JPH04262428A - Hot standby system - Google Patents

Abstract

PURPOSE:To upgrade the reliability of switching between operating host and waiting host and the switching speed in a hot standby system. CONSTITUTION:An operating type hot standby control job 120, when it detects a failure of the application that an application driving program 130 is executing based on response time or the like, outputs a host switching request 140, whereby a hot standby process 110 initiates a local host forced down means 150 to down an operating host 100 forsibly. When the down processing is completed, the local host forced down means 150 outputs a down notice 103. When receiving the down notice, a hot standby process 210 outputs a switching request 240, and in response to this request, a waiting type hot standby control job 220 uses an application transfer means 223 to take in resource and also to switch an application driving program 230 to the operating type to restart application driving.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a hot standby system that quickly takes over operations from an active system to a standby system when an abnormal situation occurs, in order to achieve high reliability in an information processing system.

0002

[Prior Art] Conventionally, in a hot standby system consisting of an active host in operation and a standby host that stands by to take over the business when the active host goes down, communication lines and job The down state of an active host is indirectly detected from events such as no response to communications performed at regular intervals using communication means consisting of inter-communications and the like. If the active host is detected to be down, the active host is forcibly brought down from the outside and then switched between the active host and the standby host. Furthermore, the business operation program, which is an application program, uses a unique method to take over the business from the active host to the standby host, taking into account the resource usage pattern.

0003

[Problems to be Solved by the Invention] Firstly, in the past, failure of an active host was indirectly detected from events such as no response to communications performed at regular intervals using communication means. However, since the communication means were not originally designed for such purposes and are used at the same level as other services, there are problems in terms of reliability and speed, and communication There has been a problem in that it is difficult to distinguish between a failure in the means and an abnormal operation of the operating host.

Second, conventionally, when switching between an active host and a standby host, the active host was forcibly brought down from the outside and then switched between the active host and standby host. There was a problem in that it was difficult to perform high-speed switching.

[0005] Thirdly, in the past, switching between the active host and the standby host was performed by the application program while taking into consideration the resource usage pattern, which resulted in the problem that many steps were required to develop the application program. .

An object of the present invention is to improve the reliability and switching speed when switching between an active host and a standby host, and to facilitate the development of application programs.

[0007]

[Means for Solving the Problems] In order to achieve the above-mentioned object, the present invention provides a system for connecting an active host and a backup standby host that is ready to start service while the active host is in operation. In the configured hot standby system, the active host has a first business operation program activated as an active type under control, and outputs a host switching request by detecting an abnormality in its own host. An active hot standby control job, a host forced down means for forcibly bringing down the host and outputting a down notification to the standby host, and a first hot standby process controlling the host forced down means. The standby host has a second business operation program activated as a standby type under control, and detects an abnormality in the active host through inter-job communication with the active hot standby control job. Standby type that outputs a confirmation request for other systems, takes in the resources necessary for business operation in response to the switching request, and switches the second business operation program that is started as the standby type to the active type to restart the operation job. In response to the hot standby control job and the other system confirmation request, it is determined whether the operating host is normal or not, and if it is determined to be abnormal, a forced down request is output, and the switching is performed in response to the down notification. a second hot standby process that outputs a request, and the first hot standby process responds to a host switching request output from the active hot standby control job and a forced down request output from the second hot standby process. In response, the self-host forced down means is activated to bring down the active host.

[0008]

[Operation] When the active hot standby control job detects that an abnormality has occurred in the active host based on the response time of the first business operation program under control, it outputs a host switching request. When a host switching request is output from the active hot standby control job, the first hot standby process on the active host side activates its own host forced down means. As a result, the own host forced down means brings down the active host and outputs a down notification.

The second hot standby process on the standby host outputs a switching request in response to the down notification. As a result, the standby type hot standby control job takes in the resources necessary for the business operation, switches the second business operation program started as a standby type to the active type, and restarts the business operation.

Furthermore, when the standby hot standby control job on the standby host side detects an abnormality in the active host through inter-job communication with the active hot standby control job, it outputs a request for confirmation of other systems. In response to this other system confirmation request, the second hot standby process on the standby host side determines whether the active host is normal or not, and if it determines that it is abnormal, outputs a forced down request.

