GROUP ALERT IN SERVER SYSTEMS

Abstract

A server system having functionality of group alerting is disclosed. Said server system comprises: a plurality of server computers having alert notification capabilities, the plurality of server computers being divided into at least one group; and a management console node managing and monitoring the plurality of server computers; wherein the alert notification is issued by a group of the at least one group of the plurality of server computers when a health problem of a server computer in said group of the at least one group of the plurality of server computers occurs.

Description

TECHNICAL FIELD

The present application relates to event alerting in a server system. In particular, the present application relates to a server system having functionality of group alerting and a method for group alerting in a server system.

BACKGROUND

In a server system, system administrators usually have to manage and monitor a lot of and a variety of server computers. System administrators must take actions when any of the client server computers meets abnormal situations in order to sustain the whole server system operating normally and healthily. Different functions, areas and locations of the client server computers may encounter different types of problems. The system administrators must handle a variety of types of problems from a lot of server computers within time limits, which is a complex and urgent task.

The Intelligent Platform Management Interface (IPMI) architecture is usually applied to a server system for alerting events of any of the client server computers. The IPMI is a standardized computer system interface used by system administrators for out-of-band management of computer systems and monitoring of their operation. It is a way to manage a computer that may be powered off or be otherwise unresponsive by using a network connection to the hardware rather than to an operating system or login shell.

In the conventional IPMI supported server system platform, an alert notification is issued by individual computer systems. This is very useful for a small business company. However, for large scale enterprises, such as Google, Yahoo, Facebook, which have very large data centers, this technology more likely focuses on a group of servers management which servers are classified into groups by areas, locations and functions. Therefore, a method for group alerting in such a server system platform is required.

CONTENT OF THE INVENTION

The present application proposes an invention of group alerting for the server system in the large scale enterprise with very large data center, such that an administrator can monitor the large data center by predefined groups.

According to one aspect of the present application, a server system having functionality of group alerting is disclosed. Said server system comprises: a plurality of server computers having alert notification capabilities, the plurality of server computers being divided into at least one group; and a management console node managing and monitoring the plurality of server computers; wherein the alert notification is issued as a group event by a group of the at least one group of the plurality of server computers when an event of a server computer in said group of the at least one group of the plurality of server computers occurs.

According to another aspect of the present application, a method for group alerting in a server system is disclosed. Said method comprises: managing and monitoring a plurality of server computers by a management console node; dividing the plurality of server computers in the server system into at least one group, each of the plurality of server computers having alert notification capabilities; and issuing the alert notification as a group event by a group of the at least one group of the plurality of server computers when an event of a server computer in said group of the at least one group of the plurality of server computers occurs.

BRIEF DESCRIPTION OF THE DRAWINGS

Various features and advantages of the disclosed embodiments will be apparent from the detailed description which follows, taken in conjunction with the accompanying drawings, which together illustrate, by way of example, features of the disclosed embodiments.

FIG. 1 illustrates a block diagram of the groups of servers in a server system of the present invention.

FIG. 2 illustrates a block diagram of another server system of the present invention.

FIG. 3 illustrates an operation mechanism of a group event in one of at least one group of a server system of the present invention.

FIG. 4 illustrates a block diagram of a further server system of the present invention.

FIG. 5 illustrates a method for group alerting in a server system of the present invention.

DETAILED DESCRIPTION

FIG. 1 illustrates a block diagram of a server system (100) of the present invention. The server system (100) comprises a management console node (110) and a plurality of server computers communicating with each other via a network (120). The plurality of server computers is divided into at least one group, i.e., server group 1 (130), server group 2 (140), . . . server group N (150). The management console node (110) may be, but not limited to, a server computer, a personal computer, or a handheld computing device. An administrator of the server system may classify the plurality of server computers in the server system to different groups by their areas, locations, or functions.

