JP3134864B2 - Recording medium recording recovery system and recovery program for system combination device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a computer system including a plurality of processors (hereinafter referred to as "hosts") and shared resources between hosts shared by the respective hosts. The present invention relates to a recovery system of a system coupling device to be controlled (hereinafter, referred to as MSCP) and a recording medium recording a recovery program.

[0002]

2. Description of the Related Art Japanese Unexamined Patent Publication No. Sho.
When each of a plurality of information processing devices (corresponding to the host of the present invention) issues an exclusive access request for a shared resource, the lock information related to the exclusive access request is held in its own control table, and one In a computer system in which an access control device (corresponding to the MSCP of the present invention) holds lock information of the shared resource and exclusively controls access to the shared resource, if a failure occurs in the access control device, A recovery method of a conventional access control device in which each information processing device restores lock information of the access control device based on lock information held in a control table after the access control device is restored is disclosed.

[0003]

However, in this conventional access control device recovery method, a failed access control device is recovered, and the lock control information held by each information processing device is used to recover the access control device. There was a problem that the operation of the system could not be resumed until the lock information was restored.

[0004] An object of the present invention is to provide a computer system comprising a plurality of hosts and a plurality of MSCPs, and when one of the MSCPs fails during the operation of the computer system,
By distributing and holding the lock information held by all the MSCPs including the failed MSCP in the remaining MSCPs, the application program (AP) can be used without waiting for the failed MSCP to recover.
, That is, to continue the operation of the computer system.

Another object of the present invention is to provide a spare MSCP in a computer system, and if any of the MSCPs fails during the operation of the computer system, all the MSCPs including the failed MSCP are transmitted. By maintaining the held lock information in a distributed manner between the remaining MSCP and the spare MSCP, the performance related to the lock processing of the MSCP before the occurrence of the failure is maintained even after the occurrence of the failure.

Another object of the present invention is that when one new MSCP is added to the computer system, the lock information held by each MSCP before the addition is increased by one as a whole after the addition. An object of the present invention is to improve the performance of the MSCP locking process by performing the lock process by holding the MSCP in a distributed manner.

[0007]

According to a first aspect of the present invention, there is provided a recovery system for a system coupling device, comprising a plurality of hosts and a plurality of lock information holding lock information for a plurality of inter-host shared resources shared by the plurality of hosts. A system coupling device, wherein each of the plurality of hosts requests any of the system coupling devices to lock any of the inter-host shared resources, and responds to the request to lock the inter-host shared resources. The lock information notified from the locked system coupling device is stored, and each of the plurality of system coupling devices locks the inter-host shared resource requested by any one of the hosts, and locks the inter-host shared resource. And the lock information is notified to the requesting host, and one of the plurality of hosts is When any one of the system coupling devices fails, the lock information held in another normal system coupling device is erased, and each of the plurality of hosts communicates with the other normal system coupling device. Is required to lock the shared resources between hosts based on the lock information stored in the host.

According to a second aspect of the present invention, there is provided a recovery system for a system coupling device, comprising: a plurality of hosts; a plurality of system coupling devices which hold lock information for a plurality of shared resources between hosts shared by the plurality of hosts; Wherein each of the plurality of hosts requests any of the system coupling devices to lock any of the inter-host shared resources, and in response to the request, The lock information notified from the system coupling device that has locked is stored, and each of the plurality of system coupling devices locks the inter-host shared resource requested by any of the hosts and locks the inter-host shared resource. While holding lock information indicating that the lock information is notified to the requesting host, any of the plurality of hosts, the If any one of the number of system coupling devices fails, the lock information held in the other normal system coupling device is erased, and each of the plurality of hosts communicates with the other normal system. Requesting any one of the plurality of system coupling devices obtained by adding the spare system coupling device to the coupling device to lock the inter-host shared resource based on the lock information stored in the own host. And

A third system coupling device recovery system according to the present invention comprises a plurality of hosts and a plurality of system coupling devices which hold lock information for a plurality of shared resources between hosts shared by the plurality of hosts. , Each of the plurality of hosts requests any one of the system coupling devices to lock any of the inter-host shared resources, and responds to the request from the system coupling device that has locked the inter-host shared resource. Memorize the notified lock information,
Each of the plurality of system coupling devices locks the inter-host shared resource requested by any of the hosts, holds lock information indicating that the inter-host shared resource is locked, and transmits the lock information to the request source. Notify the host, any of the plurality of hosts, when recognizing that a new system coupling device has been connected, erase the lock information held in the previously provided system coupling device, the For each of the plurality of hosts, for any one of the plurality of system coupling devices obtained by adding the newly connected system coupling device to the previously provided system coupling device, the lock information stored in the own host is stored in the lock information. Requesting to lock the inter-host shared resource based on the request.

A fourth system coupling device recovery system according to the present invention includes a plurality of hosts and a plurality of system coupling devices which hold lock information for a plurality of inter-host shared resources shared by the plurality of hosts. , Each of the plurality of hosts requests any one of the system coupling devices to lock any of the inter-host shared resources, and responds to the request from the system coupling device that has locked the inter-host shared resource. Memorize the notified lock information,
Each of the plurality of system coupling devices locks the inter-host shared resource requested by any of the hosts, holds lock information indicating that the inter-host shared resource is locked, and stores the lock information and its own system. Notifying the requesting host of the coupling device identifier and the requesting host, and when any of the plurality of hosts recognizes that any one of the plurality of system coupling devices has failed, the other is notified. In the case where each of the plurality of hosts holds lock information including the identifier of the failed system coupling device, based on only the lock information, the other normal system coupling device It is characterized by requesting any one to lock the inter-host shared resource.

