This invention relates generally to servers, such as database servers and other types of servers, and more particularly to the configuration of such servers.
Computer servers have become a mainstay in information technology (IT) environments. Generally, a server is a computer that is shared by multiple users. Servers can include file servers, for storing the data files of multiple users, database servers, for storing databases accessible by multiple users, web servers, for storing web pages accessible by multiple users over the Internet, as well as other types of servers. Large organizations may have tens, hundreds, or even more servers in their IT environments.
For usability, performance, security, and other reasons, servers must usually be appropriately configured prior to their first use, and reconfigured as the environments in which they operate change, and as they are updated with new hardware and/or software. Network administrators are typically expected to configure and otherwise maintain servers under their responsibility. As server technology has grown increasingly complex, the number of different configuration parameters that administrators are expected to track has correspondingly increased in size and complexity.
However, this increase in the size and complexity of server configuration parameters is reaching, if not having reached already, the point where it may be nearly impossible for a single or small group of network administrators to properly track such parameters. In response, automatic server configuration utilities are becoming more commonplace. These utilities are usually in the form of wizards, which are instructional help utilities that guide the administrators through the steps necessary to properly accomplish server configuration.
However, for both hardware and software server manufacturers, the development of such utilities can be burdensome. Each wizard utility is typically specific to a given type of server, either from a hardware or a software perspective. Changes in the hardware or software of the server may thus necessitate a large programming effort to rework the utility. The end result may be that the manufacturers have to dedicate a large amount of resources to create such utilities, or decrease the number of server configuration parameters that can be set using such utilities.
For these described reasons, as well as other reasons, there is a need for the present invention.
The invention relates to the configuration of servers using profile templates. A first method of the invention includes loading a profile template for configuring one or more servers. The profile template includes abstracted server configuration parameters and server configuration rules. The parameters are for abstractly configuring a server type of which the servers are, whereas the rules specify how values of the parameters translate to direct server values for directly configuring the servers, based on attributes of the servers. A user is requested to modify preset of the parameters, and/or enter missing values of the parameters. Finally, a configuration for each server is generated as the values of the parameters, and the server configuration rules.
A second method of the invention subsequently loads such a configuration. The configuration includes values of abstracted configuration parameters for abstractly configuring a type of server, and server configuration rules. The server configuration rules of the configuration are applied to the values of the abstracted configuration parameters of the configuration, based on attributes of a server. This generates the direct server values for directly configuring the server. Thus, the server is configured based on the direct server values generated.
A system of the invention includes a configuration server and one or more servers. The configuration server stores configurations, including values of abstracted configuration parameters and server configuration rules. The rules specify how the values of the parameters translate to direct server values for direct server configuration. Each server correspondingly loads a configuration from the configuration server, and applies the rules to the values of the parameters based on its attributes, yielding direct server values that the server uses to configure itself.
An article of manufacture of the invention includes a computer-readable medium and means in the medium. The means in the medium is for loading a configuration including values of abstracted configuration parameters, and server configuration rules. The means is further for applying the server configuration rules to the values of the abstracted configuration parameters, based on server attributes, to generate direct server values. The means is finally for configuration a server based on the direct server values generated.
Other features and advantages of the invention will become apparent from the following detailed description of the presently preferred embodiment of the invention, taken in conjunction with the accompanying drawings.
The drawings referenced herein form a part of the specification. Features shown in the drawing are meant as illustrative of only some embodiments of the invention, and not of all embodiments of the invention, unless otherwise explicitly indicated, and implications to the contrary are otherwise not to be made.
Overview: Preferred Method
With respect to the part 102 of the method 100, a profile template is initially loaded (106). The profile template includes abstracted server configuration parameters, as well as server configuration rules. The abstracted server configuration parameters are for abstractly configuring a server type, of which the servers that will individually perform the part 104 of the method 100 are. These parameters are high-level parameters, which preferably do not specifically map to individual and directly set parameters of the servers, values for which are referred to as direct server values to avoid confusion. The abstracted server configuration parameters may include a number of static direct server values, though, which remain constant regardless of a server's attributes.
As an example, the abstracted server configuration parameters may include parameters regarding the resource utilization of the servers, the load balancing requirements of the servers, and so on. More specifically, they may include such parameters as percentage of server resources, percentage of online transaction processing (OLTP) to be performed by the server, and so on. The collection of abstracted server configuration parameters preferably constitute a usage model of a given server's configuration, providing a high-level view as to how the server is to perform.
