JPH11312058A - Storage subsystem - Google Patents

Abstract

(57) [Problem] To connect with a redundantly configured control device in a multi-host multi-path configuration, regardless of whether a logical volume is shared or not, it is possible to continue processing when a failure occurs in each part. Due to an increase in overhead in a normal state, a difference in performance from that in a single configuration has become a problem. When a disk array device having a dual controller configuration and a plurality of ports is connected to a plurality of hosts in a redundant configuration, the assignment of logical volumes is divided among the controllers, and the logical volumes are shared by the controllers. Even in the event of a host, bus, controller, or cache failure, a plurality of switching means at the time of failure are provided depending on the host connection environment, and in some cases, switching is not performed. In addition, even in the redundant configuration, it is possible to realize low overhead which is not different from the single configuration even in the normal state, and to prevent performance degradation in switching processing at the time of various failures.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a storage subsystem provided with a control device connected to a plurality of hosts or a single host having a plurality of ports. And technology for improving the performance of switching processing to an alternate system in the event of a failure.

[0002]

2. Description of the Related Art Japanese Patent Laid-Open Publication No. Hei 4-21 discloses a storage device such as a controller and a disk having redundancy.
There is a storage subsystem described in US Pat. This storage subsystem is constituted by a double system, one of which operates as an active system and the other of which operates as a standby system. Then, the storage information of the active disk device is copied to the standby disk device via the shared disk device accessible from both systems, or in the event of a failure of the active controller, the standby controller causes the active controller to copy the storage information to the standby disk device. By extracting the storage information of the disk device, the data integrity in the event of a controller and disk device failure has been improved.

[0003] Conventionally, a control device that shares a storage medium with a plurality of controllers and a plurality of ports has a configuration in which an arbitrary logical volume can be accessed from a plurality of paths, as seen in a backbone open market. In this configuration, all logical volumes can be uniformly accessed from any controller.

[0004]

With the increase in the scale of computer systems, the speed of data processing, and the increase in data capacity, there is a strong demand for higher performance, higher reliability, and higher availability for storage controllers. .

[0005] In the prior art, only the improvement of reliability and availability closed in the control device has been performed by the dual system on the control device side and the duplication of the cache. Considering the entire computer system, the duplex of the control device and the access bus for the host and the control device are reduced.
Redundancy is essential, and in order to achieve this, a multiple controller configuration and a multiple port configuration are required.

However, in such a configuration, in order to realize a configuration in which all logical volumes can be uniformly accessed from any controller, as in the conventional open system market, host data is transferred between a plurality of controllers. It is essential to share the cache management information that is temporarily stored. In order to share the cache management information, it is necessary for the processor to perform exclusive lock control and exclusive information management to perform exclusive control. The rate of performance degradation is large. Therefore, the only alternative was to sacrifice performance to improve reliability and availability.

[0007] In order to maintain the performance as compared with a single configuration without a redundant configuration while taking a redundant configuration as a system, lock control between controllers can be performed while all logical volumes can be accessed from any controller. , Exclusive control and the like are unnecessary. Therefore, as a normal access mode, a method of fixing the assignment of a logical volume to each controller can be considered.
In the event of a failure between the control device bus and the control device, it is not possible to substitute the processing for the assigned logical volume, and it cannot be said that the configuration is redundant. Therefore, it is necessary that each assigned task is realized with low overhead during normal times (normal times), and it is necessary to be able to take over the job when a failure occurs.

However, as a connection form to each controller, a single port is connected by a plurality of hosts, for example, by daisy chain connection. Alternatively, the controller has a plurality of ports, and each port is connected to a different host. Furthermore, when each host is connected to another controller as an alternate path, a failure on one host or a bus failure may cause a failure.
If you try to switch the logical volume to another controller and switch all of the logical volumes in charge of the other controller at once as a switch of the charge of the controller, the other hosts will access as normal processing. Can not be performed. Therefore, when connecting to a plurality of hosts, switching means suitable for each host and bus failure is required.