According to a fifth aspect of the present invention, there is provided a recovery system for a system coupling device, comprising: a plurality of hosts; a plurality of system coupling devices which hold lock information for a plurality of shared resources between hosts shared by the plurality of hosts; Wherein each of the plurality of hosts requests any of the system coupling devices to lock any of the inter-host shared resources, and in response to the request, The lock information notified from the system coupling device that has locked is stored, and each of the plurality of system coupling devices locks the inter-host shared resource requested by any of the hosts and locks the inter-host shared resource. While holding lock information indicating that the lock information and the identifier of the own system coupling device to the requesting host, When any one of the number of hosts recognizes that one of the plurality of system coupling devices has failed, the host notifies the other hosts of the failure, and each of the plurality of hosts fails. In the case where lock information including the identifier of the system coupling device obtained is held, based on only the lock information, a plurality of system coupling devices in which the spare system coupling device is added to the other normal system coupling device. It is characterized by requesting any one to lock the inter-host shared resource.

According to a sixth aspect of the present invention, there is provided a recovery system for a system coupling apparatus, comprising: a plurality of hosts and a plurality of lock information for a plurality of shared resources between hosts accessed by an application program executed on the plurality of hosts. Each of the plurality of hosts receives a lock request for any one of the inter-host shared resources from the application program executed on the host, and receives a lock request for the identifier of the inter-host shared resource. Executing a hash function to identify a system coupling device to which a lock instruction command corresponding to the lock request is issued, and issuing a lock instruction command for the inter-host shared resource to the identified system coupling device, The lock information notified from the specified system coupling device is stored, Each of the plurality of system coupling devices receives a lock command command issued from any of the hosts, and performs lock processing if the inter-host shared resource on which the command of the lock command is executed is not locked. Execute, and hold lock information indicating that the inter-host shared resource has been locked, and notify the lock information to the host that has issued the command of the lock instruction. If any one of the coupling devices fails, the lock information held in another normal system coupling device is deleted,
Each of the plurality of hosts executes a hash function on the identifier of the inter-host shared resource included in the lock information stored in its own host and issues a lock instruction command to the system coupling apparatus. , And issues a lock instruction command for the inter-host shared resource indicated by the lock information to the specified system coupling device.

A seventh aspect of the present invention provides a recovery system for a system combination device, comprising: a plurality of hosts and a plurality of lock information for a plurality of shared resources between hosts accessed by an application program executed on the plurality of hosts. Each of the plurality of hosts receives a lock request for any of the inter-host shared resources from the application program executed on the host, and A hash function is executed on the identifier of the shared resource to specify a system coupling device to which a command of a lock instruction corresponding to the lock request is issued, and the specified system coupling device issues a lock instruction for the inter-host shared resource to the specified system coupling device. Issue a command and notify from the specified system coupling device Lock information, and each of the plurality of system coupling devices receives a command of a lock command issued from any one of the hosts, and locks the inter-host shared resource on which the command of the lock command is executed. If not, a lock process is executed to retain lock information indicating that the inter-host shared resource has been locked, and to notify the lock information to the host that has issued the command of the lock command, If any one of the plurality of system coupling devices fails, the lock information held in the other normal system coupling device is deleted, and each of the plurality of hosts A hash function is executed on the identifier of the inter-host shared resource included in the lock information stored by the host to issue a lock command. The system linking device of the elephant is specified from the plurality of system linking devices obtained by adding the spare system linking device to the other normal system linking device. The lock command is issued for the lock command.

According to an eighth aspect of the present invention, there is provided a recovery system for a system coupling apparatus, comprising: a plurality of hosts and a plurality of lock information for a plurality of inter-host shared resources accessed by an application program executed on the plurality of hosts. Each of the plurality of hosts receives a lock request for any one of the inter-host shared resources from the application program executed on the host, and receives a lock request for the identifier of the inter-host shared resource. Executing a hash function to identify a system coupling device to which a lock instruction command corresponding to the lock request is issued, and issuing a lock instruction command for the inter-host shared resource to the identified system coupling device, The lock information notified from the specified system coupling device is stored, Each of the plurality of system coupling devices receives a lock command command issued from any of the hosts, and performs lock processing if the inter-host shared resource on which the command of the lock command is executed is not locked. Executing, and holding lock information indicating that the inter-host shared resource has been locked, and notifying the lock information to the host that has issued the command of the lock instruction, and any one of the plurality of hosts is newly connected to the system. Upon recognizing that the device is connected, the lock information held in the previously provided system coupling device is erased, and each of the plurality of hosts is included in the lock information stored in the own host. The system coupling device that executes the hash function for the identifier of the inter-host shared resource and issues the lock command A command of a lock command for a shared resource between hosts indicated by the lock information specified from a plurality of system coupling devices obtained by adding the newly connected system coupling device to the provided system coupling device. Is issued.

According to a ninth aspect of the present invention, there is provided a recovery system for a system combination device, comprising: a plurality of hosts and a plurality of lock information for a plurality of inter-host shared resources accessed by an application program executed on the plurality of hosts. Each of the plurality of hosts receives a lock request for any one of the inter-host shared resources from the application program executed on the host, and receives a lock request for the identifier of the inter-host shared resource. Executing a hash function to identify a system coupling device to which a lock instruction command corresponding to the lock request is issued, and issuing a lock instruction command for the inter-host shared resource to the identified system coupling device, The lock information notified from the specified system coupling device is stored, Each of the plurality of system coupling devices receives a lock command command issued from any of the hosts, and performs lock processing if the inter-host shared resource on which the command of the lock command is executed is not locked. Executing and holding lock information indicating that the inter-host shared resource has been locked, and notifying the lock information and the identifier of the own system coupling apparatus to the requesting host, and any one of the plurality of hosts, When recognizing that any one of the plurality of system coupling devices has failed, the fact is notified to another host, and each of the plurality of hosts includes an identifier of the failed system coupling device. When the lock information is held, a hash function is executed on the identifier of the inter-host shared resource included in the lock information, and the command of the lock instruction is executed. A system coupling device to be issued is specified from among the other normal system coupling devices, and a command of a lock command for the inter-host shared resource indicated by the lock information is issued to the specified system coupling device. .