The first hot standby process on the active host activates its own host forced down means in response to the forced down request. As a result, the own host forced down means forcibly brings down the active host, and then outputs a down notification.

[0012] The second hot standby process on the standby host outputs a switching request in response to the down notification;
As a result, the standby type hot standby control job takes in the resources necessary for the business operation, switches the second business operation program that has been activated as a standby type to the active type, and restarts the business operation.

[0013]

Embodiments Next, embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 is a block diagram of an embodiment of the present invention, in which an active host 100, a standby host 200, and a down notification 1
03. An inter-host communication means 101 for transmitting and receiving forced down requests 104, etc.; and an inter-job communication means 102 for communicating between a job running on the active host 100 and a job running on the standby host 200. It is composed of.

The active host 100 includes a hot standby process 110 including a hot standby control job monitoring means 111, an active hot standby control job 120 including a hot standby establishing means 121 and a hot standby canceling means 122, and an active hot standby control job. Business operation program 13 that operates under the control of 120
0, self-host forced down means 150, and control table 160. Note that the active hot standby control job 120 is configured by an operating system.

The standby host 200 includes a hot standby process 210 including a hot standby control job monitoring means 211, another system confirmation means 212, and another system host forced down activation means 213, a hot standby establishment means 221,
Standby hot standby control job 220 including hot standby release means 222 and business handover means 223
, a business operation program 230 that operates under the control of the standby hot standby control job 220 , and a control table 260 . Note that the standby hot standby control job 220 is configured by the operating system.

The hot standby process 110 in the active host 100 is the active hot standby control job 12.
It has a function of activating the own host forced down means 150 in response to a host switching request 140 from 0 and a forced down request 104, and also has a hot standby control job monitoring means 111 that manages the contents of the control table 160. .

Active hot standby control job 120
is a function that detects the occurrence of an abnormality in the active business such as online work being performed by the business operation program 130, and requests host switching 14 by detecting the abnormality.
0 to the hot standby process 110, and also communicates with the hot standby establishing means 221 in the standby hot standby control job 220 using the inter-job communication means 102 to determine whether to enter the hot standby established state. The active hot standby control job 120 is brought into a hot standby release state in response to a hot standby release notification sent from the hot standby release means 222 via the mutually confirming hot standby establishment means 121 and the inter-job communication means 102. and hot standby canceling means 122 for returning a response to hot standby canceling means 222.

The self-host forced down means 150 has a function of forcibly bringing down the active host 100 by being started from the hot standby process 110, and sends a down notification 103 using the inter-host communication means 101 when the down processing is completed. It has a function of sending data to the standby host 200.

In the control table 160, information indicating the status is stored by the hot standby control job monitoring means 111 when the active hot standby control job 120 becomes ready and when the business operation program 130 starts operating. When the registered active hot standby control job 120 or business operation program 130 normally ends, the hot standby control job monitoring means 111 deletes the information regarding the completed job.

The hot standby process 210 in the standby host 200 has a function of outputting a switching request 240 in response to the down notification 103, and also includes a hot standby control job monitoring means 211 that manages a control table 260, Standby control job 220
The active host 100 responds to the other system confirmation request 250 from
and another host that sends a forced down request 104 to the active host 100 using the inter-host communication means 101 when the other system check unit 212 confirms the abnormality of the active host 100. The forced down activation means 213 is also provided.

[0021] Standby hot standby control job 220
has a function to detect that the business operation program 230 has fallen into a stall state, a function to detect through inter-job communication that an abnormality has occurred in the active host 100, and a function to change the business operation program 230 that is started as a standby type to an active type. The hot standby job has a function of resuming business operation by using the hot standby job, and also communicates with the hot standby establishing means 121 in the active hot standby job 120 via the inter-job communication means 102 to confirm whether to enter the hot standby state. When the standby establishment means 221 and the business operation program 230 are in a stalled state, the active hot standby control job 1 is executed using the inter-job communication means 102.
A hot standby cancellation notification is sent to the hot standby cancellation unit 122 in the hot standby cancellation unit 1.
A hot standby release means 222 that waits for a response from the hot standby control job 220 and puts the standby hot standby control job 220 into a hot standby release state, and a hot standby release means 222 that waits for a response from the hot standby control job 220 and takes in the resources necessary for business operations (the hot standby control job 220 that is used by the active host 100 for business operations). It also has a business handover means 223 that takes over disk devices to which databases, journal files, etc. are allocated) and switches networks.