In recent years, a path switching function, which is a fault tolerant function, is being supported on the host side due to a strong need for fault tolerant. Representative examples include Alternate Link of HP and Sa
fe Path (path switching software).
In this function, the host side is connected to an alternate path, and when recognizing a failure in an arbitrary path, the processing is automatically transferred to another system path (the logical volume is transferred) and the processing is restarted. The recognition of a failure is often due to a timeout or the like, and if the processing is delayed for a certain period of time in the first I / O after the processing is transferred to another system, there is a high risk of being recognized as a failure. Therefore, in the switching process in charge of the controller, the I / O is received and recognized, but the subsequent switching process time is as follows:
It is an essential condition that it is performed in a short time in which no obstacle is recognized.

[0010]

In order to solve the above-mentioned problems, each controller has a plurality of bus connection ports and a cache mechanism for temporarily storing data, and a storage medium for storing data from a host. Is a control device composed of a plurality of controllers sharing a logical volume (L
UN) Assignment of charge and I to the non-charge controller
Means for specifying the method of switching the charge of the logical volume upon receipt of / O, specifying the method of allocating the cache to the logical volume, and specifying the single / double write to the cache; Processor, which controls switching and communication to another system controller at the time of request, means for controlling data transfer for switching between double writing / single writing to a cache, a cache memory as a controller, and a LUN unit Means to be divided and used, load information for each LUN, and LUN
Is realized by a processor that determines an optimal switching method.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 shows an example of the overall configuration of a system including a disk array device according to the present embodiment. 1, 10, 20, 30, and 40 are host computers, 5
Reference numeral 000 denotes a disk array device, which is connected to each host by a SCSI bus or the like. 1000, 20
00 is a controller unit having a dual configuration, 4000 is a disk device group for storing data from the host, 300
Reference numeral 0 denotes an operation panel for specifying a LUN and the like. The disk device group 4000 often has an array configuration.

Each controller has four ports 1010, 1020, 1030, 10
40, 2010, 2020, 2030, and 2040, and are connected to each port on the host side. Further, the transfer control units 1100 and 2100
It also controls data transfer between the host 300 and the host, and also controls single / double writing. Disk array control unit 1
Reference numerals 200 and 2200 control the controller A and the controller B as a whole.

FIG. 2 shows the configuration of the cache memory. The cache memories 1300 and 2300 store the management information 140
0,2400, Controller A area 1500,25
00, areas 1600 and 2600 for the controller B. 1400 and 2400, 1500 and 2500, respectively
1600 and 2600 are duplicated. Management information 1400, 2400 of cache memory 1300, 2300
Have the same configuration due to double writing, the configuration of the management information will be described using the cache memory 1400 on the controller A side as an example. The management information 1400 includes communication information 1410 between controllers and LUN information 1 for each controller.
420, data management information 1430 for managing data in the cache. The data management information 1430 is divided into two for each controller.

The cache memories 1300 and 2300 are preferably nonvolatile for reliability. If the host
If a redundant configuration is already used, for example, two disk array devices 5000 are prepared and one is used as a current device and the other is used as a spare device, the redundancy on the disk array side is intentionally requested. Operation panel 3
000 by the user's instruction with 000
When used as overwriting, 1500 and 1600 for controller A and 250 for controller B
0, 2600 can also be used. In this case, the use efficiency of the cache memory is doubled compared to the double writing. The single-write and double-write controls are performed by the disk array control units 1200 and 2200 by the transfer control units 1100 and 2100.
This is realized by setting 100 transfer modes.
The cache memories 1300 and 2300 are used as RD caches and WR caches, and can provide high performance. In order to use the WR cache function, double writing as described above is indispensable for guaranteeing data at the time of cache memory failure and controller failure.

The hosts each have two ports and are each connected to a different controller. In FIG. 1, the host 10
Are connected to another controller via the SCSI bus 100A, to the port A: 100A in the controller A: 1000, to the port B: 12-100B-port A: 2010, and so on. This connection is a typical example of a configuration having a redundancy that can be switched to another system in the event of a failure on the host side or a bus failure.