According to a tenth aspect of the present invention, there is provided a recovery system for a system coupling apparatus, comprising: a plurality of hosts and a plurality of lock information for a plurality of inter-host shared resources accessed by an application program executed on the plurality of hosts. A system coupling device, comprising a spare system coupling device, each of the plurality of hosts receives a lock request for any of the inter-host shared resources from the application program executed on the host,
A hash function is executed on the identifier of the inter-host shared resource to specify a system coupling device to which a command of a lock instruction corresponding to the lock request is issued, and to the specified system coupling device, A lock command is issued, and the lock information notified from the specified system coupling device is stored, and each of the plurality of system coupling devices receives the lock instruction command issued from any of the hosts. If the inter-host shared resource on which the command of the lock instruction is to be executed is not locked, a lock process is executed to hold lock information indicating that the inter-host shared resource has been locked, and And the identifier of the own system coupling device to the requesting host, and any one of the plurality of hosts When recognizing that any one of the number of system coupling devices has failed, it notifies the other host of the failure, and each of the plurality of hosts locks including the identifier of the failed system coupling device. When the information is held, a hash function is executed on the identifier of the inter-host shared resource included in the lock information, and the system coupling device to which the command of the lock command is issued is connected to the other normal system coupling device. A lock command for a shared resource between hosts indicated by the lock information is specified from a plurality of system connection devices obtained by adding the spare system connection device to the device, and the specified system connection device is issued. .

The first recording medium of the present invention performs a process of requesting one of a plurality of system coupling devices to lock one of a plurality of inter-host shared resources shared by a plurality of hosts. Each of the plurality of hosts locks the inter-host shared resource requested by any of the hosts, holds lock information indicating that the inter-host shared resource has been locked, and transmits the lock information to the request source. The process of notifying the host of each of the plurality of system coupling devices is performed, and the process of storing the lock information notified from the system coupling device is performed by each of the plurality of hosts. If any one of the hosts fails, the processing for erasing the lock information held in another normal system coupling device is performed by the plurality of hosts. And performing a process of requesting one of the other normal system coupling devices to lock the inter-host shared resource based on the lock information stored by the host. Characterized in that a program to be executed by each of the above is recorded.

The second recording medium of the present invention performs a process of requesting one of a plurality of system coupling devices to lock one of a plurality of inter-host shared resources shared by a plurality of hosts. Each of the plurality of hosts locks the inter-host shared resource requested by any of the hosts, holds lock information indicating that the inter-host shared resource has been locked, and transmits the lock information to the request source. The process of notifying the host of each of the plurality of system coupling devices is performed, and the process of storing the lock information notified from the system coupling device is performed by each of the plurality of hosts. If any one of the hosts fails, the processing for erasing the lock information held in another normal system coupling device is performed by the plurality of hosts. The host can be shared by any one of the plurality of system coupling devices obtained by adding a spare system coupling device to the other normal system coupling device based on the lock information stored in the own host. A program for causing each of the plurality of hosts to perform a process of requesting to lock a resource is recorded.

The third recording medium of the present invention performs a process of requesting one of a plurality of system coupling devices to lock one of a plurality of inter-host shared resources shared by a plurality of hosts. Each of the plurality of hosts locks the inter-host shared resource requested by any of the hosts, holds lock information indicating that the inter-host shared resource has been locked, and transmits the lock information to the request source. Of each of the plurality of system coupling devices, causing each of the plurality of hosts to perform a process of storing the lock information notified from the system coupling device, and newly connecting the system coupling device. When recognizing that the lock has been performed, a process for erasing the lock information held in the system coupling device previously provided is performed by any of the plurality of hosts. Against done so, any of a plurality of system coupling device wherein by adding newly connected system coupling device to a system coupling device was equipped the prior to,
A program is recorded which causes each of the plurality of hosts to perform a process of requesting to lock the inter-host shared resource based on the lock information stored in the host.

The fourth recording medium of the present invention performs a process of requesting one of a plurality of system coupling devices to lock one of a plurality of inter-host shared resources shared by a plurality of hosts. Each of the plurality of hosts locks the inter-host shared resource requested by any of the hosts, holds lock information indicating that the inter-host shared resource has been locked, and stores the lock information and its own system. The process of notifying the requesting host of the identifier of the coupling device is performed by each of the plurality of system coupling devices, and the process of storing the lock information notified from the system coupling device is performed by each of the plurality of hosts. When recognizing that any one of the plurality of system coupling devices has failed, the process of notifying the other hosts of the failure is performed by the plurality of hosts. If the lock information including the identifier of the failed system coupling device is held, the information is shared by any of the other normal system coupling devices based on only the lock information. A program for causing each of the plurality of hosts to perform a process of requesting to lock a resource is recorded.

