Alarm display system and method of cluster storage system

Technical Field

This invention involves a kind of applied to the alarm display system of the computer system and method, in particular relates to a storage system for the cluster (Cluster) sharing equipment alarm processing of a abnormal event by the cluster storage system alarm display system and method.

Background Art

At present, in a personal computer (PC   Cluster) trooped in storage system alarm system generally is set to some key soft, hardware object and related event is monitored, when the monitored object when the abnormal state of the alarm system will be the way of adopting the specific abnormal condition presented to the user it knows, for example, can be made of the page display, transmitting E-mail, SNMP (simple network management protocol, to define a network in the node (node) the problem of management agreement TCP/IP) prompt the user, can be realized. Cluster as a whole, comprising a shared equipment (such as magnetic disc) and independent equipment (e.g., CPU, memory, etc. (Memory)), alarm module of each node is the same, when detecting the abnormal equipment to practice in the prior art is that each node will separate processing abnormal event, the shared equipment also adopt the same way, such that when the result of the detection of the different nodes in the shared device when the abnormal events will take different processing modes, the most obvious is the problem of the page display may occur with a device to a different abnormal display. This is undoubtedly a is not accurate, the alarm processing mode is not appropriate, will also be at the same time the alarm information that the user is very disconcerting.

In particular the increase in the cluster nodes, system monitoring software and hardware object an increasing number of cases, this kind of the prior art there are undoubtedly an alarm processing mode the defect.

Content of the invention

The invention solves the technical problem of providing a cluster storage system alarm display system and method of, using reasonable single abnormal event processing mode, the computer cluster storage system by sharing equipment of all abnormal event according to the extent of the harm to the system level partition and analysis and treatment, and then realize the abnormal events the shared device of reasonable to users, accurate alarm prompt.

In order to achieve the above-mentioned purpose, the invention offers a cluster of the alarm display system of the storage system, comprising: an alarm information acquisition and storage module, and to the interrupt mode by polling each node shared device for monitoring the detection of the abnormal event, and the abnormal event of the alarm information of the alarm information is stored to a database, at the same time, the abnormal events in the database stored in the alarm event; a node load analysis module, is used for the abnormal events to the use of the shared device on the load information of each node sequencing operation, load the nodes in the node with the lowest node and node of the node of the highest load, and designated by the node load the minimum node of the abnormal events to the shared device for an alarm analysis processing; an alarm information analysis processing module, is used for the alarm event database to each node stored in the abnormal event of the detected alarm information and stored in the database of the abnormal events the alarm information corresponding to the analysis processing, according to the alarm information analysis results and determining abnormal event alarm processing priority, then according to this alarm processing priority of the node load analysis module the obtained node load the highest node, the shared device and occurrence of abnormal events most affected by abnormal event provided to the node information of the node, and the alarm to the node of the user; and an alarm module, according to the different monitoring object and warning level of the warning mode of the selective calling different alarm to the user.

Moreover, in order to achieve the above-mentioned purpose, the present invention provides a cluster storage system alarm display method, comprising the following steps: monitoring an interrupt mode and through polling of each node shared device detected by the abnormal event, and acquires load information of each of the plurality of nodes; and the obtained abnormal event of abnormal event alarm information to be stored; an abnormality has occurred in the event of the use of the shared device sequencing of a node's load operation, load the nodes in the node with the lowest node and node of the node of the highest load, and designated by the node load the minimum node of the abnormal events to the shared device for an alarm analysis processing; all the nodes on the stored detected the abnormal event of the abnormal events and corresponding to the alarm information of the analysis processing, according to the alarm information analysis results and determining abnormal event alarm processing priority; and according to the priority of the alarm processing node of the node of the highest load, the occurrence of abnormal shared device and most affected by abnormal event provided to the node information of the node, the node and the alarm to the user.