However, all the ports of the disk array device 5000 may be connected to different hosts.
In this case, there is no redundancy on the host side. Also, if the host
In such a two-path configuration, the alternate path may be connected to another port in the same controller. In the case of this configuration, in the event of a failure on the controller side, there is no alternate path and access is not possible, but on the host side, in the event of a bus failure, switching to the same controller is possible, switching processing described later, and request by communication Since neither process is required, the feature is that there is no performance degradation due to switching.

Hereinafter, the assignment of logical volumes in the configuration of FIG. 1 will be described. Assignment of a logical volume is performed exclusively between controllers. For example, LUNs 0, 1, 2, and 3 are assigned to the controller A1000, and LUNs 4, 5, 6, and 7 are assigned to the controller B2000. The assigned logical volume is the operation panel 3000
May be set in advance, or the LUN number of the received command may be assigned after the startup of the disk array device. As described above, by dividing the assigned LUN and the cache memory used for each controller, exclusive control or the like peculiar to multi-controller hybrid becomes unnecessary, and high performance can be provided.

A plurality of LUNs in the controller may be shared between a plurality of hosts, or may be independently accessed for each host. At this time, whether the cache memory allocation in the controller is centrally managed in the controller,
The user can specify whether to divide and use each LUN. When LUN is set independently for each host,
If only a LUN under an arbitrary host occupies the cache memory due to an access pattern on the host side, a difference in capability between the hosts, and the like, the performance of another LUN from another host decreases. To avoid this phenomenon,
The cache memory can be divided in advance for each LUN. This designation can also be set by the operation panel 3000. If the frequency of access for each LUN changes over time, for example, the
In the case of accessing the main LUNs 2 and 3 during the night of 0 and 1, it is possible to dynamically change the allocation in LUN units according to the load state. LUN
The designation of whether the unit assignment is statically arranged or dynamically assigned can also be set on the operation panel. According to the present invention, performance to a plurality of logical volumes in a multi-host connection environment can provide an equal service to the host in any access pattern from the host.

Next, an assignment method for each LUN and a method of changing the assignment method according to the load state will be described with reference to FIG. The minimum management unit for managing the use / non-use of the cache memory is called a segment. Data management information is LUN
The number of available segments 1431, the number of currently used segments 1432, the number of unused segments 143
3. The actual ADR of the cache for each segment,
It consists of segment management information 1434 for managing data ADR from the host. The segment management information has information equal in number to the number of segments. For example, if the size of the cache memory 1500 for the controller A is
Assume that it is for 0 segments. Now, it is assumed that four LUNs are assigned to the controller A and each of them is designated to be divided into 250 segments. Therefore, 250 is set as the number 1431 of usable segments for each LUN. When the I / O is received from the host and the cache is used, the segment management information 1434 is set by subtracting the number of used segments 1432 and the number of unused segments 1433. However, when the cache is used as an RD cache, it is managed as unused, and when used as a WR cache, it is managed as being used.

In addition, I / O for each LUN received in a certain period
Count the number of O. This information may be held in the disk array control unit or in the management information of the cache memory. Determine the load for each LUN based on this information,
L from LUN less than average I / O number for each LUN
The empty segment is transferred to UN. That is, the number of usable segments 1431 is changed, and the number of unused segments is 1
433, and by shifting the location of the segment management information. This change processing may be performed at regular intervals or when the I / O load from the host is low. With this control, a cache can be dynamically allocated according to the load of each LUN.

Next, switching processing at the time of a failure will be described. FIG. 3 shows an example of LUN assignment. In this assignment example, not only is the LUN divided for each controller,
LUNs are set independently for each host and each port.
With this configuration, when the SCSI bus 100A of the host 10 fails, access to the host-side LUN 0 is performed from the alternate path 100B. At this time, the controller B
“0” is compared with the LUN information 1420 for each controller to recognize that it is not in charge. At this time, when all LUNs for the controller A are switched to the controller B at once, the processing when the remaining hosts 20, 30, and 40 access the LUNs 1, 2, and 3 from the port A side is also performed on the controller B side. You will be switched to the person in charge. Therefore, when the controller A receives the access to the LUNs 1, 2, and 3 from the hosts 20, 30, and 40, it is determined that the access is to the non-assigned LUN, and the controller A, which was originally in charge of these LUNs, processes the LUN. The process of switching the charge occurs. As described above, in the method of switching LUNs under the controller at once,
As described above, in a multi-host connection environment, there is a problem that, when an arbitrary host system fails, switching processing frequently occurs, and performance is greatly reduced.

