JPH07121304A - Disk control method - Google Patents

Abstract

(57) [Summary] [Purpose] Data is guaranteed even if the disk usage rate exceeds 100%. [Structure] The disk controller 12 is a working DASD 13
When the usage rate of a exceeds 80%, for example, writing of new data is prohibited and only rewriting is permitted. If the usage rate exceeds 100% at the time of rewriting data, the excess is written to the spare data storage unit 13e, and
Corresponding to the CKD track, the spare and active disk type and the FBA block number are registered in the mapping table 12e. When reading, the disk control device 12 obtains the information corresponding to the CKD track instructed by the higher-level device 11 from the mapping table 12e, and based on the FBA block number and the spare or working disk included in the information. The data is read and restored, the restored data is written in the memory 12c, a desired record is searched in the memory, and transferred to the higher-level device 11.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a disk control method, and more particularly to a disk control method for compressing data input from a host device and writing the data to the disk, and restoring data read from the disk and transferring the data to the host device. Regarding

[0002]

2. Description of the Related Art DASD (Direct Acce
ss Storage Device) track format is CK
There are D format and FBA format. FIG. 11 is an explanatory diagram of the CKD format, and FIG. 12 is an explanatory diagram of the FBA format. In the CKD format (FIG. 11), there are several areas separated by gaps on the track, and an index mark is attached at the beginning. H
A home address area is an area in which the address of the track (CKD track) in which it is present is described and is provided after the index mark. After the home address area, a count area C, a key area K,
Variable-length records R0 and R composed of data area D
1, R2, ... are written in plural. In the counter area C, the track address (CCHH, CC: cylinder address, HH: head number), the record number R and the subsequent key area, the length of the data area, etc. are written, and the track address and record number are specified. By doing so, individual records can be specified. A key for searching is written in the key area K, but it is not always necessary. User data is written in the data area D.

In the FBA format (FIG. 12), all logical blocks (FBA blocks) have the same length and each F
The BA block has unique identification numbers 1, 2, ... In addition, at the beginning of each FBA block, a 64-byte I
A section D (contents of which will be described later) is provided. In a computer system using a magnetic disk subsystem as an I / O subsystem, when an OS (operating system) handles data by a CKD control method, a CKD format disk device (CKD disk device) is connected as an I / O device. It

By the way, in recent years, the capacity of an FBA disk device holding an interface such as SCSI has dramatically increased with the progress of technology, and a capacity comparable to that of a conventional CKD disk device is being realized. Such an FBA disk device is small in size and low in price, and retains the performance enough to be used as an external device of a large computer system. Therefore, O
S handles the data by the CKD control method, and the magnetic disk device reads / writes the data in the FBA format.
A writing computer system has been proposed. In such a computer system, a magnetic disk controller provided between the channel and the FBA disk device converts the CKD / FBA format. FIG. 13 is an explanatory diagram of a computer system, (a) is a configuration diagram, and (b) is a format conversion explanatory diagram. 1 is host HST, 2 is channel CH, 3 is FB
The A disk units DEV and 4 are magnetic disk control units DCU provided between the channels and the FBA disk units. CKD to FBA format conversion, CKD
Each record of the format is divided into the size of the FBA block, and the ID part is added to the head to form an FBA block (FBA block conversion). In this case, a large number of records may be included in one FBA block, or only a part of the records may be included, depending on the relationship between the size of the FBA block and the record size.

The 64-byte ID portion inserted at the beginning of the FBA block has the contents shown in FIG. That is, in the ID part, the head sector value of the CKD track stored in the FBA block, a block flag that stores block-specific information (for example, a valid record exists in the block, etc.), the CKD track logic of the block. CCHH, individual record number existing in the record, information of each existing record (record descriptor),
The number of bytes (record pointer) from the head of the CKD track data stored in the block to the count / HA / EOT is recorded. The bit descriptors Bit7 to Bit0 have the following meanings.