In the control table 260, when the standby type hot standby control job 220 becomes ready and when the business operation program 230 becomes ready for business handover, the hot standby control job monitoring means 21
1, information indicating this is registered, and when the standby hot standby control job 220 or the business operation program 230 terminates normally, the hot standby control job monitoring means 211 deletes the information regarding the job to be terminated.

Next, the operation of this embodiment will be explained.

When the systems of the active host 100 and the standby host 200 are started, the operator starts the active hot standby control job 120 and the standby hot standby control job 220. Active hot standby control job 120, standby hot standby control job 2
When the process 20 becomes ready, it notifies the hot standby processes 110 and 210 that it has become ready.

Upon receiving this notification, the hot standby process 110, 210 uses the hot standby control job monitoring means 111, 211 to update the control table 160, 2.
60, it is registered that the active type hot standby control job 120 and the standby type hot standby control job 220 have become ready.

Active hot standby control job 120
When the operator confirms through a console or the like that the hot standby control job 220 is ready, the operator starts the business operation programs 130 and 230 as active type and standby type, respectively. As a result, the active host 100 side starts active business such as online business, and the standby host 200 side becomes ready to take over the business. After confirming this, the active hot standby control job 120 notifies the hot standby process 110 that the active business has started, and the standby hot standby control job 220 notifies the hot standby process 110 that the active work has started.
0 is notified that the preparation for business handover is complete.

Upon receiving this notification, the hot standby control job monitoring means 111 in the hot standby process 110 registers in the control table 160 that the active work has started, and the hot standby control job monitoring unit 111 in the hot standby process 210 registers the start of the active system job in the control table 160. The means 211 registers in the control table 260 that the preparation for business handover has been completed.

Active hot standby control job 120
When confirming that the active work has started, the hot standby establishment means 22 in the standby type hot standby control job 220 uses the hot standby establishment means 121.
1 through the inter-job communication means 102. Then, by confirming that the standby host 200 side is ready to take over the business, the hot standby state is established. Further, when the standby type hot standby control job 220 confirms that the job handover preparation is completed, the hot standby establishment means 221 is used to communicate with the hot standby establishment means 121 in the active type hot standby control job 120 and the inter-job communication means. Communication is performed via 102. Then, by confirming that the active host 100 side has started the active service, a hot standby state is established. In addition, in the hot standby established state, the active hot standby control job 120 and the standby hot standby control job 220 send and receive data at fixed time intervals in order to indirectly detect abnormalities on the active host 100 side. Let's do it.

[0029] Standby type hot standby control job 220
In the hot standby established state, when it is detected that the business operation program 230 started as a standby type has fallen into a stall state, the hot standby release means 22
2 is used to send a hot standby release notification to the hot standby release means 122 in the active hot standby control job 120 via the inter-job communication means 102. When the hot standby release means 122 receives this notification, the active hot standby control job 120 enters the hot standby release state and returns a response to the hot standby release means 222 via the inter-job communication means 102. When the hot standby release means 222 receives the response from the hot standby release means 122, the standby hot standby control job 220 enters the hot standby release state.

Furthermore, in the active hot standby control job 120, when the hot standby is established, an abnormality occurs in the active host 100, such as an extremely delayed response to the active business being performed by the business operation program 130. If detected based on the fact that the hot standby process 110 has
raise.

Upon receiving the host switching request 140, the hot standby process 110 activates its own host forced down means 150. As a result, the own host forced down means 150 forcibly downs the active host 100,
When the down processing is completed, a down notification 103 is sent to the standby host 200 using the inter-host communication means 101.

