US20020169739A1

US20020169739A1 - Method and apparatus for composite analysis using hierarchically organized directives

Info

Publication number: US20020169739A1
Application number: US10/151,803
Authority: US
Inventors: Adam Carr; Paul Holland; Mark McDowell
Original assignee: Hewlett Packard Co
Current assignee: Hewlett Packard Development Co LP
Priority date: 2001-05-10
Filing date: 2002-05-21
Publication date: 2002-11-14
Also published as: US20040015908A1; US6952689B2; US20030177106A1; US7146350B2; US20040006546A1; DE10220938A1; US20030135476A1; US20030140278A1; US20020169734A1; US20040015907A1; US7028221B2; US20020169738A1; US20030154177A1

Abstract

A method and apparatus is provided for auditing the configuration of an enterprise using low-level analyzers that include a program and a descriptor of the type of collector data the program can accept, and high-level analyzers that are descriptors which call upon, or incorporate by reference, other analyzer descriptors. A framework, guided by one or more analyzer descriptors and by any descriptors those descriptors may call upon or incorporate by reference, repeatedly executes the set of programs associated with some of the analyzer descriptors against a different node or nodes of an enterprise during each such execution. When executing a program, the framework provides the program with the collector data of the type designated by the descriptor associated with the program.

Description

CROSS REFERENCE TO RELATED APPLICATIONS

This application is a continuation-in-part of application Ser. No. 09/851,963 filed on May 10, 2001.[0001]

FIELD OF THE INVENTION

The present invention relates generally to the field of auditing a configuration of an enterprise, and more particularly, to a method and system for analyzing the configuration information of an enterprise based on expert knowledge.

BACKGROUND OF THE INVENTION

The core business of companies today is more and more dependent on information technology (“IT”) services. As such, the impact of outages or interruptions in service is more and more significant. Businesses therefore require continuous operation of an infrastructure or system. Because of this requirement, an infrastructure that is “always on” is becoming the norm rather than the exception. The technology underlying that infrastructure must be configured to maintain continuous and optimal operation (configured “always right”) of that infrastructure.

To obtain optimal configuration, a company may desire to perform an audit of the configuration of its infrastructure. The infrastructure may, for example, include a cluster configuration consisting of three nodes. In general, a cluster configuration is a group of nodes that are arranged to “back up” and substitute for the execution of a program in the event a node is unable to execute that program or the node is unavailable. This company may perform the audit or may typically hire an outside company firm to perform the audit. In this example, the audit usually involves the following five steps:

Resources—finding someone with the right skill level to do this system configuration audit.

Configuration—looking at a complex enterprise configuration either directly or by collecting all the relevant configuration information needed for the investigation.

Analysis—coming to conclusions.

Reporting—documenting the findings and presenting this information to a company requesting the audit.

Action Plan—creating an action plan to address the issues found.

Unfortunately, there are several issues or problems along the way that need to be overcome to make a successful audit.

Resources. A company must find the right personnel. To audit a system configuration a person requires practical experience with the underlying technology. The successful candidate must have sufficient technical and professional skills along with practical experience. Typically, no single person possesses expertise in all required areas. Time and money constraints limit the number of resources on an audit activity, limiting in turn the content of the delivery (depth, breath, . . . ) and also the quality. The audit may come to a halt if multiple people decide to change jobs. High turnover usually translates into a knowledge drain unless such knowledge is documented. In the event of a departure, a company must invest in training which is time consuming. This is no guarantee that the company actually captured the knowledge and expertise of the departing employee. This knowledge is an asset to the company that is lost forever. Not only must this knowledge be captured and kept, but it must also be utilized effectively. It must be accessible, and automatic access is highly desirable.

Configuration. The next issue concerns how to obtain the configuration information. In some circumstances the information has been previously obtained. However, this configuration information may likely be in a format unusable for analysis. In the absence of this information, it must be retrieved. In the event the audit is being performed by an outside company, manual interaction with a customer's system should be avoided even if the customer permits access. There is a risk that the interaction may cause a problem with the customer's platform. The customer will likely blame the outside company for problems with the system regardless of fault.

Today, there are software tools available for the collection of configuration information. If such a tool was developed locally, reliability and maintenance is a concern largely because such tools are incomplete when they are developed. Also, the quality of a local tool will be limited by the local expertise.

Now, the tools described require installation on a customer system which may make the system unavailable or may cause it to crash. In addition, the customer may challenge the reason why he/she requires the tool to be installed on his/her systems. Customers with rigid change management in place will not allow the installation of any tools on short notice.

Analysis. Even if one is successful collecting configuration information, there are other obstacles. The analysis performed at the customer site will likely require multiple visits. This is valuable time lost. In addition, there are typically limited available resources on-site. Further, analysis typically requires the application of several analyzers to identify issues. Then there is the question of what parts of the node must be checked or analyzed and what issues should be identified. Both questions are typically answered by a single individual. Because of the limited knowledge of that individual, the system checks and issues identified may not be fully exhaustive. It is important, however, to rely on a well defined list of items to be checked and criteria of satisfaction to ensure a reliable and stable environment. It would be advantageous to use input from multiple sources of expertise, but this is rarely ever practical.

As it concerns a cluster configuration, the analysis of the configuration is not limited to the individual systems. In cluster configurations, differences between system or node configurations (e.g., installed software, memory size, kernel configuration, etc.) are important. Therefore, the nodes in a cluster must be compared against each other. With no tools in place to do this, it will be a manual effort to extract certain configuration information and possibly to write scripts to do a comparison and provide the results in a presentable way. This is a laborious and time consuming process, but a necessary task.

Reporting. If the analysis is accomplished and a list is generated identifying issues (problems) with the cluster configuration, there are still other obstacles. The issues identified are not organized in any logical way or according to customer requirements. Information should be sufficient to provide an adequate description that is to the point, professional and accurate, and that caters to different audiences (technical & non-technical).

More information may be included as desired. Descriptions in the reports, however, may reflect an individual's personal vision as opposed to a company's uniform recommended practice. This will result in inconsistencies among deliverables, sending mixed messages to customers. Further, it is also important to assign the correct description to each piece of information. This would appear obvious but becomes less so when handling similar pieces of information for different systems.

Now, the report presented should be consistently formatted. Technical personnel should not have to spend their time writing reports, when their technical skills may be used in better ways. This would therefore require additional resources for personnel skilled in technical writing. Further, the reports must cater to the audience requesting the audit. The audience may include technical and non-technical management, sales people, and many others. A system is desired that is capable of crafting different reports, with the results of the analysis in accordance with an auditor's request.