The fifth recording medium of the present invention performs a process of requesting one of a plurality of system coupling devices to lock one of a plurality of inter-host shared resources shared by a plurality of hosts. Each of the plurality of hosts locks the inter-host shared resource requested by any of the hosts, holds lock information indicating that the inter-host shared resource has been locked, and stores the lock information and its own system. The process of notifying the requesting host of the identifier of the coupling device is performed by each of the plurality of system coupling devices, and the process of storing the lock information notified from the system coupling device is performed by each of the plurality of hosts. When recognizing that any one of the plurality of system coupling devices has failed, the process of notifying the other hosts of the failure is performed by the plurality of hosts. In the case where lock information including the identifier of the failed system coupling device is held, a spare system coupling device is assigned to the other normal system coupling device based on only the lock information. A program is recorded which causes each of the plurality of hosts to perform a process of requesting any one of the plurality of system coupling devices to lock the inter-host shared resource.

A first recording medium group according to the present invention divides the program recorded on any one of the first to fifth recording media into a plurality of portions, and divides the plurality of portions into a plurality of recording media. It is characterized in that it is recorded.

[0023]

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

Referring to FIG. 1, according to a first embodiment of the present invention, a host 11, 1 operating under program control is provided.
, 1k and their hosts 11, 12, ..., 1k
, And 2m, which are connected to each other. It is assumed that there are two or more MSCPs.

As modifications of the first embodiment and the second to fourth embodiments described below, the hosts 11, 12,..., 1k and the MSCPs 21, 22,.
2 m and recording medium 3 such as a magnetic disk device or a semiconductor memory
0 via a signal line, and the hosts 11, 12,.
, 1k and the MSCPs 21, 22,..., 2m may execute the following processing by a program loaded from the recording medium 30 via a signal line.

Each host 11, 12,..., 1k includes lock control means 111, 121,..., 1k1 (hereinafter collectively referred to as lock control means 1n1) and lock information storage means 112, 122,. Hereinafter, these are collectively referred to as lock information storage unit 1n2) and lock information storage unit 11
, 1k3 (hereinafter collectively referred to as lock information storage unit 1n3).

The lock control means 1n1 is connected to the hosts 11, 1
2. In response to a lock request or an unlock request issued by an AP (not shown) executed in 1k for exclusive control of a shared resource between hosts (not shown), a lock command corresponding to the lock request or the unlock request or The command of the unlock instruction is MSCP21, 22, ..., 2
Issued for m.

The lock information storage means 1n2 returns the lock command or unlock command issued to the MSCP 21, 22,... 2m by the lock control means 1n1 from the MSCP 21, 22,. The lock information stored in the lock information storage unit 1n3 is updated based on the execution result of the command.

Next, the operation of the first embodiment will be described.

First, the operation when the lock control unit 1n1 issues a lock command command for the inter-host shared resource will be described with reference to FIGS. FIG.
FIG. 7 is a diagram showing a state in which the inter-host shared resource is locked by a lock command issued by the lock control unit 1n1. FIG. 3 is a diagram showing a state in which the inter-host shared resource to be locked by the lock command issued by the lock control means 1n1 has already been locked by the lock command issued by the other lock control means 1n1. is there.

When an AP executed on the host 11 issues a lock request for the inter-host shared resource a,
The lock control unit 111 issues a lock command to one MSCP (FIG. 2A). The MSCP to which the command of the lock instruction is issued executes a hash function on the identifier of the inter-host shared resource, for example, the identification number, so that any one is determined. In the present embodiment, it is assumed that a lock command has been issued to the MSCP 22. The MSCP 22 that has received the lock instruction command locks the inter-host shared resource a after confirming that no host has locked the inter-host shared resource a, and “the host 11 locks the inter-host shared resource a. 2 is stored (see FIG. 2).
(2)) and notifies the host 11 that the lock command has succeeded (FIG. 2 (3)). Lock control means 111
Receives the notification indicating that the lock command has succeeded, causes the lock information storage unit 112 to store information indicating that "the shared resource a between hosts has been locked" in the lock information storage unit 113 (FIG. 2 (4)).

Subsequently, when an AP executed on the host 12 issues a lock request for the same inter-host shared resource a, the lock control means 121 sets the MSCP2
2 (if the hash function is executed for the same shared resource identifier between the hosts, the same MSCP is selected if there is no change in the MSCP configuration) (FIG. 3 (1)). The MSCP 22 that has received the lock command receives the inter-host shared resource “a” and
3 and suspends the execution of the command of the lock instruction, and stores information indicating that "host 12 is waiting for lock of shared resource a between hosts a" (FIG. 3).
(2)), the host 12 is notified that the lock command is waiting (FIG. 3 (3)). Lock control means 12
1 receives the notification that the lock command has been waited, and stores the “shared resource a between hosts a” in the lock information storage unit 122.
Lock waiting ”is stored in the lock information storage unit 123 (FIG. 3D).

Next, an operation when the lock control means 1n1 issues a command of an unlock instruction for the inter-host shared resource will be described with reference to FIGS.

In the state shown in FIG. 3, when the AP executed on the host 11 issues an unlock request for the inter-host shared resource a, the lock control unit 111
22 (if the hash function is executed for the same shared resource identifier between the hosts, the same MSCP is selected if there is no change in the MSCP configuration) (FIG. 4 (1)). . Upon receiving the unlock command, the MSCP 22 unlocks the inter-host shared resource a, and erases the information of “the host 11 has locked the inter-host shared resource a” stored therein (see FIG. 4 (2)
a)), and notifies the host 11 that the unlock command has succeeded (FIG. 4 (3a)). Lock control means 11
1 receives the notification that the unlock command has succeeded, and causes the lock information storage unit 112 to delete information indicating that “the host 11 has locked the inter-host shared resource a” from the lock information storage unit 113 (FIG. 4). (4a)).