In the present invention, in order to solve the above problem,
It is possible to realize switching in LUN units. LU
When switching is performed in N units, the controller B
When receiving the I / O of LUN0, by setting only the target LUN to be the LUN0, the switching process does not occur even when accessing the controller A from another host, and the performance can be maintained.

On the other hand, in the above configuration, if the controller A fails, all the hosts switch the access right to the port B. Therefore, the batch switching is more in charge of the processing of the LUN than the switching for each LUN. Is reduced, and performance degradation due to LUN processing charge switching processing can be prevented.

In the case where the LUN is allocated as shown in FIG. 4, that is, when the LUN is shared between the ports in the controller, that is, used between the hosts, when the controller side fails, Although the processing can be continued by the batch switching, the switching operation occurs frequently even in the case of the host side failure or the SCSI bus failure even in the above two switching processings, and the performance is greatly reduced. For example, the SCSI bus 100A of the host 10
If a failure occurs, the host 10 issues the processing on the controller A side to the controller B using the alternate path 100B. That is, a request for LUNs 0, 1, 2, and 3 is issued to the controller B.

The controller B is responsible for the I / O of the non-assigned LUN.
The switching process is performed in response to receipt of the request. In the case of batch switching, all LUNs are transferred to the controller B side. After that, LUNs are sent from other hosts 20, 30, and 40.
When a processing request for 0, 1, 2, 3 is made via the controller A, batch switching occurs again. That is, the I / O from the host 10 and the hosts 20, 30, 4
Batch switching occurs at every I / O from zero. Even in LUN unit switching, exactly the same switching is repeated.

In order to solve the above problem, the present invention also proposes a method of not switching when an I / O to a non-assigned LUN is received. This method uses the communication information between controllers 1
Using 410, a request is made to the partner controller to secure a cache in the partner controller area and to read from and write to the disk device 4000. The partner controller is
When the request processing is completed, the inter-controller communication information 141
Report this to 0. Then, the own controller executes only the transfer with the host. According to the present invention, in a configuration in which a LUN in a controller is shared between a plurality of hosts, it is possible to continue processing from an alternate path in the case of a host or a bus failure without frequent switching operations.

In the present invention, the batch switching processing method, LU
The user can select an optimal method from among three modes of a switching processing method for each N and a request processing method to the partner controller without switching in consideration of the access environment of the own system configuration. This selection is made using the operation panel 3000.
Done through. Furthermore, even when the user selects the switching processing method for each LUN, if a failure occurs in the controller, the LUN processing charge may be switched by the batch switching method.

In the present invention, the disk array device 50
An automatic mode for automatically selecting an I / O pattern from the 00 side can also be set. The selection method in the automatic mode will be described below.

When assigning LUNs, LUs are assigned to each controller.
If N is assigned, that is, if it is not set for each port, it is impossible to determine whether the LUN in the controller is shared between the ports (hosts) or independent. Therefore, the type of the access LUN from each port for a certain period is counted, and it is determined whether it is a share after a certain period or independent. If independent, switch in LUN units is selected; if shared, request method is selected. still,
Since the access pattern may change with time, the above processing is performed after a certain period.

The batch switching is a dual configuration with a single port, and is effective when a single host configuration with an alternate path configuration is used. Also, in the multiple-host configuration as shown in FIG. 1, even when accessing by sharing the LUN, it is notified that the host has been switched. With the switching process of the arbitrary host,
This is effective in a host environment in which other hosts can be controlled to switch to an alternate path at once.

In the storage subsystem of the present invention, when assigning a LUN for processing to each port of a controller, a plurality of ports in one controller are assigned to handle processing of the same LUN. be able to. Also in this case, based on the method of allocating LUNs to a plurality of ports in the controller, the LUs that the controller is not responsible for processing from the host computer
When a processing request for N is received, a switching method for all LUNs, a switching method for each LUN, or a request method can be automatically selected.