Action Plan. Typically, the company performing the audit must prepare an action plan to resolve the issues that were determined. If the company does not act aggressively, the company may lose the business. However, in many situations, the company may have little assistance available from the audited company for the creation of an action plan. The preparation of the action plan is facilitated when a scenario is available with the steps to resolve a specific issue, together with the means to find additional reference material.

In summary, in today's world, companies are relying more and more on their IT systems for core parts of their business. This produces ever increasing requirements for reliability, availability, scalability, and performance. Therefore, companies are increasingly becoming more demanding because the consequences are severe when their IT systems suffer downtime. Also, the speed at which things are done is increasing, and accordingly the turnaround time for consulting deliveries is decreasing.

It would be desirable to provide an automated system that overcomes the disadvantages described above.

It would also be desirable to achieve an automated system that performs a complete audit of a system configuration generating the necessary and desired audit reports automatically.

SUMMARY OF THE INVENTION

Briefly summarized, the invention in one of its embodiments resides in a method carried out in the environment of a system that analyzes data gathered by collectors monitoring nodes. It is a hierarchical method used to define and to organize the analysis tasks to be performed.

This method comprises performing, for each low-level analysis task, analysis steps that comprise accepting data from one or more collector types, analyzing the data, and generating reports, guided by an analyzer descriptor written in XML or a comparable format identifying the collector types whose data the analysis steps are to process. Then, for each higher-level analysis task of which there is at least one, performing one or more sets of analysis steps for either low-level or higher-level analysis tasks, guided by an analyzer descriptor written in XML or a comparable format that calls upon or that incorporates by reference other analyzer descriptors. Next, guided by one or more analyzer descriptors and, in the case of higher level analyzer descriptors, by the analyzer descriptors they call upon or incorporate by reference directly or indirectly, repeatedly performing the set of analysis steps corresponding to the one or more of these analyzer descriptors one or more times against collector data gathered from a different node or set of nodes during each such exercise of the set of analysis steps. And during this process, presenting for acceptance and analysis during the performance of each data acceptance step data gathered from collectors of the type indicated by the guiding analyzer descriptors.

Another embodiment of the invention is a system that analyzes data gathered by one or more collectors monitoring one or more nodes. The system comprises an analyzer program and descriptor for each low-level analysis task that the system performs, the program containing instructions that accept data from one or more collector types, analyze the data, and generates reports, and the analyzer descriptor written in XML or a comparable format and identifying the collector types whose data the program's instructions can accept. For each hierarchically higher-level analysis task, the system further comprises at least one analyzer program, and an analyzer descriptor written in XML or a comparable format that calls upon, or that incorporates by reference, other analyzer descriptors. And the system also comprises a framework that is guided by a set of one or more analyzer descriptors and, in the case of hierarchically higher-level analyzer descriptors, by the analyzer descriptors they call upon or incorporate by reference directly or indirectly. Under this guidance, the framework executes repeatedly one or more times the set of analyzer programs corresponding to some of these analyzer descriptors against collector data gathered from a different node or set of nodes during each such execution of the set of programs. And also under this guidance, during each execution of each program, the framework presents the program with only the type of collector data the instructions of the executing program are designed to accept, as indicated by the analyzer descriptor corresponding to the executing program.

BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS

FIG. 1 presents an overview block diagram of an enterprise analyzer system that periodically collects information from the nodes in an enterprise, analyzes that information, and uses the issues which result from the analysis as a basis for generating a variety of reports. [0027]
FIG. 2 is a hardware block diagram illustrating the nodes in an enterprise linked to an analyzer server and a report generator server. [0028]
FIG. 3 is a hardware block diagram illustrating the computers and other nodes in an enterprise having a support node. [0029]
FIG. 4 illustrates the organization of a collector database. [0030]
FIG. 5 illustrates a process of creating a collector. [0031]
FIG. 6 illustrates a process of installing a collector on an enterprise. [0032]
FIG. 7 is a flow diagram of the process of periodically executing collectors within an enterprise. [0033]
FIG. 8 is a software block diagram of the elements of the analysis server together with certain elements of the report generator server and an HAO Server. [0034]
FIG. 9 illustrates the organization of the tracker database. [0035]
FIG. 10 illustrates the process of an auditor designing an assessment task to analyze an enterprise and to generate reports. [0036]
FIG. 11, a continuation of FIG. 10, illustrates the process an auditor goes through to define the reports that are to be generated. [0037]
FIG. 12 is a flow diagram illustrating the process of creating a new analyzer. [0038]
FIG. 13 is a continuation of FIG. 12 illustrating the process of creating a new analyzer. [0039]
FIG. 14 is a table listing the user interface tools which are desirable to aid in analyzer creation. [0040]
FIG. 15 is a flow diagram of the general process executed by the analyzer harness when it performs the analysis of the nodes in an enterprise. [0041]
FIG. 16 illustrates the process of creating a report template. [0042]
FIG. 17 is a continuation of FIG. 16 illustrating the process of creating a report template. [0043]
FIG. 18 is a continuation of FIGS. 16 and 17 illustrating the process of creating a report template. [0044]
FIG. 19 is a flow diagram illustrating how reports are generated. [0045]
FIG. 20 is a continuation of FIG. 19 illustrating how reports are generated. [0046]
FIG. 21 is an overview block diagram of the elements involved in the execution of analyzers including the analyzer loader, the analyzer descriptor, and the analyzer argument manager. [0047]
FIG. 22 presents a flow diagram of the detailed steps carried out by elements of the analyzer harness (shown in FIG. 21) when it executes an analyzer. [0048]
FIG. 23 is a continuation of FIG. 22 illustrating the steps carried out by the analyzer harness when it executes an analyzer. [0049]
FIG. 24 is a continuation of FIGS. 22 and 23 illustrating the steps carried out by the analyzer harness when it executes an analyzer. [0050]
FIG. 25 illustrates an XML hierarchy which can be used to generate varying scope assessments of an enterprise. [0051]
FIG. 26 illustrates a process of defining a task in the environment of the XML hierarchy shown in FIG. 25.[0052]