The server configuration rules then specify how values of the abstracted configuration parameters translate to the direct server values for directly configuring the servers, based on the servers' attributes. Thus, the server configuration rules map the abstracted server configuration parameters to concrete, direct server values that are directly used to configure the servers. Because the rules perform this mapping based on attributes of the servers, they are applied to the server configuration rules at boot time of each server. The attributes may include the processing speed of a server, the memory of a server, the storage of a server, and so on.
The profile template loaded is preferably not a complete configuration, but rather a configuration skeleton that may provide only a subset of the needed data to configure a server, whereas a given server configuration needs to define all such data. Therefore, the user is requested (108) to modify any preset values of the abstract server configuration parameters, and/or enter any missing values of the parameters. A given profile template may be for high-performance OLTP servers, for instance, and as such may specify preset values for the parameters that the user may modify. The template may also have missing, or blank values, for other parameters that the user must enter to create a server configuration. Preferably, 108 of the method 100 is performed by the user at a client interacting with the configuration server.
The profile template may itself be hierarchically related to one or more parent profile templates, from which it initially inherits some of the abstracted parameters and configuration rules that are then modified, deleted, or supplemented by the template. For example, there may be a baseline OLTP server profile template, on which a high-performance OLTP server profile template is based. The latter template inherits some of its abstracted parameters and configuration rules from the former template, which are then modified, deleted, or supplemented as necessary.
Once any missing values of the abstracted server configuration parameters have been entered, and the user has modified any preset values as desired, the configuration server generates a configuration for each of the servers (110). The configuration for each server includes the values of the abstracted server configuration parameters, including the values entered or modified by the user, and the server configuration rules. Preferably, the configurations for the servers are stored at the configuration server itself. In one embodiment, 110 is alternatively performed within the part 104, preferably when booting, rather than within the part 102, so that a server can be reconfigured based on changes made to the server profile without the user having to explicitly initiate 110.
With respect to the part 104 of the method 100, each server loads its configuration (112). The configuration may be loaded when booting, or at other times as well, including when a significant event occurs, such as a hardware fault or the reconfiguration of a resource. Where the configuration is stored at the configuration server, a server thus requests and receives its configuration from the configurations server. The server then applies the server configuration rules of the configuration to the abstracted server configuration parameter values of the configuration (114), based on the server's attributes. This enables the server to generate direct server values by which it can be configured. The server then configures itself using these direct server values (116).
For instance, the abstracted server configuration parameters may specify how the user, such as an administrator, wishes a collection of servers to perform. Each server then loads its configuration and applies the server configuration rules to the values of these parameters to determine exactly how it should configure itself in light of these parameters and the server's attributes. The direct server values thus are the specific and concrete parameters of the servers by which the servers configure themselves. The user does not have to set or track these values him or herself, but rather only has to specify the more abstract, high-level server configuration parameters.
Background Technology: Example System
The configuration server 202 has an internal or external storage 208, on which the server 202 stores the profile templates for constructing the configurations for the servers 204, and also preferably stores the configurations themselves for the servers 204. User interaction is preferably involved for constructing the profile templates and the configurations. This may be accomplished by the user communicating directly with the client 210, which communicates with the configuration server 202 over the network 206. Similarly, when any of the servers 204 is booting or otherwise requires its configuration, it communicates with the configuration server 202 over the network 206.
Profile Template Construction via Inheritance
Profile templates are initially constructed prior to their use to construct specific configurations. The construction of a profile template may be included as an initial part of the method 100 of
Within the system 200 of
The diagram 300 of
More specifically, the template 302 includes the abstracted server configuration parameters 308A and the server configuration rules 310A. Therefore, the template 304 inherits the parameters 308A and the rules 310A from the template 302, and also adds the parameters 308B and the rules 310B. Similarly, the template 306 inherits the parameters 308A and 308B from the template 304, as well as the rules 310A and 310B from the template 304. Added to the template 306 are the parameters 308C and the rules 310C.
Configuration Construction from Profile Template
Once a profile template has been constructed, specific server configurations are constructed using the profile template. The construction of server configurations is indicated as the part 102 of the method 100 of
The diagram 400 of
Configuring Server Using Previously Constructed Configuration
Once a server configuration has been constructed, the server for which the server configuration is meant can be configured, preferably at boot time, or at another time, such as when a significant event occurs. Configuring a server using a previously constructed configuration is indicated as the part 104 of the method 100 of
Within the system 200 of
The diagram 500 of
Example Server Configuration
An example server configuration is next described. As may be appreciated by those of ordinary skill within the art, the example configuration described is for illustrative purposes only, and does not represent a limitation on the invention itself. The basic elements of the configuration can be described in general terms by the following grammar.
configuration::=serverld (configParam)*(configRule)*;
configParam::=paramld paramType value;
configRule::=ruleld target [paramsReqd]((configExpr)*|(configVector)*);
configExpr::=[condition]expression;
configVector::=[condition](configItem)*;
configItem::=metaRef item;
A server identifier serverId specifies a configuration as a set of abstracted server configuration parameter values configParam, and a set of server configuration rules configRule. Each configuration parameter value configParam is a name-value pair with a value for the type paramType, such as a string or a number. The parameter identifier paramld is used in the expression of rules, and is substituted at time of rule application or evaluation with the associated value.