On the other hand, since the MSCP 22 stores the information that "the host 12 is waiting for the lock of the inter-host shared resource a", it is issued from the host 12 after notifying the host 11 that the unlock command has succeeded. Shared resource a between hosts a
Is executed, and the inter-host shared resource a is locked. As a result, the MSCP 22 deletes the information “the host 12 has waited for the lock of the inter-host shared resource a” stored therein, and newly stores the information “the host 12 has locked the inter-host shared resource a”. (Fig. 4
(2b)), the host 12 is notified that the lock command has succeeded (FIG. 4 (3b)). Lock control means 12
1 receives the notification that the lock command has succeeded, and stores the information of “waiting for lock of the shared resource a between hosts” stored in the lock information storage unit 123 in the lock information storage unit 122 to the “shared resource between hosts”. a is locked "(FIG. 4 (4b)).

Finally, regarding the operation at the time of the MSCP failure,
This will be described with reference to FIGS.

Referring to FIG. 5, the inter-host shared resource a is locked by the host 12 and the host 11 is waiting for lock, the inter-host shared resource b is locked by the host 11 and the host 1k is waiting for lock, Resource c
Is assumed to be locked by the host 1k and the host 12 is in a lock waiting state. In this state,
A case where a certain MSCP (the MSCP 22 in this embodiment) has failed will be described.

A host (in this embodiment, host 1
When the lock control unit 111 detects the failure of the MSCP 22 (FIG. 6 (2)) (hereinafter, this host 11 is referred to as a representative host), the lock control unit 111 notifies all the other hosts of the failure (FIG. 6). 6 (3)). The lock control unit 111 of the host that has issued the notification and the lock control unit 1n1 of all other hosts that have received the notification transmit a lock request or an unlock request for the inter-host shared resource from the AP executed by each host. Even if issued, the lock command or the unlock command is not issued to the MSCP, and the lock process or the unlock process for the inter-host shared resource is suspended (FIG. 6 (4)). As a result, the operation of the system is temporarily stopped, and the execution of the AP that accesses the MSCP is also temporarily stopped. Lock control means 11 of representative host 11
1 declares the start of the recovery process to all the MSCPs that have not failed (FIG. 6 (5)). The MSCP having received this declaration erases all the lock information stored therein (this is referred to as initialization) (FIG. 6 (6)).

Lock control means 111 of representative host 11
Sends the notification of the completion of the MSCP initialization to the lock control means 1n1 of all the other hosts when the initialization of all the non-failed MSCPs is completed (FIG. 7 (1)). The lock control unit 111 of the representative host 11 and the lock control units 1n1 of all the other hosts that have received this notification lock the shared resources among the hosts among the information stored in the lock control information storage unit 1n3. The lock command is issued to one of the non-failed MSCPs based on the information indicating that the lock has been performed (FIG. 7 (2)). One of the MSCP to which the command of the lock instruction is issued is determined by executing a hash function on the identifier of the inter-host shared resource, for example, the identification number. When the lock control means 1n1 of all other hosts has issued the lock command, it notifies the lock control means 111 of the representative host 11 that the reproduction of the lock information has been completed (FIG. 7 (3)).

Lock control means 111 of representative host 11
When the lock control unit 1n1 of all the other hosts receives the notification that the reproduction of the lock information is completed, the lock control unit 1n1 of all the other hosts notifies the lock control unit 1n1 of the start of the reproduction of the information waiting for the lock. (FIG. 8 (1)). The lock control unit 111 of the representative host 11 sends this notification, and the lock control units 1n1 of all the hosts other than the representative host receive the notification of the start of reproduction of the information waiting for the lock, and then store the lock information. The lock command is issued to the MSCP based on the lock waiting information in the unit 1n3 (FIG. 8 (2)). Also in this case, one of the MSCPs to which the command of the lock instruction is issued is determined by executing a hash function on the identifier of the inter-host shared resource, for example, the identification number. When all the lock waiting information is reproduced, the lock control means 1n1 of all other hosts notifies the lock control means 111 of the representative host 11 that the reproduction of the lock waiting information is completed (FIG. 8 (3)). .

Lock control means 111 of representative host 11
Upon receiving notification from the lock control means 1n1 of all the other hosts that the reproduction of the lock waiting information has been completed, the lock control means 1n1 of all the other hosts is notified of the resumption of the lock control (FIG. 9 ( 1)). The lock control unit 111 of the representative host 11 and the lock control units 1n1 of all the other hosts that have received the notification resume the reception of the lock request or the unlock request for the inter-host shared resource from the AP executed on the host. (Fig. 9
(2)). As a result, the operation of the system that has been temporarily stopped is restarted.

With the above, the operation of the first embodiment of the present invention is completed.

According to the first embodiment of the present invention, when one MSCP fails, the lock information held by each host is retransmitted to another MSCP, thereby waiting for the restoration of the failed MSCP. And the operation of the computer system can be continued.

Next, a second embodiment of the present invention will be described. The second embodiment is a case where a spare MSCP exists in the computer system of the first embodiment.
This process will be described with reference to FIG. 10 with respect to differences from the first embodiment. In the present embodiment, the number of MSCPs other than the spare may be one instead of a plurality.

The MSCP (in this embodiment, MSCP2
When the lock control unit 111 of the host (the host 11 (representative host in the present embodiment)) recognizes the failure (2) in FIG. 10 (1) in FIG. 10 (2) in FIG. It is checked whether a spare MSCP exists in the computer system (FIG. 10 (3)). In the present embodiment, it is assumed that a spare MSCP 2p exists. As in the first embodiment, MSCP2 is sent to all other hosts.
2 is notified (FIG. 10 (4)), and lock processing and unlock processing of the inter-host shared resource are suspended in all hosts (FIG. 10 (5)). The lock control unit 111 of the host 11 issues a recovery start declaration to the MSCP other than the failed MSCP (FIG. 10 (6)), and holds the recovery start declaration in the MSCP other than the failed MSCP as in the first embodiment. Is initialized (FIG. 10 (7)). Further, a spare MSCP is added to an MSCP other than the failed MSCP.
2p is added, and the lock information and the lock waiting information are reproduced in those MSCPs. Here, the MSCP to which the command of the lock command is issued is MSCP other than the failed MSCP.
The target is the one obtained by adding the spare MSCP 2p to the SCP, and one of them is determined by executing a hash function on the identifier of the inter-host shared resource, for example, the identification number.