DETAILED DESCRIPTION

Definition of Terms [0053]
The following terms used in this application shall have the respective meanings ascribed to them below unless otherwise expressly defined in this application. [0054]
Analyzers. Analyzers are processes, defined by rules or programs, that analyze selected configuration information gathered by one or more collectors to one or more nodes, and that identify and report issues which may require attention. [0055]
Analyzer Harness. An analyzer harness is a framework or system which encapsulates, or wraps, and then executes an analyzer, providing the analyzer with collector data relating to a specific node or nodes of a specific enterprise each time an analyzer is encapsulated or wrapped and then executed. [0056]
Audit. This term is normally used for a formal examination or verification of financial accounts, in this context it means a very thorough examination or verification. One can expect a high depth of coverage and typically (but not necessarily) a significant breadth of coverage (i.e., multiple domains). Given the details, an audit typically takes a significant amount of time. Resources that perform Audits typically are subject matter experts with a lot of experience. Depending on the areas of coverage this is done by one person or by a team. Typically, this will require a team of multiple specialists. [0057]
Collectors. A collector is a command, or a series of commands, that causes programs installed at one or more nodes to gather configuration information about the node and that return reports defining the node's configuration. [0058]
Configuration. Any information specific to the static or the dynamic configuration of one or more nodes (or field computers, hardware, software, firmware, databases, storage systems, etc.) at a given point in time. [0059]
Configuration Tracker. A tool which gathers configuration information from one or more nodes and which highlights changes between snapshots of the information gathered at different times. [0060]
Daemon. A Unix term that refers to a specific type of program or agent designed to work in the background. [0061]
Enterprise. Collection of computers, software, and networking that comprises the computing environment of a business. [0062]
Enterprise Name. A name that identifies a particular enterprise and that may be used as a retrieval key to access the configuration information of the enterprise. [0063]
Field Computers or Field Nodes. Field computers or nodes are computers or nodes installed at enterprise sites and whose operations are monitored from a central site. [0064]
Framework. This is a symbolic structure in which expert knowledge and experience is placed such that it may be automatically used to analyze enterprise configuration information gathered by collectors to identify issues that may require management attention. More specifically, in the context of this embodiment, expert system rules called analyzers are placed into a harness or framework and supplied with configuration information from enterprise nodes and are thereby enabled to report any issues that may have arisen within the enterprise. [0065]
High Availability Observatory (“HAO”). In this embodiment of the invention, the High Availability Observatory is a software entity installed on a support node within an enterprise that periodically causes the execution of collectors on the enterprise nodes, captures the reports which the collectors generate, and transfers those reports back to a centralized tracker database for later use and analysis. It is noted that the HAO is merely one of many possible ways in which configuration information may be collected and analyzed in accordance with the teachings of the present invention. [0066]
Issue. An issue is any matter that may need to be investigated by or reported to the management of an enterprise. An analyzer performs tests upon configuration data gathered from the nodes of an enterprise by collectors. Those tests determine if there are any issues that need to be drawn to management's attention. [0067]
Issues Database. An issues database is a database containing issue information generated by analyzers when they process configuration information generated by collectors operating on the nodes of an enterprise. [0068]
Managed Element. A managed element is one or more physical devices (CPU, node, computer, etc.) or logical devices (program, volume, thread, process, etc.) that are monitored and/or managed. Each analyzer operates upon a particular class of managed elements, which may be a single node. A set of analyzers can be assigned to analyze all the members of a managed element of an enterprise. [0069]
Node. A node is a particular device in an enterprise, such as a server, work station, printer, router, switch, or hub. A multi-processor may be configured either as a single node or as multiple nodes. [0070]
Tracker Database. A tracker database is a database containing configuration information gathered from one or more nodes of one or more enterprises. [0071]
Overview of the System [0072]
Referring to FIG. 1, an overview block diagram is shown of an [0073] automated method 100 for repeatedly analyzing the configuration of an enterprise. The method 100 may generally be broken into three stages: collection, analysis, and reporting, as is shown in FIG. 1 by the dashed lines.
In [0074] step 102 of the method, enterprise configuration information is collected from field nodes. The collection step utilizes a set of collectors 104 to gather the desired configuration information. The collectors 104 are each commands or programs stored on a support node in the enterprise. The collectors 104 are run at periodic intervals on each node of the enterprise. The collectors 104 gather desired configuration information and store it as files in the tracker database 106. Specifically, for each node, there is a configuration information file stored on the tracker database 106 associated with each and every collector that generates configuration information.
In [0075] step 108, configuration information is analyzed by an analyzer harness 806 (FIG. 2) to identify issues within the field nodes. As part of this process, the analyzers 110 are used to investigate particular issues associated with the field nodes. The analyzers 110 are developed or written by experts to identify an issue or issues. The experts possess knowledge and information which is captured in written form as analyzers addressed to specific issues. The analyzers are then stored in an analyzer database 804 (FIG. 2) for later use by the analyzer harness 806.
An important aspect of the present invention is that the reports generated by this arrangement focus on specific issues rather than upon the status of an enterprise. Accordingly, the support engineers do not have to wade through large amounts of configuration information to isolate problems. Providing support for the enterprise is thus transformed from configuration gathering and manual analysis to executing a specified set of analyzers and allowing the framework “expert” to diagnose the problem and identify the pertinent issues. [0076]
As will be discussed in greater detail below, the [0077] analyzer harness 806 executes the desired analyzer 110 with the configuration information stored in the tracker database 106 and generates a report 2126 (FIG. 21) in XML format. The report 2126 identifies issues relating to the field nodes. This issue identifying report 2126 is then stored as a file in an issue database 112. If an issue has not arisen with respect to any node, the issue will be absent from the report 2126.
At this stage, the [0078] report 2126 generated by step 108 may be used to generate a full report along with text descriptions of the issues, as desired by an auditor. The XML report 2126 from the analyzer harness 806 is sent to a report generator 206 (FIG. 2). The report generator 206 may use report templates and rules database 204 to generate reports 208 (FIG. 2) addressing specific issues for each field node. These reports 208 identify the analyzers 110 used, the node from which the collector information is gathered, optionally the actual output of the collector 104, and the issue identified.
Now, the issues stored in the [0079] issues database 112 may be used in another way to review the overall performance of the enterprise. To this end, in step 114 the issues are analyzed using rules written by the experts, and a report is generated as desired by the auditor. Generally speaking, the reports are generated from templates stored in the report templates and rules database 204. The reports may be presented in step 118 to enterprise management, technical management, the field engineering team, and to a workflow system or healer system (self-healing technology). This self healing technology will help the computer industry achieve the goal of providing servers and other node that repair themselves automatically, thereby staying in service without the need of human intervention over many years. The present invention, by automatically identifying issues that can indicate the possible future failure of nodes in enterprises and by generating reports of the type that can control the operations of other nodes which can download software to