Bit7: Standard record (record in which the values of physical CCHH and logical CCHH match) Bit6: HA field Bit5: EOT field Bit4: Record fits in this block Bit3: Interruption occurred when writing the count part Bit2: Interruption occurred when writing the key part Bit1: Interruption occurred when writing the data part Bit0: Unused.

In the FBA format, each logical block (FBA block) has an identification number. Therefore, if it is possible to identify which FBA block the CKD format data corresponds to, it is not necessary to continuously write the CKD format data in the FBA disk device. In other words, there is no need to worry about the place or order of writing. There is a method called mapping as a means for identifying which FBA block the CKD format data corresponds to. FIG. 14 is an explanatory diagram of mapping, 5 is a mapping table, and B1 to Bn are logical blocks obtained by dividing CKD format data (track position is CCHH) on a predetermined CKD track into n pieces. Assuming that each logical block B1, B2, B3 to Bn is written in the position of FBA block number 3, 23, 21, ... 40, the mapping table 5 corresponds to CCHH and FBA block numbers 3, 23, 21.・ ・ ・
40 is registered. In addition, a predetermined track (C
When a read command for a record on (CHH) is issued, the FB corresponding to CCHH is referred to by referring to the mapping table 5.
A block numbers 3, 23, 21, ... 40 are obtained, FBA blocks designated by the obtained numbers are sequentially read and stored in a memory, and a desired record position is obtained by referring to the ID section on the memory. Transfer the record to the host device.

[0008]

By the way, as one of the techniques which has been attracting attention in recent years, by processing or compressing data, the size of the data is reduced without damaging the contents of the original data and the data is stored in the external storage medium. Various compression / decompression techniques have been developed for storing and restoring original data from compressed data in actual data processing. For example, as a typical method, there is a run-length coding method or a universal coding method, which is a method of performing coding based on the continuity of data in one data string. In the run-length encoding method, for example, when focusing on the data “a”, “aa” is replaced with “a2”, “aaa”.
This is a method of compressing by writing “a” as “a5.” It is conceivable to adopt a method of compressing and writing data when writing a large amount of data to the physically same disk and restoring when writing data.

However, the problem in adopting the data compression technique is that the size of the compressed data cannot be obtained without actually performing the compression.
Therefore, when certain stored data is read, a part of the data is changed, and then the data is written again, the size of the compressed data may be different and the data may not be stored in the same place. Taking the run-length method as an example, after encoding "aaaaaa" as a5 and storing it in the FBA disk, if this data is changed to "aaabaa", the encoded data becomes "a2ba2". The data size of becomes large and cannot be stored in the original location. By the way, with FBA discs, CK
Since the mapping table manages which FBA block the D format data corresponds to, it is not necessary to continuously write the FBA-blocked logical block in the FBA disk device. Therefore, even if the data size after the change becomes large by rewriting, the data size after the change does not matter because it can be written in a vacant place and managed by the mapping table. For example,
Monitor the disk usage rate, and if it exceeds 80%,
The upper device is notified of that fact, and the new data is not created thereafter, and only the rewriting of the data is allowed.
And if the size increases due to the rewriting of data,
The increased amount can be stored in the 20% area.

However, in such a method, the usage rate will eventually exceed 100% (data overflow), data cannot be written to the disk device, and the system will be down. From the above, an object of the present invention is to provide a spare data storage unit so that the usage rate is a set value (for example, 100%).
%), The excess amount is stored in a spare storage unit so that data can be guaranteed even if the usage rate exceeds 100%.

[0011]

FIG. 1 is a diagram for explaining the principle of the present invention. Reference numeral 11 is a host device, for example, a channel, 12 is a disk controller, 12c is a cache memory, and 12e.
Is a non-volatile mapping table that stores the correspondence between CKD tracks (CCHH) and FBA block numbers, 13
Reference numeral a is an operating DASD, and reference numeral 13e is a backup DASD.