The down notification 103 from the own host forced down means 150 is received by the hot standby process 210, and upon receiving the down notification 103, the hot standby process 210 refers to the control table 260. When it is confirmed that the standby hot standby control job 220 is ready for business takeover, it sends a switching request 240 to the standby hot standby control job 220. As a result, the standby hot standby control job 220 uses the business handover means 223 to import resources such as disk devices to which databases, journal files, etc. used in the active business are allocated, and to switch networks. , After that, commercial operation program 2 is started as a standby type.
30 will be switched to operational type and resume commercial operation.

Switching between the active host and the standby host is also performed in the following cases. In the hot standby established state, the active hot standby control job 120 and the standby hot standby control job 220 exchange data at regular intervals via the inter-job communication means 102. When the data from the control job 120 is interrupted, the standby hot standby control job 220 issues an other system confirmation request 250 to the hot standby process 210. As a result, the hot standby process 210 activates the other system confirmation means 212. When activated, the other system confirmation means 212 performs other system confirmation using the inter-host communication means 101, and even if the response to this is not received after a certain period of time, the active host
If the hot standby process 210 does not return from 00, the hot standby process 210 determines that an abnormality such as a stall has occurred in the active host 100, and sends a forced down request 104 to the active host 100 using the other host forced down activation means 213.

The hot standby process 110 that has received this forced down request 104 forcibly brings down the active host 100 using its own host forced down means 150, and when the own host forced down means 150 completes the down process, the host The standby host 2 via the communication means 101
A down notification 103 is sent to 00. Down notification 103
After the is sent, the same operation as described above is performed.

[0035]

Effects of the Invention As explained above, the present invention is provided with an active hot standby control job for detecting abnormalities in the active host, so that abnormalities in the active host can be detected indirectly through inter-job communication. This has the effect of improving reliability and detection speed compared to conventional methods that detect .

Further, the present invention is provided with self-host forced down means for forcibly bringing down an active host in response to a host switching request from an active hot standby control job and a forced down request from a standby host. Therefore, there is an effect that switching between the active host and the standby host can be performed at high speed.

Furthermore, since the present invention provides the standby hot standby control job with a function of importing resources necessary for business operation, the number of man-hours required for developing the business operation program, which is an application program, can be greatly reduced. It has the effect of reducing

[Brief explanation of the drawing]

FIG. 1 is a block diagram of an embodiment of the invention.

[Explanation of symbols]

100... Active host 101... Inter-host communication means 102... Inter-job communication means 110, 210... Hot standby process 111, 2
11...Hot standby control job monitoring means 120...Active hot standby control job 121, 221...Hot standby establishing means 122, 222...Hot standby canceling means 130, 230...Business operation program 150...Self host forced down means 160, 260... Control table 200... Standby host 212... Other system confirmation means 213... Other host forced down activation means 220... Standby type hot standby control job 223... Business handover means

Claims (2)

[Claims] Claim 1: In a hot standby system consisting of an active host and a standby standby host that is ready to start service while the active host is in operation, the active host is , has a first business operation program started as an active type under control, and has an active hot standby control job that outputs a host switching request by detecting an abnormality in the own host, and a hot standby control job that forcibly activates the own host. a host host forced down means for causing the host to go down and outputting a down notification to the standby host; and a first hot standby process for controlling the host host forcible down means, wherein the standby host is activated as a standby type under control. has a second business operation program that is configured to operate, and outputs an other system confirmation request by detecting an abnormality in the active host through inter-job communication with the active hot standby control job, and responds to the switching request. A standby type hot standby control job that takes in the resources necessary for the business operation and switches the second business operation program started as the standby type to the active type to resume the operation job, and responds to the other system confirmation request. a second hot standby process that determines whether the operating host is normal or not, outputs a forced down request when determining that it is abnormal, and outputs the switching request in response to the down notification; The first hot standby process activates the own host forced down means in response to the host switching request outputted by the active hot standby control job and the forced down request outputted by the second hot standby process, and A hot standby system characterized by bringing down the operating host. 2. The active hot standby control job detects an abnormality in the active host based on a response time in a job being performed by the first business operation program. hot standby system.