The [0080] analyzer server 800 includes an analyzer database 804 which stores the analyzers 110 and an analyzer harness 806 for wrapping the analyzers 110 retrieved from the analyzer database 804 with the files of configuration information retrieved from the tracker database 106. The analyzer harness 806 generates an issue report file in XML format which is placed into the issues database 112. As discussed more fully below, this report may be used to generate a full issue report as described above with respect to FIG. 1.
The [0081] legacy server 210 includes a legacy database 212. The legacy database 212 stores configuration information files obtained during prior manual collections or other means. The legacy database 212 can be linked to the analyzer harness 806. In the event that the HAO collectors are unavailable or not present to automatically and continually generate configuration information files, configuration information files can be retrieved from the legacy database 212. This information, however, is only as current as the most recent manual collection. But other automated techniques may be used to place node configuration information into the legacy database 212.
Note that the HAO is only one way in which configuration information may be gathered from an enterprise. This can and should be done for every node in each enterprise so that the benefits of configuration information collection and analysis can be brought to bear universally. [0082]
The [0083] report generator server 202 is also a part of the central site. The report generator server 202 is linked to the analyzer server 800 through the (XML) issues database 112. The report generator server 202 includes a report generator 206 for receiving the XML issue reports from the database 112. As discussed above with respect to FIG. 1, the report generator 206 may generate a full report concerning the identified issues including more detailed information about each issue.
Now, the issue reports are stored in an [0084] issue database 112 as described. A report templates and rules database 204 is part of the report generator server 202. The server 202 stores various report templates and rules which are developed by subject matter experts. These can be field personnel or product development reinitialize disk drives and to perform other maintenance services automatically, moves the computer industry a giant step closer to providing servers and other nodes that maintain and repair themselves automatically with only minimal human intervention.
Referring to FIG. 2, there is shown a hardware block diagram for implementing the method in FIG. 1. An [0085] enterprise 300 is located on a customer site, and it includes a support node 308 which may be operated by the manufacturer of the computers at the customer site. Note that the support node 308 includes the customary components of a computer system including a CPU, a display or other I/O device, a network or communications interface, RAM or ROM or other memory, as well as stable storage devices such as disk or CD-ROM drives. Other servers or nodes described may also include the same customary components.
Software known as High Availability Observatory (“HAO”) available from Hewlett-Packard, Incorporated is stored on the [0086] support node 308, and it manages the collectors 104 that gather configuration information. In the enterprise 300, the support node 308 is connected to the nodes 302 and 304 by a network which enables the collectors 104 to gather configuration information. Note that the enterprise 300 has a firewall 324 to act as a barrier to prevent unwarranted access to the enterprise 300 and to act as a central point of presence for maintenance and e-mail connectivity. Note that use of a support node in this manner is one of many ways in which configuration information may be collected and subjected to analysis.
FIG. 2 also illustrates the components located at a central site. The central site includes an [0087] analyzer server 800 situated remote from the enterprise 300. The analyzer server 800 (in this embodiment, indirectly through the tracker database 106 as is shown in FIG. 2) is linked to the enterprise 300 via an ISDN line or some other form of wide area network. An HAO server 105, a report generator server 202, and a legacy server 210 are also part of the central site. The HAO server 105 is linked to the support node 308 by the ISDN line. Files of configuration information generated by the collectors 104 are saved in a tracker database 106. These files are retrieved from the tracker database 106 by the analyzer server 800, as is described below. personnel. The application of these rules helps to determine the overall performance of the enterprise 300. At an auditor's request, the report generator 206 will retrieve (1) the rules from the report templates and rules database 204 and (2) the issues stored in the issues database 112, and will then generate a report 208 using the templates from the report templates and rules database 204. The report may be in any desired format, such as Microsoft Word, Excel, PowerPoint, or HTML or in some special format designed for delivery to some other computer or node for use in automatic control. The report may identify all the issues with the enterprise 300 to enable the auditor determine the overall performance of the enterprise 300.
Referring to FIG. 3, the cluster configuration example described above, along with an analysis of the enterprise configuration, will now be discussed in more detail. As part of the [0088] enterprise 300, three nodes 302, 304, 306 and several personal computers 314-322 are linked through a network 312. The nodes 302, 304, 306 include utilities 303, 305, 307, respectively to set up and configure the nodes properly. The enterprise 300 also includes a file server 326 with a volume 328 for storing programs. The file server 326 is linked to the nodes 302, 304, 306 to enable these nodes to access the programs from the volume 328 on the file server 326. As discussed above, the HAO support node 308 is also part of the enterprise 300, and it includes a collector database 310. The HAO support node 308 is linked to the HAO server 105 on the central site via an ISDN line or other means of connection. A firewall 324 is installed on the enterprise 300 for the reason described above.
The three [0089] nodes 302, 304, 306 are arranged in a cluster configuration. In this configuration, the nodes are each configured as a back-up to each other to ensure that executable programs can be run on a back-up node if the node assigned to run a program fails. To configure the cluster configuration properly, the utilities of each node must be set up to ensure that the node recognizes or “sees” the program on the volume 328. In other words, the volume 328 of the file server 326 must be “attached” to each of the nodes 302, 304, 306 to enable each to run a program or access a program on the volume 328. If a PC 314, for example, is using a program through the network 312 that is normally run on the node 302, the volume 328 must be attached to the node 304 to run that program in the event the node 302 fails or is otherwise unable to execute that program. The cluster configuration is realized using a Hewlett-Packard, Incorporated program known as MC/ServiceGuard. The MC/ServiceGuard program creates a cluster configuration which ensures that this type of back-up occurs.
Now, in this example, it is desired to analyze the configuration information of the cluster arrangement. To this end, the [0090] collectors 104 from the collector database 310 on the HAO support node 308 are sent to the nodes 302, 304, and 306 to gather configuration information. This information, in file form, is then returned and is sent to the central HAO server 105 (FIG. 2) to be stored as files on the tracker database 106. The tracker database 106 identifies the node from which the information was gathered, and the files contain facts concerning what was collected, such as storage or memory capacity, etc. At an auditor's request, the analyzer harness 806 (of analyzer server 800) retrieves the desired configuration information files and the desired analyzers 110 from the analyzer database 804 and wraps them together, forming a framework that can generate issue reports 2126 in XML format. Each issue report 2126 indicates typically that a particular problem exists within a particular node. For example, the report 2126 may report that the node 304 does not recognize the volume 328 on the file server 326 because it has not attached this volume 328. If the node 304 cannot recognize or “see” the volume 328, node 304 cannot run a program in the event that the node 302 fails.