[0012]

The disk controller 12 compresses the data input from the higher-level device 11 and writes it in the DASD 13a, restores the data read from the DASD 13a, and transfers it to the higher-level device 11. In such a system, a spare data storage unit (for example, a spare DASD 13 that is invisible to the user in advance when data overflow occurs).
e) is provided. Then, the disk controller 12 is DAS
It is monitored whether the usage rate of D13a exceeds a set value (for example, 100%), and if it exceeds the set value, the overflow data is written to the spare DASD 13e and the upper device is notified that the data overflow has occurred. To do. From the above, the usage rate is 10
Data can be guaranteed even if it exceeds 0%

Further, the disk control device 12 is a working DA.
When the usage rate of the SD 13a exceeds the first set value (for example, 80%), writing of new data is prohibited and only rewriting is permitted, and when the data is rewritten, the usage rate is set to the second set value ( For example, 100%) is monitored, and if it exceeds the second set value, the amount exceeding the second set value is written in the spare data storage unit 13e and the spare is made corresponding to the CKD track. , Discrimination of active disk and FBA block number are registered in the mapping table 12e. Further, the disk control device 12 reads the host device 11 at the time of reading.
The information corresponding to the instructed CKD track is obtained from the mapping table 12e, and the backup included in the information,
F based on the FBA block number and the type of active disk
The BA block data is read out and restored, the restored data is written in the memory 12c, a desired record is searched in the memory and transferred to the host device 11. If you do this,
Even if the usage rate exceeds 100%, the data can be guaranteed, and since the data storage position is stored in the mapping table, the data can be read correctly and the correctness of the data can be guaranteed.

[0014]

EXAMPLES Overall configuration FIG. 2 is a configuration diagram of a computer system of the present invention. 10 is a host device (HST), 11 is a channel device (CH),
12 is a magnetic disk controller (DCU), 13a, a,
13b ... are magnetic disk devices in operation (DAS
D) and 13e are spare magnetic disk devices (DASD) invisible to the user. Magnetic disk controller 12
In FIG. 12, reference numeral 12a denotes a channel adapter (CA) that controls an interface with a host device, 12b denotes a device adapter (DA) that controls an interface with a DASD, 12
c is a cache memory (CM) provided to reduce access time, 12d is a compression / decompression unit that compresses data written in DASD and restores data read from DASD, and 12e is a CKD track (C
CHH) and the FBA block number are stored in a nonvolatile mapping table (MPT), 12f is a resource manager (RM) that controls the overall resource management and processing operations, and includes a channel adapter CA, a device adapter DA, The cache memory CM, the resource manager RM, and the compression / decompression unit are connected to each other by a bus 12g so that data can be exchanged between them.

The channel adapter 12a, the device adapter 12b, and the resource manager 12f are each composed of a microprocessor and have the structure shown in FIG. In the figure, 51 is a microprocessor (MP
U), 52 is a control storage unit (CS) having a RAM configuration, 53 is a control storage unit (CS) having a ROM configuration, 54 is a driver / receiver (DV / RV) connected to the bus 12g, and 55 is a bus interface logic ( BIL), 56 is a driver / receiver (D) connected to an external interface
V / RV), 57 is a buffer or table storage unit (T
S) and 58 are individual LSIs (gate arrays). In the mapping table 12e, as shown in FIG. 4, information for distinguishing an operating DASD from a backup DASD (operating DASD: A, B, ..., Backup DASD: E),
CKD track (CCHH) and CK on the track
The correspondence with the number of the FBA block that stores the FBA format data obtained by converting the D format data into the FBA format is stored. DV is the device address of DASD recognized by the host device.

Data Reading FIG. 5 is an explanatory diagram of a data reading procedure. DASD
It is assumed that the data is compressed and written in the (FBA disk device). When the seek command SK is generated from the channel device CH, the command is notified to the resource manager RM via the channel adapter CA.
The resource manager RM queues the seek command in the command queue, and after that, the seek operation is pretended to be completed and an operation end signal is returned to the channel device CH via the channel adapter CA. The reason for queuing is that there are multiple positioning commands (seek,
Set sector etc.) is queuing and search ID
This is because when a command is received, the commands that have been queued are executed together to speed up the process.