Thus, the operation of the second embodiment of the present invention is completed.

According to the second embodiment of the present invention, the first
In the same manner as in the embodiment, when a failure occurs in a certain MSCP, the lock information held by each host is retransmitted to another MSCP so that the operation of the computer system can be performed without waiting for the recovery of the failed MSCP. It has the effect of being able to continue, and furthermore, a spare MSCP is added to the first embodiment, and lock information or lock waiting information is reproduced without changing the number of MSCPs.
This has the effect that the performance before the occurrence of a failure related to the lock processing of the CP can be maintained.

Next, a third embodiment of the present invention will be described. In the third embodiment, an MSCP is added to the computer system according to the first embodiment. This process will be described with reference to FIG. 11 while referring to differences from the first embodiment. In the present embodiment, the number of MSCPs excluding the MSCP to be added may be one instead of a plurality.

In the first embodiment, the recovery process of the MSCP is started when an MSCP failure occurs. In the present embodiment, the MSCP is added when the MSCP is added.
The process of reconstructing the lock information stored in the SCP is started. The fact that the MSCP 2q has been additionally connected to the computer system indicates that one of the hosts (the host 11 in the present embodiment)
When the lock control unit 111 (representing the host) recognizes (FIG. 11A), as in the first embodiment,
The lock control unit 111 notifies the other host that the MSCP 2q has been added (FIG. 11 (2)), and suspends the lock processing and unlock processing of the inter-host shared resource in all the hosts (FIG. 11 (3) )). The lock control unit 111 of the host 11 issues a recovery start declaration to all the MSCPs (FIG. 11 (4)), and
The information held in P is initialized (FIG. 11 (5)).
Further, the lock information and the lock waiting information are reproduced for all the MSCPs to which the MSCP 2q is added. Here, the MSCP from which the command of the lock command is issued is MSCP.
All MSCPs to which 2q is added are targeted, and any one is determined by executing a hash function on the identifier of the inter-host shared resource, for example, the identification number.

Thus, the operation of the third embodiment of the present invention is completed.

According to the third embodiment of the present invention, the first
Instead of the failure of the MSCP as in the embodiment of the present invention, the lock information held by each MSCP before the addition is added to the whole information after the addition of the MSCP newly to the computer system. MSC increased by one
MSCP to perform lock processing by holding and distributing in P
Has the effect of improving the performance related to the lock processing.

Next, a fourth embodiment of the present invention will be described, focusing on differences from the first and second embodiments.

Each MSCP stores information indicating that “the host 1k has locked the inter-host shared resource x” when the inter-host shared resource is locked, and also indicates that the lock command to the host 1k is successful. Notify the MSCP identification number.

A host (host 11 in this embodiment)
(Representative host)) lock control means 111
When a failure of the CP (the MSCP 22 in the present embodiment) is detected, the failure is notified to all other hosts.

In the first and second embodiments, the lock control means 1n1 of all the hosts suspends the lock processing or the unlock processing for the inter-host shared resource, and retains the lock processing or unlock processing for all the MSCPs. The present embodiment is characterized in that only the lock information relating to the failed MSCP 22 is reproduced without erasing the lock information held by the normal MSCP.

That is, the lock control means 111 of the representative host 11 and the lock control means 1 of all the other hosts
n1 is information (MSCP) related to the failed MSCP 22 in the information stored in the lock control information storage unit 1n3.
22 (attached with an identification number of 22), a lock command is issued to any non-failed MSCP based on the information. One of the MSCP to which the command of the lock instruction is issued is determined by executing a hash function on the identifier of the inter-host shared resource, for example, the identification number.

Thus, the operation of the fourth embodiment of the present invention is completed.

According to the fourth embodiment of the present invention, the first
And, as in the second embodiment, the lock information held by all the MSCPs is once erased,
Instead of reproducing the lock information for the SCP, the lock information held by the normal MSCP is not erased, and only the lock information held by the failed MSCP is reproduced to another normal MSCP. As a result, compared to the first and second embodiments, the elimination of the lock information held by the MSCP has an effect that the recovery process can be performed efficiently.

In the first to fourth embodiments,
In order for the lock control unit 1n1 to specify the MSCP to which the command of the lock command is reissued, a hash function is executed on the identifier of the inter-host shared resource to which the command of the lock command is issued, for example, the identification number, Any one
Although one MSCP has been specified, the method of specifying the MSCP is not limited to executing the hash function, and the command of the lock instruction is
May be issued in order, or may be issued preferentially to the MSCP with a low (low-frequency) lock processing load at that time.

[0060]

As described above, according to the recovery system of the system coupling device of the present invention, even if the MSCP fails during the operation of the computer system, the lock information held by all the MSCPs including the failed MSCP can be obtained. The rest of the MS
By distributing and holding the data in the CP, the AP can be used without waiting for the failed MSCP to recover.
, Ie, the operation of the computer system can be continued.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a first exemplary embodiment of the present invention.

FIG. 2 is a diagram showing a state in which a shared resource between hosts is locked by a lock command issued by a lock control unit 1n1 in the first embodiment of the present invention.