Another [0091] exemplary report 2126, discussed below, might indicate that the storage or memory capacity of the volume 328 attached to the node 304 is a less than 5% of its total capacity. This capacity level may not be sufficient to provide for expected growth in the number or size of the files stored on this volume.
Detailed Discussion of System Elements [0092]
With reference once again to FIG. 1, the [0093] collectors 104 are commands or sets of commands that can run on each of the nodes of an enterprise, testing the configurations of those nodes, both static and dynamic, and generating collector reports reporting their findings. At step 102, the reports generated by these collectors, in the form of files, are collected and are transmitted to a centralized tracker database 106 where they are stored. Many types of collectors can be used in this manner. A family of collectors 104 that are run at a particular enterprise site are defined by the service engineers at that site and may be the very same programs that the service engineers run when they visit the site and manually run tests to learn about the system configuration.
At the central site, the [0094] analyzers 110 read and analyze the collector information contained in the tracker database 106. Each analyzer is created by an expert in system testing and maintenance to evaluate the information generated by one or more collectors and to look for one or a few specific potential problems that may be present. These problems are called “issues”. As a simple example used throughout the remainder of this specification and in the appendices, an issue that might be of concern to the management of an enterprise could be the fact that one or more volumes attached to one or more nodes are filled to within five percent of their maximum capacity and are in danger of possibly running out of room to store additional files. Different analyzers test for different conditions and focus upon different issues. At step 108, the analyzers are called upon to analyze the configuration information in the tracker database 106 and to thereby generate issue information which is stored in an issues database 112.
In addition to the [0095] collectors 104, which define what tests are run upon the individual nodes of an enterprise, and in addition to the analyzers 110, which search through the configuration information gathered by the collectors 104 looking for issues that may need to be brought to the attention of management, the present invention contemplates the use of additional rules and report templates 116. These enable the performance, at step 114, of a further analysis of the issue information using rules written by experts; and the generation of a variety of useful reports directed to different audiences.
As indicated at [0096] step 118, one of these reports might be a summary report, written in laymen's language, directed to the non-technical management of an enterprise to advise them of critical problems and trends. This report may indicate general conditions that exist, and it may suggest reasons why this management should be concerned about these conditions through the use of simple examples and explanations. Other possible reports would be reports for technical management which might be fairly detailed, indicating in technical terms precisely what problems exist and what solutions might be considered. A separate set of reports might be generated for the field engineering team documenting the status of the enterprise, indicating points that need to be watched, and possibly suggesting field corrections.
An [0097] issue description database 117 may be provided which can supply the report generating software with highly detailed descriptions of every issue that may arise, including examples and also explanations of why these issues are of concern to management, written in non-technical language that can be understood by non-technical managers. The issue descriptions in the database 117 thus supplement the very brief issue descriptions which result from analysis and which appear in the issues database 112.
In addition to being based upon the occurrence of issue information in the [0098] database 112, reports may be generated directly from the underlying tracker database 106 such that the reports may indicate the status of the enterprise as well as the occurrence of issues of interest to management. In addition, status information from the tracker database 106 may be used to supplement and augment reports otherwise based upon the occurrence of issues.
If work orders for the field engineers are generated through computers or other nodes in a work flow system, a machine-readable output of the report generator might be sent directly to the work flow system to cause it to generate work orders that cause service personnel to be sent immediately to the site to carry out time critical maintenance. For example, reports of failed disk drives in disk arrays can be transformed automatically into work orders that schedule a repair team to visit the site and to replace the drives before a mission-critical enterprise is shut down by further failures. If automated tools or “healer system” exist for making corrections to the nodes in the field, then some reports may be generated which become instructions to the healer system that cause immediate correction, through automated means, of certain problems to ensure continued operation of servers that are mission critical to the enterprise. For example, corrupted files can be downloaded and reinstalled automatically, and non-bootable systems may be repaired. [0099]
Technical reports can provide action plans which explain precisely how a problem can be addressed or a machine repaired. Through computer linkage, an electronic report can trigger the automatic ordering of parts and the scheduling a technician to visit the site to install the parts. It is only a small step beyond this to provide for automatic repair and correction of defects such that the nodes become substantially self healing. [0100]
FIG. 2 presents an overview hardware diagram of both an [0101] enterprise 300 located at one or more business sites and of an analyzer server 800 and report generator server 202 which might typically be located at a central site but which could be also located at an enterprise site.
In FIG. 3, the [0102] nodes 302 and 304 and possibly many other nodes exist at the enterprise site 300. These nodes would typically be servers, but they could be routers, printers, and other computational devices that are to be monitored. Also at the enterprise site 300, a support node 308 is networked together to the other nodes 302 and 304. The support node 308 includes tools (the HAO system) that enable the collectors 104 to run periodically upon the nodes 302 and 304 and other enterprise nodes, thereby generating information defining the static and dynamic configuration of the nodes 302 and 304 and other nodes. The support node program 402 (FIG. 4) captures the reports generated by the collectors 104 and transfers them through the firewall 324 and over some form of wide area network to the tracker database 106 within an HAO server 105 that is typically located at a central site. Accordingly, the tracker database 106 contains an up-to-date series of collector reports defining the configuration of the nodes 302 and 304 within the enterprise. This configuration information is available for service personnel to peruse and to study when trying to decide what types of services are needed to support enterprise 300 operations.
Some [0103] enterprises 300 may not be equipped with support nodes having the HAO program 402 and collectors 104 or their equivalent. Some or all of the configuration information for such enterprises 300 may be maintained upon a legacy database 212 (FIG. 2) within a legacy server 210. This information may also be used to support an analysis of an enterprise.
At the central site, an [0104] analyzer server 800 contains an analyzer database 804 which contains analyzers 110. These are rules or programs that implement tests defined by technical experts. The analyzer server 800 also contains an analyzer harness 806 which is able to harness individual analyzers 110 or rules together with configuration information gathered from specific nodes 302 and 304. This information is retrieved from the tracker database 106 or the legacy database 212. The analyzers 110 analyze this information and thereby determine whether there are any issues that need to be addressed. The analyzer harness 806 may receive an issue report each time an analyzer 110 is run if there are any conditions that need attention. The analyzer harness 806 adds to those issue reports information identifying the particular nodes 302 and 304 and the issue or error messages associated with the analyzer 110 and passes the issue reports out as XML reports 2126. These reports are is both human and machine readable. They list all of the issues which may need management attention. These are stored in an (XML) issues database 112.