To summarize the above, the advantage of this invention lies in:

The present invention provides a cluster storage system alarm display system and method of, can be according to the system that may exist in the apparatus a plurality of impact on the effectiveness of the integrated consideration to unusual and shared equipment alarm, thereby realizing to remind the user reasonable, accurate processing of the shared device cluster storage system the beneficial effect of the abnormal events. Cluster storage system of the present invention the warning display system and method, through the use of different nodes shared equipment the abnormal events detected by the alarm prioritization of the nodes and the node of the analysis of the load, in other words the load is low the node of the cluster storage system for the analysis of all the shared device and the abnormal events most affected by abnormal event node, each node is then supplied to, the alarm to the node of the user, thereby effectively avoiding the use of shared devices on their own different nodes, in the same period of the different shared device the problem of abnormal alarm processing. Furthermore, cluster storage system of the present invention the warning display system and method, through the use of node the lower load alarm abnormal event of node processing, more can be balanced in the cluster storage system pressure of the equipment.

The following combination of the embodiment of Figures and the detailed description of the invention, but not as a limitation of this invention.

Description of drawings

Figure 1 is of the present invention the warning display cluster storage system block diagram of the system; and

Figure 2 a cluster storage system of this invention the alarm display method flow chart.

Wherein, the Figure mark:

10: alarm information acquisition and storage module

20: alarm information database

30: alarm event database

40: node load analysis module

50: alarm information analysis processing module

60: alarm module

Mode of execution

The following, some of the context of the present invention with the preferred embodiment described in detail.

Please refer to the fig. 1, expressed in the picture a of the present invention the warning display cluster storage system block diagram of the system, and the alarm display system is used for the cluster (Cluster) storage system the abnormal events sharing apparatus for alarm processing. As shown, the invention is a cluster of the alarm display system of the storage system, comprising:

Alarm information acquisition and storage module 10, and to the interrupt mode by polling each node shared device for monitoring the detection of the abnormal event, and the abnormal event of the alarm information of the alarm information to the database 20 in, at the same time, the abnormal event is stored in the alarm event database 30 in; node load analysis module 40, in the abnormal event of the use of the shared device on the load information of each node sequencing operation, load the nodes in the node with the lowest node and node of the node of the highest load, and designated by the node load the minimum node of the abnormal events to the shared device for an alarm analysis processing, the users in the cluster storage system using pressure different the node load of each node will also be different, leading to each of the plurality of nodes there is a great difference between the flow rate of the data, when the shared device is abnormal, wherein the node load the maximum, the maximum data flow (i.e. node load the highest) by the node of the impact of the abnormal events will be the greatest, the factors which can be used for the above-mentioned two points of each node calculated load information, in order to obtain the and uses the shared device abnormal event of the most affected node; alarm information analysis processing module 50, is used for the alarm events database 30 stored in each node of the abnormal events detected and the alarm information database 20 of abnormal event is stored in the alarm information corresponding to the specific analysis of the processing, according to the alarm information analysis results and determining abnormal event alarm processing priority, then according to this alarm processing priority of the node load analysis module 40 the obtained node load the highest node, the shared device and occurrence of abnormal events most affected by abnormal event provided to the node information of the node, and the alarm to the node of the user; and

An alarm module 60, according to the different monitoring object and warning level of the warning mode of the selective calling different alarm to the user, wherein the different monitoring object refer to different and the abnormal events most affected by abnormal events the node, according to the user specified, different monitoring object and alarm level to take different alarm action. Different monitoring object and the alarm grade and serious, and the like according to the category of the degree or importance levels, corresponding to the different alarm modes, may, for example, the page display, light emitting diode (LED) warning, buzzing warning, sending SNMP   Trap (SNMP, that in English:   Management Protocol   Network Simple, that the: simple network management protocol, is a series of stack and norms, from which collected in the devices on the network the method of network management information. SNMP of the equipment to the network management station to report problems and error provides a method. The managed device may at any time the agent program to the network management station to report error conditions, such as prefabricated predetermined threshold transboundary degree, and so on. Agent program does not need to wait for the management work station in order to obtain these error conditions when a polling implementation report. These error conditions is known as SNMP   Trap) prompt, prompt (E-mail) sending e-mail, such as log record alarm to the user. User for different type or grade degree of alarm selection designated the corresponding alarm mode, the system according to the different monitoring object and alarm level to select a corresponding alarm mode. The selection of the alarm mode will also be in operation in the system, according to the historical data of the item for dynamically adjusting the, so as to obtain the most to satisfy the user expectations and the warning mode of the most practical.