The management console node (110) manages and monitors the operation and the health/utilization status of the plurality of server computers. When an event of a server computer in the plurality of server computers occurs, an alert notification will not be issued by said server computer having the event. On the contrary, the alert notification will be issued by a group of the at least one group of the plurality of server computers, to which said server computer belongs. Hence, the issued alert notification reports a group event of a specific group in the server system rather than an individual event of a specific server computer in the server system. Namely, if a server computer (131) in server group 1 (130) has an event, the alert notification will be issued by server group 1 (130); if a server computer (141) in server group 2 (140) has an event, the alert notification will be issued by server group 2 (140); and if a server computer in server group N (150) has an event, the alert notification will be issued by server group N (150), instead of said server computer itself. The alert notification is sent to the management console node (110), such that the management console node (110) can locate the group including said server computer having the event.

Each of the plurality of server computers in the server system (100) may have a built-in IPMI supported Baseboard Management Controller (BMC) and does not need any extra hardware to support the group alerting of the server system of the present invention.

FIG. 2 illustrates a block diagram of another server system (200) of the present invention. Similar to the server system (100) illustrated in FIG. 1, the server system (200) comprises a management console node (210) and a plurality of server computers communicating each other via a network (220). The plurality of server computers is divided into at least one group. i.e., server group 1 (230), server group 2 (240), . . . server group N (250). However, based on different criteria of dividing the plurality of server computers into different groups, the same server computer (234) may belong to different groups. For example, the administrator divides portions of the plurality of server computers into server group 1 (230) according to their location A and divides other portions of the plurality of server computers into server group 2 (240) according to their function B. If the server computer (234) is located at the location A and has the function B, the server computer (234) may belong to both of server group 1 (230) and server group 2 (240). The server computer (234) may issue its events through server group 1 (230) or server group 2 (240) depending on the group alert policies. For example, the server computer (234) may issue an event of CPU utilization over 90%, which belongs to a functional event, through server group 2 (240) and issue an event of unexpected shutdown, which belongs to a health event at a specific location, through server group 1 (230).

FIG. 3 illustrates an operation mechanism of a group event in one (300) of at least one group of a server system of the present invention. Each group in the server system comprises a domain node (310) and at least one client server computer (320, 330, 340). The domain node (310) may be, but not limited to, a server computer, a personal computer, or a handheld computing device.

Based on different criteria of dividing the plurality of server computers into different groups, a domain node in one group of the server system may also be a client server computer in another group of the server system. For example, as illustrated in FIG. 2, the server computer (234) may act as the domain node of server group 2 (240) and issue the group events of server group 2 (240) to the management console node (210). The server computer (234) may also be a client server computer of server group 1 (230) and send its own events to the domain node (not shown in FIG. 2) of server group 1 (230) rather than the management console node (210). Namely, the server computer (234) is simultaneously the domain node of server group 2 (240) and one of the client server computers of server group 1 (230).

When an event of a client server computer (320) occurs, the client server computer (320) cannot send a normal alert to the management console node. Instead, the client server computer (320) sends a private event (350), which can only be received by the domain node (310). After receiving the private event (350) issued by the client server computer (320), the domain node (310) needs to decide that the current event (350) has reached predefined threshold values and needs to be sent to the management console node as a group event (360) according to a predefined policy. The issued group event (360) may comprise, for example, a group health alert, a group utilization alert, a group hardware failure alert and other group alerts indicating any event that impacts the normal operations of the server computers. Examples of group alerting policies are listed in the following table. However, the illustrated examples of group alerting policies do not limit the claimed scope of the present invention. Any modification of the alerting policies may easily be conceived by a person skilled in the art based on the actual requirements when managing and monitoring a server system.

Group Alert (Health/Utilization/Hardware Failure)