The XML issue report is passed on to a [0105] report generator 206 within a report generator server 202. The report generator 202 includes a report templates and rules database 204 that controls the generation of a wide variety of different reports for different audiences, as was explained above. For example, some reports may simply restate each of the issues in a nicer format and sorted in a particular order for presentation to a technical team that needs to gain a detailed understanding of all the issues which have arisen. Other templates may be assembled under the control of elaborate rules which first scan and analyze the information contained within the XML issue report, possibly comparing it to previous similar reports generated on earlier occasions, to spot trends and various conditions that are in need of attention, thereby functioning as a configuration tracker. These reports may generate high level explanations of the state of the enterprise computers and that may be read and understood by lay people, including the management of the enterprise. Trends in the information not apparent in the detailed issue report may also be spotted and brought to light in this manner. Accordingly, a wide variety of technical and non-technical reports 208 in a wide variety of differing formats and levels of detail may be generated and passed on to management and to technical personnel, as well as to automated systems that may automate the scheduling of servicing or even cause the actual repair of the nodes in the field.
Collectors—Gathering Configuration Information [0106]
The next part of this detailed description focuses upon FIGS. [0107] 3 to 7 which explain how the collectors are created and then used to gather configuration information and to feed that information into the tracker database 106.
FIG. 3 presents a more detailed description of an [0108] enterprise 300. The enterprise 300 includes three servers 302, 304, and 306 which are identified as “nodes” in FIG. 3. Illustrative of the many possible servers, routers, printers, switches, and other audible devices that may be included within an enterprise, the nodes are tied together by a network 312 which also connects to a large number of work stations or PCS 314, 316, 318, 320, and 322. These work stations or PCS also qualify as nodes and may also be monitored by the present invention.
A number of [0109] utilities 303, 305, and 307 are installed upon each of the nodes 302, 304, and 306. While many utilities may be present on a particular machine, of particular concern to this description are utilities that might be useful to service personnel when investigating the configuration of the machine. These are utilities that can be run by service personnel and that would then generate reports which the service personnel might view on the monitor or print out and then use in the course of diagnosing various system problems.
While many such utilities would be useful, for the purpose of simplifying the present description of the invention, only one utility will be discussed in detail and used as an example throughout the description of the invention which follows. Clearly, many dozens of other utilities could be used in the same manner to generate all kinds of reports and to support all kinds of analysis of these nodes and the generation of all kinds of reports. [0110]
The selected utility is one called “bdf”. With reference to the Appendices, and in particular to Appendix E, the command [0111]
bdf-il [0112]
typed on a line by itself as a command issued to any computer or node gives a report on what [0113] volumes 328 from various file servers 326 are “mounted” upon a particular node 302, 304, or 306 whose configuration is being determined. A report is generated such as that shown in Appendix E. This report indicates the volume that is mounted, its size, how much of its capacity is utilized by files, how much is available for further storage, and the percent of the storage area that is utilized.
With references to lines [0114] 14 and 16 of Appendix E, it will be seen that two of the volumes have 96 percent of their available storage space filled with files. Given the importance of maintaining additional available storage space in file systems for the addition of additional files, service personnel examining this machine would probably conclude that the volumes were overly full and that adjustments must be made to ensure that additional files added to these volumes do not cause the volumes to overflow and to produce errors and to not accept any more files.
Within the [0115] support node 308, there exists a collector database 310, the contents of which are illustrated in FIG. 4. This database 310 includes HAO collector program 402 which is capable of launching a collector command, such as the “bdf” command illustrated in the preceding paragraph, upon any of the nodes 302, 304, or 306 and also upon the PCs 314, 316, etc. This collector program periodically causes collectors to run upon each of the nodes that are being supervised by this particular support node 308. This is done in accordance with a collection schedule 404. The nodes examined are included in a node list 406. The collector portion of the collector database 310 is shown at 408. It contains a list of all the collectors that are to be executed, together with a logical name and an ID number for each collector, as is shown for the “disk usage” collector “bdf-il” assigned the ID number “100024”.
In some cases, a collector may not be a simple command, such as illustrated above, but it may a series of commands to run certain programs or utilities in a certain order. Such a list of commands can be stored in a file which can be called a complex collector. Complex collector files are stored at [0116] 410 and are also stored within the utilities areas 303, 305, and 307 on the nodes 302, 304, and 306. In the case of a complex collector, the list of collectors 408 includes the name of the file that contains the list of commands which comprise the complex collector, rather than the actual collector commands themselves.
FIG. 5 illustrates the process of creating a new collector. At [0117] step 502, a new collector is created by simply adding it to the list of collectors 408. The collector command 506 is typed in along with a name for the collector 504 and its I.D. number 508. As soon as a new collector is added to the list of collectors 408, it becomes active. Of course, there may be several lists for different collectors that are run more frequently or less frequently on any given enterprise system.
Once a collector is created, it may be installed upon an [0118] enterprise 300, at step 602, by installing it upon the support node 308 of the enterprise within the collector database 310, and in particular in the portion 408 of the database 310 reserved for collectors. Complex collectors, consisting of multiple commands, must be stored as script files on the nodes 302, 304, and 306 with the names of the script files added to the list of collectors 408. Finally, at step 604, the collector is scheduled for execution periodically in the schedule 404.
FIG. 7 illustrates the execution of collectors against the [0119] nodes 302, etc. in an enterprise 300. Execution is typically triggered by a timer 702. (Execution may also be triggered by things like a configuration item which the system has determined has changed its state such that the change in state of the configuration item causes a related analysis to be carried out.) When a timer expires, the node list 406 within the collector database 310 is referenced, and the nodes are processed one at a time. At step 704, a node is selected. At step 706, the list of collectors 408 is referred to, and a first collector is selected for execution. A command is sent to the node 302 being audited to execute the designated collector. This can be done, for example, using a utility such as Telnet over an Internet network, but great care must be taken that the collectors are safe to execute and do not interfere with normal operations of mission critical nodes. In the HAO embodiment of the invention, a collection agent and a collection daemon (not shown) are installed on each node (see U.S. Pat. No. 6,148,402 which issued on Nov. 14, 2000 to Randall B. Campbell). The support node collector program 402 sends a collector 104 to the collection daemon on the node 302 to be audited. The daemon passes the collector 104 on to a collection agent. The collection agent executes the collector 104, captures its standard output and standard error outputs (STDOUT and STDERR), and returns the collected output information to the daemon, which returns it to the collection program 402 running upon the support node 308.