Wherein, the above-mentioned abnormal event of abnormal event alarm information includes the abnormal event of the grade of the historical frequency of information, this level can be abnormal event of the system under abnormal events degree of influence of the sharing of the apparatus is divided into low, medium, high three levels, wherein:

Shared equipment may affect the normal operation of the abnormal event of low level may be set to, for example, insufficient space of the equipment and other abnormal event;

Shared software may lead to equipment damage or unable to use the abnormal event is set to middle level, such as a redundant array of inexpensive disk (Redundant   Arrays   of   Disks   Inexpensive, abbrebyted RAID) subject may lead to damage of the equipment data read/write errors thereby affecting the integrity of the data, at the apparatus could not be used at will cause; and

Hard of sharing may lead to equipment damage or unable to use the abnormal event is set to the high-grade, such abnormal event requiring immediate processing, such as automatically removed from the system in a timely manner in the damaged equipment, or reminding the user of the manual pull out timely replacement of damaged intact equipment, in order to not affect the data of the user.

And the above-mentioned abnormal event of the historical frequency is a pre-set period of time (for example time Δ t) the number of times of the abnormal events occur, this can be a time period preset according to the need is preferably 20 seconds.

Now please refer to the fig. 2, the Figure of the invention of a cluster storage system alarm display method flowchart, and the alarm display method is used for the cluster (Cluster) storage system the abnormal events sharing apparatus for alarm processing. As shown, a of this invention a cluster storage system alarm display method, comprising the following steps:

Way monitoring and interrupt through polling of each node shared device detected by the abnormal event, and acquires load information of each node (step 100);

The obtained abnormal event and abnormal event of the alarm information to be stored (step 200);

In the abnormal event of the use of the shared device sequencing of a node's load operation, load the nodes in the node with the lowest node and node of the node of the highest load, and designated by the node load the minimum node of the abnormal events to the shared device for an alarm analysis processing (step 300), the users in the cluster storage system using pressure different the node load of each node will also be different, leading to each of the plurality of nodes there is a great difference between the flow rate of the data, when the shared device is abnormal, wherein the node load the maximum, the maximum data flow (i.e. node load the highest) by the node of the impact of the abnormal events will be the greatest, the factors which can be used for the above-mentioned two points of each node calculated load information, in order to obtain the and uses the shared device abnormal event of the most affected node;

Each node on the stored detected the abnormal event of the abnormal events and corresponding to the alarm information of the analysis processing, according to the alarm information analysis results and determining abnormal event alarm processing priority (step 400); and

According to the priority of the alarm processing node of the node of the highest load, the occurrence of abnormal shared device and most affected by abnormal event provided to the node information of the node, the node and the alarm to the user (step 500).

Furthermore, the above present invention of a cluster storage system in the warning display method can also further comprise according to different monitoring object and different selective calling alarm grade to the user warning mode of the step of the alarm (not shown in the Figure), wherein, different monitoring object refer to different and the abnormal events most affected by abnormal events the node, according to the user specified, different monitoring object and alarm level to take different alarm action. Different monitoring object and the alarm grade and serious, and the like according to the category of the degree or importance levels, corresponding to the different alarm modes, may, for example, the page display, light emitting diode (LED) warning, buzzing warning, sending SNMP   Trap (SNMP, that in English:   Protocol   Network Management Simple, that the: simple network management protocol, is a series of stack and norms, from which collected in the devices on the network the method of network management information. SNMP of the equipment to the network management station to report problems and error provides a method. The managed device may at any time the agent program to the network management station to report error conditions, such as prefabricated predetermined threshold transboundary degree, and so on. Agent program does not need to wait for the management work station in order to obtain these error conditions when a polling implementation report. These error conditions is known as SNMP   Trap) prompt, prompt (E-mail) sending e-mail, such as log record alarm to the user. User for different type or grade level of alarm may be selectively designated after the corresponding alarm mode, can be according to the different monitoring object and alarm level to select a corresponding alarm mode. This alarm mode selection can also be based on the historical data of the item for dynamically adjusting the, so as to obtain the most to satisfy the user expectations and the warning mode of the most practical.