Upon receiving the seek operation end signal, the channel device CH issues the set sector command SS. This command is notified to the resource manager RM via the channel adapter CA. The resource manager RM queues the set sector command in the command queue, then makes it appear that the set sector operation has ended and returns an operation end signal to the channel device CH via the channel adapter CA. Upon receiving the set sector operation end signal, the channel device CH receives the search ID command SI.
Issue D. This command is notified to the resource manager RM via the channel adapter CA.

When the resource manager RM receives the search ID command SID, the device adapter DA receives the CKD track position (CCHH) included in the seek command.
Is issued to instruct staging, and a retry signal is issued to disconnect the channel device CH from the magnetic disk controller DCU. By receiving the retry signal,
The channel device CH is temporarily disconnected from the magnetic disk control device DCU, and during that time, it is possible to execute services for another magnetic disk control device. The channel device CH
Receives a retry signal, disconnects, and then receives a command request signal from the magnetic disk controller DCU, the command issued immediately before (search I in this example).
D command) is retransmitted.

When staging is instructed, the device adapter DA performs staging processing described later, reads the FBA format data (FBA block) corresponding to the instructed CKD track (CCHH) from DASD, and inputs it to the compression / decompression unit 12d. To do. The compression / decompression unit 12d restores the input compressed data and restores the cache memory 12
Expand to c. When the expansion of the restored data into the cache memory is completed, the device adapter DA notifies the resource manager RM of the staging end. As a result, the resource manager RM sends a command request signal to the channel device CH via the channel adapter CA. Upon receiving the command request, the channel device CH reissues the search ID command SID issued immediately before. When the channel adapter CA receives the search ID command SID, it searches for the designated record in the cache memory 12c, and upon completion of the search, sends an operation end signal of the search ID to the channel device CH. The channel device CH issues the read command RD when receiving the operation completion signal of the search ID. Channel adapter CA uses read command RD
When the record is received, the searched record is read from the cache memory 12c and transferred to the channel device CH.

FIG. 6 is a flow chart of the staging process of the device adapter DA. When staging is instructed by the resource manager RM (step 101), the device adapter DA refers to the mapping table 12e and the FBA corresponding to the instructed CKD track (CCHH).
The track number is retrieved (step 102). Then,
Read the retrieved FBA block number with DASD (F
(BA disk device) (step 103). D
The ASD reads the designated FBA block and inputs it to the device adapter DA. The device adapter DA inputs the input FBA block data to the compression / decompression unit 12d (step 104), and then the FB to be read.
It is checked whether the A block remains (step 105),
If any remain, the process returns to step 103, and the subsequent processing is repeated. On the other hand, when the reading of all the FBA blocks corresponding to the CKD track is completed, the compression / decompression unit 12d is instructed to perform the decompression (step 106). As a result, the compression / decompression unit 12d performs the decompression process on the compressed data and stores the decompressed data in the cache memory 12c (step 107).
When the expansion of the restored data in the cache memory 12c is completed, the device adapter DA notifies the resource manager RM of the staging completion (step 108).

Data Writing FIG. 7 is an explanatory diagram of a data writing procedure. Incidentally, DAS
It is assumed that data is compressed and written in D (FBA disk). When a seek command SK is issued from the channel device CH, the command is sent to the channel adapter CA.
Is notified to the resource manager RM via. The resource manager RM queues the seek command in the command queue, and after that, the seek operation is pretended to be completed and an operation end signal is returned to the channel device CH via the channel adapter CA. Upon receiving the seek operation end signal, the channel device CH issues the set sector command SS. This command is the channel adapter CA
Is notified to the resource manager RM via. The resource manager RM queues the set sector command in the command queue, then makes it appear that the set sector operation has ended and returns an operation end signal to the channel device CH via the channel adapter CA.

Upon receiving the set sector operation end signal, the channel device CH issues the search ID command SID. This command is notified to the resource manager RM via the channel adapter CA. Upon receiving the search ID command SID, the resource manager RM passes the CKD track position (CCHH) included in the seek command to the device adapter DA to instruct staging and issues a retry signal to the channel device CH.
By receiving this retry signal, the channel device CH is temporarily disconnected from the magnetic disk control device DCU. When staging is instructed, the device adapter DA executes the staging process described in FIG. 6, and the instructed C
DAS the FBA block on the KD track (CCHH)
It is read from D and input to the compression / decompression unit 12d. The compression / decompression unit 12d decompresses the input compressed data and expands it in the cache memory 12c.