At [0120] step 710, if there are more collectors, then program control branches back to step 706 where additional collectors are caused to be executed upon that same node 302. Then, at step 712, if there are more nodes 304, etc. to be audited, then program control returns to step 704 where the next node 304 is selected, and the analysis proceeds as described above.
Finally, after all the [0121] collectors 104 have been run at all the nodes 304, etc. to be audited, and all of the information from the collectors 104 has been gathered, the collector output is transferred to the central tracker database 106 at a central site for further analysis.
Analyzing the Configuration Information [0122]
The analysis process is set forth in block diagram form in FIG. 8, this Figure shows the various databases and programs that are involved in the analysis and report generation processes. [0123]
With reference to FIG. 8, the [0124] analyzer server 800 is shown at the top of the figure and contains the analyzer database 804 and the analyzer harness 806. In the lower right portion of FIG. 8, the HAO server 105 is shown which contains the tracker database 106 that contains all of the configuration information with respect to the nodes 302 and 304, etc. in the enterprise 300. To the left in FIG. 8, the report generator server 202 is shown together with the database 204 that contains the report templates and rules and the report generator 206.
The first step in the analysis process is that of creating the [0125] analyzers 110 and creating the report generation rules and templates 116 that will generate the reports 208. As shown in the Figure, content experts 812 use analyzer creation utilities 802 to create the various documents that define the analyzers 110 and store them in the analyzer database 804. These and other content experts 812 also generate the rules that govern report generation as well as the templates for the reports, and they store them in the database 204 within the report generator server 202. Each analyzer 110 focuses upon a particular issue or set of related issues that can arise within the nodes 302, etc. in the enterprise 300. Each analyzer is designed to generate an XML report whenever an issue is found to be present and in need of consideration by management. When all of the issue information for a particular set of nodes are present within the (XML) issues database 112, then all of this information may be analyzed by a higher-level set of templates and rules 118 that are stored in the database 204 and that control the generation of high level reports 208 summarizing the condition of the enterprise 300 in ways that the management of the enterprise can understand.
Once the [0126] analyzers 110 are created and installed and the report templates and rules 116 are put in place, the system may then be called upon to do an assessment of the enterprise 300. An auditor 814, who may be an engineer or some other person desirous of learning about the condition of the nodes 302, etc. in the enterprise 300, requests an audit by using a task definition system 810 to create an assessment task. At 814, an assessment task A is shown. The assessment task 814 includes, in its definition, a list of the enterprises that are to be analyzed, a list of the nodes at each enterprise which are to be subjected to analysis, and a list of the analysis that is to be performed in the form of the actual names of the analyzers which are to be executed. The framework, based on the assessment type and its list of sub-domains, generates the list of analyzers automatically. In addition, the assessment task 814 includes a list of the reports that are to be generated following the analysis. Report generation may be done at the time of the analysis, or the reports may be generated at a later time in a separate session.
Once a [0127] task 814 is defined and initiated, the list of enterprises, nodes, and analyzers are passed to the analyzer harness 806. The analyzer harness 806 then proceeds by picking up the analyzers 110 from the database 804, one at a time, and with each analyzer 110 the analyzer harness 806 proceeds through the nodes 302, etc. one at a time. For each node, the harness 806 creates a framework linking the analyzer 110 to configuration information files that are retrieved from the tracker database 106. Using this framework, the harness 806 wraps the analyzer 110 in this environment and causes it to be executed in the context of the list of configuration information files that contain configuration information gathered from the node 302 that is being currently analyzed. During its execution, the analyzer 110 calls upon special subroutines that generate short XML reports of any issue which warrants management attention and also of any error condition which may arise. After the analyzer 110 terminates, the analyzer harness 806 takes these small issue XML reports and expands them, using issue text templates retrieved from the analyzer database 804 and also information as to the identity of the node and the identity of the assessment task, and creates an expanded XML report which is stored in the (XML) issues database 112 after the analysis have been run against all of the nodes 302, etc. In this manner, an extended issue report is generated in an XML format that is both human readable and also that lends itself to being incorporated into a database for automated retrieval and manipulation.
The list of reports from the [0128] task definition 814 is passed to the report generator 206. The report generator 206 also has access to the report templates and rules database 204 and to the XML issue report which can be retrieved from the (XML) issues database 112. Using all of these materials, an expert system engine within, or supplementing, the report generator 206 evaluates the rules and, under their guidance, examines the issue information, generating high-level conclusions for management concerning the general state of the enterprise. Then, using the report templates, the report generator 206 prepares a variety of reports, as has been explained, setting forth the status of the enterprise 300 and its nodes 302, etc. These are then fed to various recipients of the reports 817.
FIG. 9 illustrates the contents of the [0129] tracker database 106. In this embodiment, a tracker database 106 is established for the United States, another for Europe, and another for Asia. These tracker databases 106 are connected by conventional networking to the various enterprises 300 to which they relate so that they can collect information 24 hours a day and thus remain current. As illustrated in the figure, the tracker database 106 information is organized first by enterprise, within the enterprise by node, within the node by the collector, and within collector by the output files each collector has generated, each time and date stamped such that there may be a historical collection of configuration information. With this organization, the analyzer harness 806 is able to request the information file generated by any collector 104 when running on any node 302, etc. within any enterprise 300.
FIG. 10 presents a flow chart of the process by which an [0130] auditor 814 defines the assessment task that is to be carried out by the analysis server 800 during any given processing operation.
The [0131] auditor 814 begins by specifying the type of assessment that is to be carried out. At 1002, a first type of assessment is the “full audit request.” The full audit request looks at all aspects of the enterprise 300, right down to the smallest details. It also generates high level reports for management and for others, drawing general conclusions about the state of the enterprise, the nature of the problems found, and the possible solutions that may be considered. It is the full audit request or full audit of portions of an enterprise that takes full advantage of the present invention's ability not only to collect lots of configuration information using collectors 104 about an enterprise 300 and to analyze that information using multiple analyzers 110, but also to have rules associated with a report generator 206 that can look over the issues developed by the analyzer 110, scan the issue information looking for statistical indications and trends, and also compare present and past issue information to see trends over time, generating meaningful reports about where the enterprise 300 is headed and what can be done to prevent its failure and to improve its flow operations and reliability. Once this information is known, communicated, and polished, it is possible to move into the realm of self-healing systems. In this realm, an issue is detected automatically, the recommended action is known, and the issue is fixed before it becomes a problem. The customer is notified of the fix rather than the issue.