TABLE 1
Examples of Group Alerting Policies
			GTH
			Alert
		STH	Percentage	When to
		Single system	in group	trigger
Type	Event	threshold	threshold	event
Group Health	Over CPU	CPU temperature	10% (In a	STH =
Critical Alert	temperature	over limit (100° C.)	group of 100	true
			servers, that	and
			is 10)	GTH =
				true
Group Health	System	System	3% (In a	STH =
Critical Alert	unexpected	unexpected	group of 100	true
	shutdown	shutdown	servers, that	and
			is 3)	GTH =
				true
Group Health	Over CPU	CPU temperature	1% (In a	STH =
Warning Alert	temperature	over limit (100° C.)	group of 100	true
			servers, that	and
			is 1)	GTH =
				true
Group Utilization	Over CPU	CPU average	80% (In a	STH =
Alert	utilization	utilization over	group of 100	true
		90% in last one	servers, that	and
		minute	is 80)	GTH =
				true
Group Utilization	Over network	average network	20% (In a	STH =
Alert	throughput	throughput over 80 Mb/s	group of 100	true
	utilization	in last five	servers, that	and
		minutes	is 20)	GTH =
				true
Group Hardware	Disk failure	Disk failure	3% (In a	STH =
Failure Warning		(Recoverable)	group of 100	true
Alert			servers, that	and
			is 3)	GTH =
				true
Group Hardware	Disk failure	Disk failure	1% (In a	STH =
Failure Critical		(Unrecoverable)	group of 100	true
Alert			servers, that	and
			is 3)	GTH =
				true
Group Hardware	Power failure	Redundant Power	3% (In a	STH =
Failure Warning		supply failure	group of 100	true
Alert			servers, that	and
			is 3)	GTH =
				true
Group Hardware	Power failure	System shutdown	1% (In a	STH =
Failure Critical		due to power	group of 100	true
Alert		supply failure	servers, that	and
			is 3)	GTH =
				true
Group Hardware	Unexpected	Unexpected	1% (In a	STH =
Failure Critical	shutdown	system shutdown	group of 100	true
Alert	failure		servers, that	and
			is 3)	GTH =
				true

The administrator may set different predefined group alerting policies for each domain node in each server group of the plurality of server computers (as listed in Table 1). For example, the administrator may set an over temperature threshold value of 100° C. for the CPU of each client server computer in server group 1 (130). The client server computer will send a private event (350) to a domain node of server group 1 (130) when it reach a CPU temperature over 100° C. The domain node of server group 1 (130) will accumulate and count total private events sent by the client server computers. The administrator may further set a group over temperature event of 10% of the client server computers alerting the event of over CPU temperature for the domain node. The domain node will send a group event (360) to the management console node (110) when it accumulates that over 10% of the client server computers have issued the private events of over CPU temperature. After the management console node (110) receiving the group event from the domain node, the administrator may immediately take actions to prevent the client server computers from being crushed. For example, the administrator may immediately cool down the room temperature for server group 1 (130), share work loads of server group 1 (130) to other server groups, or take any other actions that may reduce the temperature of the CPU of the client server computers in server group 1 (130). For server group 2 (140), due to lower temperature endurance of the CPU of the server computers, the administrator may set a lower over temperature threshold value of 80° C. (for example) for the CPU of each client server computer and set a group over temperature threshold value of 5% of the client server computers alerting the event of over CPU temperature for the domain node, such that the administrator may be informed with a lower threshold for the issue of CPU temperature in server group 2 (140).

The other example is that the administrator may set an average network throughput threshold value of over 80 Mb/s in last five minutes for each client server computer and set an over network throughput utilization group event of over 20% of the client server computers reaching the threshold value of average 80 Mb/s throughput in the last five minutes. The administrator may monitor whether each of the groups of the server system incurs abnormally large data transmissions at the network, which may be caused by a virus attack or by a simultaneous file downloading in a specific area.

FIG. 4 illustrates a block diagram of a further server system (400) of the present invention. Similar to the server system (100) illustrated in FIG. 1, the server system (400) comprises a management console node (410) and a plurality of server computers communicating each other via a network (420). The plurality of server computers is divided into at least one group, i.e., server group 1 (430), server group 2 (440), . . . server group N (450). Due to the scale of the server system (400) being large and complex, the administrator may further classify the server computers in server group 1 (430) or in server group 2 (440) to different subgroups by their more detailed areas, locations, or functions. Each subgroup in each group of the server system also comprises its own domain node and at least one client server computer, such that the management console node (410) may focus on fewer groups in order to more efficiently manage and monitor the server system.