Each configuration rule configRule has an identifier ruleId, a target that specifies the direct server values that will be generated based on any required parameters or attributes paramsReqd, and a collection of expressions configExpr or vectors configVector. Each expression or vector has a conditional clause that invokes the expression or vector only when it evaluates to true. This provides a cascaded if-then arrangement for the rules. A vector combines several direct server values configItem into a single structure, for ease of use. Expressions are represented as a common expression language with standard arithmetic operations, logic operations, and potentially specialty operations, such as minimum, maximum ceiling, floor, and so on.
Using this grammar, a sample server configuration is as follows.
configuration (serverId=‘exampleServer’)
configParam (paramld=‘Usage’ paramType=‘string’)
value=‘Heavy’
configParam (paramld=‘ServerName’ paramType=‘string’)
value=‘exampleServer’
configParam (paramld=‘StorageDir’ paramType=‘string’)
value=‘/work/storageDirs/databases’
configRule (ruleld=STORAGE)
configExpr
target=#config:STORAGE
expression=composePath($StorageDir, ‘rootdbs’+‘.’+$ServerName)
configRule (ruleld=TYPE)
configExpr (condition=#/fmx-data/node/NodeName.eq. ‘A-OS’.and.
$Usage.eq. ‘Heavy’)
target=#config:TYPE
expression=‘MAX’
configExpr (condition=#node/NodeName.eq. ‘B-OS’.or.
$Usage .eq. ‘Medium’)
target=#config:TYPE
expression=‘MED’
configExpr (condition=#node/NodeName.eq. ‘C-OS’.or.
$Usage.eq. ‘Light’
target=#config:TYPE
expression=‘MIN’
In this example, two direct server values are being set, storage and type. The storage direct server value shows how the expression capability can be used to combine multiple abstracted server parameters into a single result. The type direct server value shows how the conditional expression capability can be used to have a series of rules that can adapt to various attribute values. Thus, the parameter Usage and attribute node NodeName are evaluated to set the type of server to MAX, MED, or MIN. Parameters are indicated by the indicator $, whereas attributes are indicated by the indicator #.
Advantages over the Prior Art
Embodiments of the invention allow for advantages over the prior art. Profile templates may be constructed, preferably by using previously constructed profile templates to inherit baseline abstracted parameters and configuration rules. Specific instances of server configurations may then be constructed from the profile templates, where a configuration is interpreted at server boot time, or at another time when the configuration is loaded, to particularly configure a server based on its attributes.
An administrator does not, therefore, have to manually track all the direct server values that proper server configuration may require setting. Rather, the administrator only has to manipulate values for the abstracted parameters. At boot time or at another time when the configuration is loaded, the server configuration rules are applied to the parameters based on the server's attributes to generate appropriate direct server values. The administrator thus accomplishes server configuration at a high level.
Furthermore, a developer does not have to manually construct a utility to assist the user in setting the direct server values. The developer may only have to construct new profile templates, which themselves are preferably based on previously constructed profile templates so that redundant effort is avoided. The developer thus avoids having to code utilities, and instead only has to create those parts of new profile templates that vary from already existing profile templates.
Alternative Embodiments
It will be appreciated that, although specific embodiments of the invention have been described herein for purposes of illustration, various modifications may be made without departing from the spirit and scope of the invention. For instance, the invention has been described within the context of a system topology that includes a configuration server, a number of servers to be configured based on configurations stored on the configuration server, and a client for constructing these configurations in interaction with the configuration server. However, other system topologies are also amenable to the invention. Accordingly, the scope of protection of this invention is limited only by the following claims and their equivalents.
Number | Name | Date | Kind |
---|---|---|---|
5165018 | Simor | Nov 1992 | A |
5819030 | Chen et al. | Oct 1998 | A |
5872928 | Lewis et al. | Feb 1999 | A |
6449715 | Krivoshein | Sep 2002 | B1 |
6536014 | McClannahan et al. | Mar 2003 | B1 |
6672505 | Steinmetz et al. | Jan 2004 | B1 |
20020112038 | Hessmer | Aug 2002 | A1 |
Number | Date | Country | |
---|---|---|---|
20030225867 A1 | Dec 2003 | US |