When the expansion of the restored data into the cache memory is completed, the device adapter DA notifies the resource manager RM of the staging end. As a result, the resource manager RM sends a command request signal to the channel device CH via the channel adapter CA. Upon receiving the command request, the channel device CH reissues the search ID command SID issued immediately before. When the channel adapter CA receives the search ID command SID, it searches for the designated record in the cache memory 12c, and upon completion of the search, sends an operation end signal of the search ID to the channel device CH. The channel device CH issues the write command WD when receiving the operation end signal of the search ID. The channel adapter CA rewrites the searched record with the record of the write command, sends a write end signal to the channel device CH, and instructs the compression / decompression unit 12d to perform compression. As a result, the compression / decompression unit 12
d is 1 CKD stored in the cache memory 12c
The data for the track is compressed, and the end of compression is notified to the resource manager RM. In response to the compression end notification, the resource manager RM instructs the device adapter DA to write the compressed data to DASD (write back instruction). The device adapter DA performs the write-back processing described later according to the write-back instruction, and writes it to DASD.

Write Back Process FIGS. 8 to 10 are flowcharts of the write back process by the device adapter DA. When the device adapter DA receives the write back instruction (step 201), it compresses the compressed data into FBA blocks (step 202). Then, it is judged whether the data is new write data or rewrite data (step 203). Whether it is new or rewritten is determined by whether the FBA block number is registered in the mapping table 12e corresponding to the designated CKD track (CCHH). If it is rewriting, it is determined whether the compressed data size (FBA block number) after rewriting is larger than the data size (FBA block number) before rewriting (step 204). If they are equal, the FBA designated by the mapping table 12e is determined. Data is written in the block (step 205). If it is smaller, the data is written in a predetermined FBA block designated by the mapping table 12e, and the FBA block in which the data is not written is deleted from the mapping table 12e.

On the other hand, if the data is new data in step 203, it is judged whether the usage rate exceeds 80% by writing the data in DASD (step 211). The usage rate η is the total FBA of DASD.
Number of blocks is N, FBA where data is currently written
If the number of blocks is n, it can be calculated by η = (n / N) × 100 (%). If the usage rate exceeds 80%, the resource manager RM is notified of this (step 21).
2). As a result, when there is an access request from the host device, the resource manager RM sends status information indicating that the usage rate has exceeded 80% to the host device. Based on this status information, the higher-level device will only perform rewriting without writing new data thereafter.

FB for writing new data
Prepare the A block in the operating DASD (Step 2
13) Then, the correspondence relationship between the CKD track (CCHH) and the FBA block number is registered in the mapping table 12e (step 214). Thereafter, the FBA block number and the FBA block data are sequentially passed to the DASD according to the SCSI interface to write the FBA block data (step 215). If the data size is larger than that before rewriting in step 204,
It is judged whether the usage rate exceeds 80% (step 221),
If it does not exceed 80%, the F for the larger size
Prepare BA block in operational DASD (Step 2
22). Then, the number of the FBA block is registered in the CKD track (CCHH) of the mapping table 12e (step 223). Thereafter, the device adapter DA sequentially transfers the FBA block number and FBA block data corresponding to the CKD track (CCHH) to DASD according to the SCSI interface to write the FBA block data (step 224).

On the other hand, if the usage rate exceeds 80% in step 221, it is judged whether the usage rate exceeds 100% (step 225). If not, it is determined that the usage rate exceeds 80%. The resource manager RM is notified (step 226), and the processes after step 222 are repeated. When the usage rate exceeds 100%, the fact is notified to the resource manager RM (step 22).
7). As a result, the resource manager RM sends status information indicating that the usage rate has exceeded 100% to the upper device when an access request is issued from the upper device later. Next, the device adapter DA prepares an overflow FBA block in the spare DASD 13e (step 22).
8). Next, the number of the FBA block is additionally registered in association with the CKD track (CCHH) of the mapping table 12e (step 223). For example, C in FIG.
The FBA block number E1 of the spare DASD is additionally registered as in CHH = 0409.