According to the illustrated server system of the present invention, a method (500) for group alerting in a server system is also illustrated in FIG. 5. Said method (500) for group alerting in the server system may manage and monitor a plurality of server computers by a management console node (510), divide the plurality of server computers in the server system into at least one group, in which each of the plurality of server computers has alert notification capabilities (520), and issues the alert notification as a group event by a group of the at least one group of the plurality of server computers when an event of a server computer in said group of the at least one group of the plurality of server computers occurs (530).

The claimed server system and method of the present application have many advantages over the conventional server system platform. For example, the group alerting allows an administrator to monitor the server system by group instead of individual server computers, such that identifying and locating group problems in a very large data center becomes easier than locating each of the individual server computers. Only deployment of a new domain node in each of the groups is required without adding any new hardware and changing any hardware configurations on each existing server computer. The group alerting for the server system of the present application may be implemented merely by a firmware update without any extra efforts.

The aforesaid detailed descriptions illustrate the preferred embodiments of the present application. However, the scope of the claimed invention should not be limited by any of the above-described exemplary embodiments, but should be defined only in accordance with the following claims and their equivalents.

Claims

1. A server system having functionality of group alerting, comprising: a plurality of server computers having alert notification capabilities, the plurality of server computers being divided into at least one group; anda management console node managing and monitoring the plurality of server computers;wherein the alert notification is issued as a group event by a group of the at least one group of the plurality of server computers when an event of a server computer in said group of the at least one group of the plurality of server computers occurs.

2. The server system of claim 1, wherein the group event is sent to the management console node by said group of the at least one group of the plurality of server computers, such that the management console node can locate said group issuing the group event.

3. The server system of claim 1, wherein each of the at least one group of the plurality of server computers comprises a domain node and at least one client server computer.

4. The server system of claim 3, wherein the domain node is also one of the at least one client server computer.

5. The server system of claim 3, wherein one of the at least one client server computer sends a private event to the domain node when an event of said one of the at least one client server computer occurs, and then the domain node decides whether to issue the group event to the management console node according to a predefined policy.

6. The server system of claim 5, wherein the group event comprises a group health alert, a group utilization alert and a group hardware failure alert in response to the predefined policy.

7. The server system of claim 1, wherein each of the plurality of server computers has a built-in Intelligent Platform Management Interface (IPMI) supported Baseboard Management Controller (BMC).

8. The server system of claim 1, wherein the plurality of server computers is divided into the at least one group by areas, locations, or functions.

9. The server system of claim 1, wherein the plurality of server computers is divided into the at least one group by an administrator.

10. The server system of claim 1, wherein one or more groups of the at least one group of the plurality of server computers may further be divided into at least one subgroup, each of the at least one subgroup comprises its own domain node and at least one client server computer.

11. A method for group alerting in a server system, comprising: managing and monitoring a plurality of server computers by a management console node;dividing the plurality of server computers in the server system into at least one group, each of the plurality of server computers having alert notification capabilities; andissuing the alert notification as a group event by a group of the at least one group of the plurality of server computers when an event of a server computer in said group of the at least one group of the plurality of server computers occurs.

12. The method of claim 11, wherein the group event is sent to the management console node by said group of the at least one group of the plurality of server computers, such that the management console node can locate said group issuing the group event.

13. The method of claim 11, wherein each of the at least one group of the plurality of server computers comprises a domain node and at least one client server computer.

14. The method of claim 13, wherein the domain node is also one of the at least one client server computer.

15. The method of claim 13, wherein one of the at least one client server computer sends a private event to the domain node when an event of said one of the at least one client server computer occurs, and then the domain node decides whether to issue the group event to the management console node according to a predefined policy.

16. The method of claim 15, wherein the group event comprises a group health alert, a group utilization alert and a group hardware failure alert in response to the predefined policy.

17. The method of claim 11, wherein each of the plurality of server computers has a built-in Intelligent Platform Management Interface (IPMI) supported Baseboard Management Controller (BMC).

18. The method of claim 11, wherein the plurality of server computers is divided into the at least one group by areas, locations, or functions.

19. The method of claim 11, wherein the plurality of server computers is divided into the at least one group by an administrator.

20. The method of claim 11, wherein one or more groups of the at least one group of the plurality of server computers may further be divided into at least one subgroup, each of the at least one subgroup comprises its own domain node and at least one client server computer.