FIG. 3 shows a first embodiment of the present invention in which a shared resource between hosts which is to be locked by a lock instruction command issued by a lock control unit 1n1 has already issued a lock instruction command issued by another lock control unit 1n1. FIG. 6 is a diagram showing a state in which the lock has been performed.

FIG. 4 is a diagram showing an operation when the lock control unit 1n1 issues a command of an unlock instruction for a shared resource between hosts in the first embodiment of the present invention.

FIG. 5 shows the MSCP according to the first embodiment of the present invention.
It is a figure which shows operation | movement of the recovery process.

FIG. 6 shows the MSCP according to the first embodiment of the present invention.
It is a figure which shows operation | movement of the recovery process.

FIG. 7 shows the MSCP according to the first embodiment of the present invention.
It is a figure which shows operation | movement of the recovery process.

FIG. 8 shows the MSCP according to the first embodiment of the present invention.
It is a figure which shows operation | movement of the recovery process.

FIG. 9 shows the MSCP according to the first embodiment of the present invention.
It is a figure which shows operation | movement of the recovery process.

FIG. 10 shows the MSC according to the second embodiment of the present invention.
It is a figure showing operation of recovery processing of P.

FIG. 11 shows an MSC according to the third embodiment of the present invention.
It is a figure showing operation of recovery processing of P.

, 1k host 111, 121,..., 1k1 lock control means 112, 122,..., 1k2 lock information storage means 113, 123,. , 2m, 2p, 2q MSCP 30 recording medium

──────────────────────────────────────────────────続 き Continued on front page (58) Field surveyed (Int. Cl. ⁷ , DB name) G06F 15/177 678 G06F 11/20 310

Claims (16)