Wherein, the above-mentioned abnormal event of abnormal event alarm information includes the abnormal event of the grade of the historical frequency of information, this level can be abnormal event of the system under abnormal events degree of influence of the sharing of the apparatus is divided into low, medium, high three levels, wherein:

Shared equipment may affect the normal operation of the abnormal event of low level may be set to, for example, insufficient space of the equipment and other abnormal event;

Shared software may lead to equipment damage or unable to use the abnormal event is set to middle level, such as a redundant array of inexpensive disk (Redundant   Arrays   of   Disks   Inexpensive, abbrebyted RAID) subject may lead to damage of the equipment data read/write errors thereby affecting the integrity of the data, at the apparatus could not be used at will cause; and

Hard of sharing may lead to equipment damage or unable to use the abnormal event is set to the high-grade, such abnormal event requiring immediate processing, such as automatically removed from the system in a timely manner in the damaged equipment, or reminding the user of the manual pull out timely replacement of damaged intact equipment, in order to not affect the data of the user.

And the above-mentioned abnormal event of the historical frequency is a pre-set period of time (for example time Δ t) the number of times of the abnormal events occur, this can be a time period preset according to the need is preferably 20 seconds.

Next, the binding table 1 and table 2 of the present invention some of the detail in the technical proposal described technical features:

Centralized when group a node detects that the shared device when the abnormality exists, the abnormal event is found node will at the same time using the shared equipment to the node (such as the table 2 is shown in the shared equipment Device1 node Node1, Node2   [...] ...) and send the alarm information of the node load of the node information, wherein the sending of the alarm information, as shown in table 1 shown in the, comprising abnormal event (Error1, Error2   [...] ...) with the grade of the historical frequency of information. Using the shared equipment through the analysis of the node of the decision operation of the load the lowest node to process the alarm event, the load minimum node will according to the height of the level of the abnormal event, and consider this abnormal event of an occurrence frequency history to determine how to deal with the abnormal event, especially when at the same time the different detected abnormal events, will also according to the above-mentioned alarm information to determine which priority should be given to the abnormal event. At the same time, as shown in table 2 is shown in, the load minimum node will also be according to different final weigh to the abnormal events by which the impact of the node will be the maximum (when the shared device is abnormal at that load the maximum, the flow rate of the data of the largest node is usually the impact of maximum), the most affected by abnormal events the reasonable node of abnormal information is presented to the user, so that the users can be according to the actual situation of the system to carry out corresponding adjustments and processing, for example, can be the node most affected by the load situation of the adjusted properly.

Table 1:

Table 2:

Through two below the embodiment of the technical proposal of this invention to describe the specific implementation of:

Implementation of the example one:

In t1 Node1 time cluster Device1 shared device node detects that a low-grade abnormal event E1 and E2, node Node1 Device1 immediately to the use of equipment of the detection is sent to the node of the abnormal event information and load information of the node of the node, at the same time the remaining node is also detection of the same abnormal event E1 and E2, and the unusual event information and load information of the node is sent to the shared equipment Device1 each node of the cluster. The minimum load by the node is the node of the analysis processing NodeL the abnormal information, taking into account the node Node5 the highest of the load, and abnormal event E2 is higher than the frequency of the history of abnormal event E1, therefore node the first priority NodeL abnormal event E2 carries out alarm treatment, and then the abnormal event E1 carries out alarm treatment, at the same time will be in the page display: node Node5 the most affected by an abnormal event, in order to remind the user.

Implementation of the example two:

In t2 Node1 time cluster Device1 shared device node detects that a low-grade abnormal event E1 and the high-grade abnormal event E2, node Node1 Device1 immediately to the use of equipment of the detection is sent to the node of the abnormal event information and load information of the node of the node, at the same time the remaining node is also detection of the same abnormal event E1 and E2, and the unusual event information and load information of the node is sent to the shared equipment Device1 each node of the cluster. The minimum load by the node is the node of the analysis processing NodeL the abnormal information, taking into account the node Node5 the highest of the load, thus node the first priority NodeL abnormal event E2 carries out alarm treatment, and then the abnormal event E1 carries out alarm treatment, at the same time will be in the page display: node Node5 the most affected by an abnormal event, in order to remind the user.

Of course, the invention can also have other various embodiments, without departing from the invention spirit and its real circumstances, the technical personnel familiar with the field of the invention according to the various corresponding change and deformation, but these relative changes and deformation should be to the right of the invention the attached scope of protection requested.