Next, the processing procedure of the batch switching processing will be described. It is assumed that the controller B receives the non-assigned LUN. At this time, the fact is notified to the controller A in the inter-controller communication information 1410. Controller A
Recognizes the batch switching based on the information, and interrupts the current process. If interrupted, the fact is notified to the controller B using the inter-controller communication information 1410. Upon receiving the report from the controller A, the controller B changes the LUN information 1420 for each controller, and sets the controller B in charge of all LUNs under the controller A. Further, all the management information for the controller A in the data management information 1430 is also managed on the controller B side. As described above, in the case of batch switching, switching can be performed only by updating management information, and high-speed switching means can be provided.

Next, a switching method for each LUN will be described. The case where the controller B receives the non-assigned LUN and switches from the controller A will be described. It is assumed that the controller B has received the non-assigned LUN. At this time, the fact is notified to the controller A in the inter-controller communication information 1410. Based on the information, the controller A
Recognizing the switching of N units, the current processing is interrupted. The suspension guarantees that the controller A is not currently accessing the cache memory. If interrupted, the fact is notified to the controller B using the inter-controller communication information 1410. Controller B is
Controller A receives a report from Controller A
The controller B temporarily manages the data management information 1430 for the application and B. Controller B is the switching target LUN
The number of in-use segments 1432 and the management information 1434 of the segment are referred to, and the number of unused segments for each LUN in the data management information on the controller B side, and the controller The data in the segment in use of A is transferred to the controller B
To the free segment on the side. This copy processing itself can be realized by the transfer control unit 2100 in the controller B.

Along with copying the data itself, the data management information for each LUN of the LUN to be switched on the controller A side is deleted, and the segments originally allocated as the number of usable segments to the LUN are distributed to other LUNs. The distributed segments are registered as the number of unused segments. Conversely, in the data management information on the controller B side, LUN information to be switched is added. The L
The number of used segments is registered in the UN as the number of segments being used, and the number of available segments is partially shifted from the number of available segments in other LUNs and registered as empty segments. The number of available segments of other LUNs is also reduced partly, and the number of free segments allocated for copying is also reduced from the number of free segments. When this series of processing is completed, the controller B instructs the controller A to resume the processing and communicates with the inter-controller communication information 141.
Use 0 to convey.

Further, in order to realize this control, the number of segments in use in the cache, that is, the number of WR data not reflected in the disk unit 4000 is managed by the WR cache function, and reaches a certain amount or more. In such a case, by giving the highest priority to the writing process to the disk device, management is performed so as not to increase beyond a certain amount. This control is realized by the disk array control units 1200 and 2200 referencing the number of in-use segments for each LUN at a fixed cycle and, when determining that the number has reached a certain amount or more, giving priority to the write processing. . With this control, there is always a certain amount or more of free segments in the cache,
Switching by the copy method in LUN units becomes possible. This switching process requires a process of copying WR data between caches, so that the processing time increases depending on the cache mounting capacity and the amount of WR data at that time. However, in recent years, the capacity of the data bus has become very high,
Since the copying process between caches is generally one having a capacity of several hundred MB / S, the switching process time is completed in about several seconds, so there is no problem in consistency with the host.

[0037]

According to the present invention, in a storage subsystem having a plurality of ports for performing bus protocol control with a host per controller and having a dual controller and cache, in a connection between a plurality of hosts, a controller is provided for each controller. By independently assigning logical volumes to each other, in any connection form, exclusive control or the like by control of a plurality of controllers is not required, and higher performance can be provided as compared with a single configuration.
Also, in the event of a host-side failure, connection bus failure, or controller-side port failure, if the host connects an alternate path to another path in the same controller, switching processing is not required, and alternate processing without performance degradation is performed. It becomes possible. Further, when an alternate path is connected to another controller, there is redundancy even in the event of a controller failure, so that switching processing is possible in all failures. LU from host
Depending on the form of access to N, it is possible to provide an optimal switching means with a short switching time.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of a control device according to an embodiment of the present invention.