Then, the CKD track (CCHH)
The FBA block number and the FBA block data corresponding to the above are sequentially passed to the operating DASD or the spare DASD according to the SCSI interface to write the FBA block data (step 224). In the above, the case where the DASD is provided as the spare storage unit has been described.
The spare area can be configured to be provided in the operating DASD. That is, up to 90% of DASD can be used as the operating area and the remaining 10% can be used as the spare area. This method has the advantage of being easy to control. Further, a spare storage unit can be provided in the memory in the magnetic disk control device, and this method is effective when the amount of overflow data is small. Further, although the case where the present invention is applied to the FBA disk has been described above, the present invention is not limited to the FBA disk, but can be applied to a CKD disk or the like. Although the present invention has been described above with reference to the embodiments, the present invention can be variously modified according to the gist of the present invention described in the claims, and the present invention does not exclude these.

[0029]

As described above, according to the present invention, in the disk control method for compressing the data input from the host device and writing it to the disk, and restoring the data read from the disk and transferring it to the host device, the spare data A storage unit is provided, and it is monitored whether the disk usage rate exceeds a set value (for example, 100%). If the disk usage rate exceeds the set value, the data is written to the spare data storage section, so that the usage rate is 1
Data can be guaranteed even if it exceeds 00%. Further, since the data storage position is stored in the mapping table, the data can be read correctly and the validity of the data can be guaranteed.

[Brief description of drawings]

FIG. 1 is a diagram illustrating the principle of the present invention.

FIG. 2 is a system configuration diagram of the present invention.

FIG. 3 is a configuration diagram of a channel adapter, a device adapter, and a resource manager.

FIG. 4 is a configuration diagram of a mapping table.

FIG. 5 is an explanatory diagram of a data reading procedure.

FIG. 6 is a flowchart of a staging process of a device adapter.

FIG. 7 is an explanatory diagram of a data writing procedure.

FIG. 8 is a flowchart (1) of a write-back process of the device adapter.

FIG. 9 is a flowchart (part 2) of the write-back process of the device adapter.

FIG. 10 is a flowchart (part 3) of the write-back process of the device adapter.

FIG. 11 is an explanatory diagram of a CKD format.

FIG. 12 is an explanatory diagram of an FBA format.

FIG. 13 is an explanatory diagram of a system using an FBA disk device.

FIG. 14 is an explanatory diagram of mapping.

[Explanation of symbols]

 11 ... Channel 12 ... Disk control device 12c ... Cache memory 12e ... Mapping table 13a ... Operating DASD 13e ... Spare DASD

Claims (2)

[Claims] 1. A disk control method for compressing data input from a host device and writing the data to a disk, and restoring data read from the disk and transferring the data to the host device, wherein a spare data storage unit is provided and a disk is used. A disk control method characterized by monitoring whether the rate exceeds a set value and, if it exceeds the set value, writing data to a spare data storage unit. 2. Data on a CKD track (CKD
Data), then the compressed data is FBA-blocked and written to the disk, and the FBA block data read from the disk is restored and a predetermined record is transferred to the host device. When the set value of 1 is exceeded, writing of new data is prohibited and only rewriting is permitted, and at the time of rewriting the data, it is monitored whether the usage rate exceeds the second set value (> the first set value). If the second set value is exceeded, the excess of the second set value is written in the spare data storage unit, and the spare, working disk, and FBA block number are associated with the CKD track. Is registered in the mapping table, and at the time of reading, information corresponding to the CKD track instructed by the higher-level device is obtained from the mapping table and included in the information. It is characterized in that FBA block data is read out and restored on the basis of the spare and operational disk types and the FBA block number, the restored data is written to the memory, the desired record is searched on the memory and transferred to the host device. Disk control method.