(57) [Claims] 1. A system comprising: a plurality of hosts; and a plurality of system coupling devices which hold lock information for a plurality of inter-host shared resources shared by the plurality of hosts, wherein each of the plurality of hosts includes: Requesting the system coupling device to lock any of the inter-host shared resources, storing the lock information notified from the system coupling device that has locked the inter-host shared resource in response to the request; Each of the system coupling devices locks the inter-host shared resource requested by any of the hosts, holds lock information indicating that the inter-host shared resource has been locked, and transmits the lock information to the requesting host. Notify any of the plurality of hosts, and if any one of the plurality of system Erasing the lock information held in the system coupling device, wherein each of the plurality of hosts is
A recovery system for a system coupling device, which requests to lock a shared resource between hosts based on lock information stored in its own host. 2. The system according to claim 1, further comprising a spare system coupling device, wherein each of the plurality of hosts is connected to one of the plurality of system coupling devices obtained by adding the spare system coupling device to the other normal system coupling device. 2. The recovery system according to claim 1, wherein a request is made to lock the inter-host shared resource based on the lock information stored in the own host. 3. When one of the plurality of hosts recognizes that a new system coupling device has been connected, the host deletes the lock information held in the previously provided system coupling device. Each of the plurality of hosts, for any one of the plurality of system coupling devices obtained by adding the newly connected system coupling device to the previously provided system coupling device, based on the lock information stored in the host. 2. The recovery system according to claim 1, wherein a request is made to lock the shared resources between the hosts. 4. A system comprising: a plurality of hosts; and a plurality of system coupling devices that hold lock information for a plurality of inter-host shared resources shared by the plurality of hosts, wherein each of the plurality of hosts includes: Requesting the system coupling device to lock any of the inter-host shared resources, storing the lock information notified from the system coupling device that has locked the inter-host shared resource in response to the request; Each of the system coupling devices locks the inter-host shared resource requested by any of the hosts, holds lock information indicating that the inter-host shared resource has been locked, and stores the lock information and the information of the own system coupling device. An identifier is notified to the requesting host, and one of the plurality of hosts is determined to be one of the plurality of system coupling devices. When recognizing that the failure has occurred, the other hosts are notified of the fact.If each of the plurality of hosts holds lock information including the identifier of the failed system coupling device, only the lock information is stored. A recovery system for the system coupling device, which requests one of the other normal system coupling devices to lock the inter-host shared resource based on the following. 5. A system further comprising a spare system coupling device, wherein each of the plurality of hosts holds lock information including an identifier of a failed system coupling device, based on only the lock information. 5. The system coupling apparatus according to claim 4, wherein one of the plurality of system coupling apparatuses obtained by adding the spare system coupling apparatus to another normal system coupling apparatus is required to lock the inter-host shared resource. Recovery system. 6. A plurality of hosts, comprising: a plurality of hosts; and a plurality of system coupling devices which hold lock information on a plurality of shared resources between hosts accessed from an application program executed on the plurality of hosts, Receives a lock request for any of the inter-host shared resources from the application program executed on the host, executes a hash function on the identifier of the inter-host shared resource, and responds to the lock request. A system coupling device to which the command of the lock instruction is to be issued is specified, a command of the lock instruction for the inter-host shared resource is issued to the specified system coupling device, and the lock information notified from the specified system coupling device. Wherein each of the plurality of system coupling devices is: When a command of the lock command issued from the host is received and the shared resource between the hosts to execute the command of the lock command is not locked, a lock process is executed, and it is determined that the shared resource between the hosts has been locked. While notifying the host that has issued the command of the lock instruction, when any one of the plurality of hosts fails in any one of the plurality of system coupling devices. Then, the lock information held in another normal system coupling device is erased, and each of the plurality of hosts hashes with the identifier of the inter-host shared resource included in the lock information stored in its own host. The system coupling device to execute the function and issue the command of the lock instruction is specified from among the other normal system coupling devices, and Recovery system of system coupling device, characterized in that the the system coupling device issues a command of the lock instruction to the inter-host shared resource indicated by the lock information. 7. The system further comprises a spare system coupling device, wherein each of the plurality of hosts executes a hash function on an identifier of the inter-host shared resource included in the lock information stored in the own host to lock the plurality of hosts. The system coupling device to which the command of the command is issued is specified from among the plurality of system coupling devices obtained by adding the spare system coupling device to the other normal system coupling device, and the lock information is transmitted to the specified system coupling device. 7. The recovery system for a system-coupling device according to claim 6, wherein a command of a lock instruction for the inter-host shared resource indicated by is issued. 8. When any one of the plurality of hosts recognizes that a new system coupling device has been connected, the host deletes the lock information held in the previously provided system coupling device. Each of the above-mentioned hosts is provided with a system coupling device to which a hash function is executed on the identifier of the inter-host shared resource included in the lock information stored in the own host to issue the command of the lock instruction. From the plurality of system coupling devices obtained by adding the newly connected system coupling device to the system coupling device, and issuing a lock command command for the inter-host shared resource indicated by the lock information to the specified system coupling device. 7. The recovery system for a system combination device according to claim 6, wherein: 9. A host system comprising: a plurality of hosts; and a plurality of system coupling devices that hold lock information for a plurality of shared resources between hosts accessed from an application program executed on the plurality of hosts, Receives a lock request for any of the inter-host shared resources from the application program executed on the host, executes a hash function on the identifier of the inter-host shared resource, and responds to the lock request. A system coupling device to which the command of the lock instruction is to be issued is specified, a command of the lock instruction for the inter-host shared resource is issued to the specified system coupling device, and the lock information notified from the specified system coupling device. Wherein each of the plurality of system coupling devices is: When a command of the lock command issued from the host is received and the shared resource between the hosts to execute the command of the lock command is not locked, a lock process is executed, and it is determined that the shared resource between the hosts has been locked. And the lock information and the identifier of the own system coupling device are notified to the requesting host, and one of the plurality of hosts is any one of the plurality of system coupling devices. When the host recognizes that the failure has occurred, the host notifies the other hosts of the failure, and when each of the plurality of hosts holds the lock information including the identifier of the failed system coupling device, A hash function is executed on the identifier of the inter-host shared resource included in the Recovery system of system coupling device characterized by identifying from the stem coupling device issues a command of the lock instruction to the inter-host shared resource indicated by the lock information to the specified system coupling device. 10. A backup system coupling device, wherein when each of the plurality of hosts holds lock information including an identifier of a failed system coupling device, the plurality of hosts share the lock information included in the lock information. A hash function is performed on a resource identifier, and a system coupling device to which a command of a lock instruction is issued is included in a plurality of system coupling devices obtained by adding the spare system coupling device to the other normal system coupling device. 10. The recovery system for a system-coupling device according to claim 9, wherein a command of a lock command for the inter-host shared resource indicated by the lock information is issued to the specified system-coupling device. 11. A process for requesting one of a plurality of system coupling devices to lock one of a plurality of inter-host shared resources shared by a plurality of hosts is performed on each of the plurality of hosts. Locking the inter-host shared resource requested by any of the hosts, holding lock information indicating that the inter-host shared resource has been locked, and notifying the requesting host of the lock information. Causing each of the plurality of system coupling devices to perform a process of storing the lock information notified from the system coupling device, and causing one of the plurality of system coupling devices to fail In this case, a process of erasing the lock information held in another normal system coupling device is performed by one of the plurality of hosts, For any normal system coupling device,
A recording medium on which is recorded a program for causing each of the plurality of hosts to perform a process of requesting to lock a shared resource between hosts based on lock information stored in the host. 12. An inter-host shared resource for one of a plurality of system coupling devices obtained by adding a spare system coupling device to the other normal system coupling device based on lock information stored in the host. 12. The recording medium according to claim 11, wherein a program for causing each of the plurality of hosts to perform a process of requesting locking of the host is recorded. 13. When one of the plurality of hosts recognizes that a new system coupling device has been connected, the process deletes the lock information held in the previously provided system coupling device. A shared resource between hosts based on lock information stored in its own host for any of a plurality of system coupling devices obtained by adding the newly connected system coupling device to the previously provided system coupling device. 12. The recording medium according to claim 11, wherein a program for causing each of the plurality of hosts to perform a process of requesting locking of the host is recorded. 14. A process for requesting one of a plurality of system coupling devices to lock one of a plurality of inter-host shared resources shared by a plurality of hosts is performed on each of the plurality of hosts. Locks the inter-host shared resource requested by any of the hosts, holds lock information indicating that the inter-host shared resource has been locked, and transmits the lock information and the identifier of the own system coupling device to the request source. A process of notifying the plurality of system coupling devices is performed by each of the plurality of system coupling devices, and a process of storing lock information notified from the system coupling device is performed by each of the plurality of hosts. When any one of the hosts is recognized as having failed, a process of notifying the failure to another host is performed by any of the plurality of hosts. Requesting to lock the inter-host shared resource to one of the other normal system coupling devices based on only the lock information when holding the lock information including the identifier of the failed system coupling device. Recording a program that causes each of the plurality of hosts to perform a process to be performed. 15. When a lock information including an identifier of a failed system coupling device is held, a plurality of normal system coupling devices added with a spare system coupling device based on only the lock information. 15. The recording medium according to claim 14, wherein a program for causing each of the plurality of hosts to perform a process of requesting one of the system coupling devices to lock the inter-host shared resource is recorded. 16. The recording method according to claim 11, wherein the program recorded on the recording medium according to claim 11 is divided into a plurality of portions, and the plurality of portions are respectively recorded on a plurality of recording media. Media group.