FIG. 2 is a configuration diagram of data in a cache of a controller according to an embodiment of the present invention.

FIG. 3 is an example of a method of allocating a logical volume among a plurality of hosts according to an embodiment of the present invention, in which an independent logical volume is allocated to each host and each port.

FIG. 4 is an example of a method of allocating a logical volume among a plurality of hosts according to an embodiment of the present invention, and is an example of sharing a logical volume between hosts.

FIG. 5 is a configuration diagram of management information in a cache memory according to an embodiment of the present invention.

[Explanation of symbols]

10, 20, 30, 40: Host computer 11, 21, 31, 41: Port A 12, 22, 32, 42: Port B 100A to 400A: SCSI bus A 100B to 400B: SCSI bus B 1010 to 1040, 201010 2040: Port A ~
D 1100, 2100: Transfer control unit 1200, 2200: Disk array control unit 1300, 2300: Cache memory 1000, 2000: Controller 3000: Operation panel 4000: Disk device group 5000: Disk array device

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Masaaki Kobayashi 2880 Kozu, Odawara-shi, Kanagawa Prefecture Storage Systems Division, Hitachi, Ltd.

Claims (13)

[Claims] 1. A storage device group having a storage area divided into a plurality of logical volumes, and a cache memory connected to a host computer and temporarily storing data transferred between the host computer and the storage device. A storage subsystem having a plurality of controllers having the plurality of controllers, wherein the plurality of controllers control the storage device group, wherein the controller is assigned the logical volume in charge of processing, and the cache memory A storage subsystem having a storage area. 2. The storage subsystem according to claim 1, wherein said controller divides a storage area of said cache memory for a plurality of said controllers into an area for a logical volume for which said controller is in charge of processing. A storage subsystem characterized by: 3. The storage subsystem according to claim 1, wherein when the controller receives a processing request from the host computer for a logical volume that is not in charge of processing, the controller is in charge of processing of the logical volume. A storage subsystem, which is in charge of processing of all logical volumes of the controller. 4. The storage subsystem according to claim 1, wherein when the controller receives a processing request from the host computer for a logical volume not in charge of processing, the controller is in charge of processing of the logical volume. A storage subsystem characterized by: 5. The storage subsystem according to claim 1, wherein when the controller receives a processing request for a logical volume not in charge of processing from the host computer, the controller communicates the contents of the processing request to the controller in charge of processing of the logical volume. A storage subsystem, wherein the controller that has received the communication performs processing on the logical volume, and communicates the processing result to the requesting controller. 6. The storage subsystem according to claim 1, further comprising an input means for selecting a processing method when the controller receives a processing request for a logical volume not in charge of processing from the host computer. Storage subsystem. 7. The control system according to claim 2, further comprising input means for designating a capacity of a cache memory area per controller and a capacity of an area for said logical volume. Storage subsystem with 8. The storage subsystem according to claim 1, wherein data of said plurality of logical volumes is stored in each of cache memories of said plurality of controllers. 9. The storage subsystem according to claim 1, wherein the data of the plurality of logical volumes is stored in each of the cache memories of the plurality of controllers, or the data of the logical volume for which each of the controllers is in charge of processing. A storage subsystem having input means for selecting whether to store only the storage subsystem. 10. The storage subsystem according to claim 1, wherein when a failure occurs in said controller, switching of processing of a logical volume assigned to said failed controller to assignment of a normally operating controller is performed. Characterized storage subsystem. 11. The storage subsystem according to claim 3, wherein, when the controller receives a processing request from the host computer for a logical volume that is not in charge of processing, the controller switches the charge by changing only the management information in the cache memory. A storage subsystem comprising a control device for performing the following. 12. The storage subsystem according to claim 4, wherein when the controller receives a processing request from the host computer for the logical volume not in charge of the processing, the storage device among the logical volumes to be switched on the cache memory. A storage subsystem for copying data not stored in a storage area of a group to a cache memory in a controller at a switching destination and switching a controller in charge of processing of the logical volume. 13. The storage subsystem according to claim 2, wherein said controller changes a storage area for a logical volume of said cache memory in accordance with a load of said logical volume. .