the buffer memory 11 in the same manner as described above.

Although this embodiment has been described using hardware, it is of course possible to perform the above operations using software using the storage area of the host computer.

Also, although the number of buffers has been explained as two, three or more will have the same or greater effect.

(f) Effect of the Invention As explained above, the present invention is applicable to a large-capacity storage system in which a host computer directly accesses a magnetic tape and in which physical blocks are provided to divide and manage the recording area of the magnetic tape. This method has a great effect because it provides a method for retrying without any problem even if an error that may destroy adjacent stripes occurs during data writing.

[Brief explanation of drawings]

Figure 1 is a diagram showing the format of magnetic tape.
FIG. 2 is a diagram for explaining blocks of a magnetic tape, FIG. 3 is a diagram for explaining an example of recording a data set on a magnetic tape, FIG. 4 is a block diagram of a circuit showing an embodiment of the present invention, and FIG. 6 and 7 are diagrams for explaining the data recording method of the present invention. 1 is a magnetic tape, 2 is an ID track, 3 and 3' are servo tracks, 4 is a stripe, 5 is a block, 10 and 15 are interfaces, 11 and 12 are buffer memories, 13 and 17 are control circuits, 14 is a multiplexer, 16 is a write circuit, 18 is an error detection circuit, 19 is a data cartridge, 20 is a physical block counter, 21 is a stripe counter, and 22 and 23 are physical blocks.

Claims (1)

[Claims] 1 Recording or reproduction is performed using a rotating head on stripes formed perpendicularly or diagonally to the longitudinal direction of the recording medium, a certain number of stripes are grouped together into one physical block, and the final stripe of the physical block is In a magnetic recording and reproducing device that does not record data, there are a plurality of storage means for storing written data, a means for reading and writing data from one of the storage means, and a risk of destroying adjacent stripes during data writing. means for detecting the occurrence of a certain error, the maximum amount of data to be written in one activation is the maximum capacity of one of the storage means, and even if data writing to the first storage means is completed, the data is not stored. If an error occurs that may destroy adjacent stripes while writing data to the second storage means without erasing it, 1) The area where the data was written in the first storage means and the error If the occurrence stripes are in the same physical block, rewriting is performed using the remaining data of the first storage means from the data recording area of the first storage means,
2) If the data in the second storage means is recorded across a second physical block adjacent to the first physical block, and the error occurs in a stripe in the second physical block, the second physical block is Rewrite from the first stripe of the physical block, 3)
If the data in the first storage means is recorded across a second physical block adjacent to the first physical block, the data is left in the first storage means from the first stripe of the second physical block. A retry method for a magnetic recording/reproducing device, characterized in that rewriting is performed using data that has been recorded.