A second type of assessment is called simply an “assessment.” An assessment involves the use of large numbers of [0132] collectors 104 whose output is processed by a fairly large number of analyzers 110. A detailed report of all the issues identified is then generated.
A third type of assessment is the health check, which states in fairly summary terms whether the nodes within an enterprise are OK or whether there are problems that need to be addressed. This a fairly simple “go” or “no go” type of evaluation that looks primarily at critical values only and that generates a minimal report. [0133]
The [0134] auditor 814 may select any one of several different types of coverage for each of these assessments. An overall assessment runs all the analyzers 110 and generates a large amount of issue information about all aspects of an enterprise 300. A hardware and operating system assessment runs only those analyzers 110 that focus upon the hardware and operating system elements, as opposed to their programs, file systems, networks, etc. A networking assessment focuses upon the networking aspects of an enterprise 300, including such things as network load, network slow-down, and the conditions of the various network routers and switches. A file management and storage assessment focuses upon the condition of the disk drives, the availability of space within the file system, and other such issues. The assessment hierarchy available for task definition can match the assessment XML hierarchy of XML control information for the individual analyzers, as is illustrated in FIG. 25 and as is explained at a later point.
The next step in the task definition process is [0135] step 1004, selecting a tentative list of analyzers consistent with the assessment type selected. The tentative list of analyzers 110 is selected automatically by the task definition system 810 and is displayed with check boxes beside each analyzer name; and then the auditor is permitted to view and to modify the list of the analyzers that will actually be processed by possibly adding some, and by possibly eliminating some that are not required or essential to a given audit. Only the analyzers with their check boxes checked are actually executed.
Next, the location of the assessment is designated. At [0136] step 1006, a region (United States, Europe, etc.) is selected, or possibly a legacy database 212 source of information is selected if real-time monitoring of the enterprise 300 to be audited is not presently implemented. At step 1008, one or more enterprises 300 within the selected region are selected for auditing. At step 1010, a listing of the enterprise 300 nodes 302, etc. is automatically generated and displayed, again with check boxes, such that the auditor may go in and remove some nodes from the audit, and add other nodes to the audit. Next, at step 1012 in FIG. 11, the auditors enter information identifying themselves and also comments, such as a general statement of why the audit is being conducted. For example, if the customer has been experiencing certain difficulties, the auditor may wish to summarize this at this point so that the reason for the audit appears in the final report.
At [0137] step 1014, the nature of the report or reports desired is selected. The auditor may only report a management summary report, or he or she may request a technical summary report. On the other hand, the auditor may wish to see all of the technical details. A special marketing report, indicating areas where the enterprise may be persuaded to purchase additional software or hardware, can also be generated. A detailed manufacturer's report would report on the update level and versions of all the software and hardware and might contain information concerning failures that might be useful for future maintenance of the enterprise, or that might suggest the need for improvements in system component design.
A format for the report is also selected. The text report can be generated in the form of a word processing document suitable for printing in color. High-level reports might be generated in presentation format for projection and display to management. An HTML report might link to underlying explanatory details from a general report. Additionally, the [0138] auditor 814 may wish to request that statistical information be presented in tabular, graphic, or chart form to make it easier to understand.
The report format will also vary depending upon the target of the report. Thus, the above report formats have assumed that humans are to receive the report. However, if the report is destined for a database, it would, of courses, be generated as an XML document suitable for simple database entry or possibly be fed directly into an SQL or other type of database. If a work flow system exists that can generate work orders for service personnel, it may be desirable to generate a machine-readable file containing information from the report that can be directly transferred into the work flow system to directly result in the scheduling of maintenance or repair. This can be done in particular for critical issues that need immediate attention. Alternatively, if a healer system facility is available that can go out and reconfigure or load new software onto nodes, a machine-readable version of a report may be generated in a format that is acceptable as input to the healer system so that such repairs can be commenced immediately. This is desirable to achieve fail-safe operation of the [0139] nodes 302, etc. not just in mission-critical applications, but in most or all applications. The present invention thus can be used as a crucial feedback link in the realization of truly self-healing systems.
Finally, at [0140] step 1016, a delivery mechanism is selected. The reports may be printed, of course, or they may be e-mailed to a recipient list that needs to be supplied. In the case of machine readable reports destined for work flow system or healer system or databases, the Internet address of the recipient computers must be provided.
Creating New Analyzers [0141]
The analyzer creation process is shown in flowchart form in FIGS. 12 and 13, and FIG. 14 presents a list of user interface facilities which can aid in the process of creating new analyzers. In addition, Appendices A, B, C and D present simple examples of the various components of an analyzer. [0142]
As a simple example of how an analyzer can be created, the “disk usage” [0143] collector number 100024, whose creation was illustrated at 500 in FIG. 5 and whose storage in the collector database 310 is illustrated at 408 in FIG. 4, can be used as a source of input information for an analyzer 110 created to illustrate the process.
Recall that the “disk usage” [0144] collector 104, when executed upon a node 302, generates the simple textual report that is illustrated in Appendix E. Simply stated, this report lists, for every file system volume attached to a given node, the name of the volume and, among other things, the percentage of the volume's file storage area that is filled with files. Appendices A-D illustrate the details of an analyzer that is assigned to look at the report shown in Appendix E, to extract records concerning all volumes whose occupancy level exceeds the 90 or 95 percent or that are full, and which are therefore in danger of overflow. This analyzer 110 will test the numeric information on each line of the report shown in Appendix E, locate the percentage figure, compare it to 90 and 95 and 100 percent, and then generate a issue report if and only if the amount of storage occupied exceeds the 90 or 95 or 100 percent limit. As will be explained, this issue report is generated by the analyzer within a framework environment where the analyzer does not really “know” which node 302, etc. it is evaluating. Accordingly, the report is then embedded into a larger XML report generated by the analyzer harness 806 that identifies the node and that also includes textual messages explaining the nature of the issue which has arisen.
Referring now to FIG. 12, the analyzer creation process preferably begins with the [0145] content expert 812 using a word processor to prepare a textual description of the analyzer 110, explaining what it is to do in human-readable terms. This textual report appears in Appendix A for the exemplary analyzer that is designed to accept and to analyze the report shown in Appendix E provided by the collector “bdf-il” which evaluates the usage of volume storage space.
The expert content that is put into the analyzers covers formal and systematic knowledge that can be easily communicated and codified and is readily available (explicit knowledge) and personal information that is based on experience and learning and is often unrecorded and unarticulated (tacit knowledge) and is more difficult to formalize. [0146]
Next, [0147] step 1204, the analyzer creation utilities 802 gather information identifying the content expert 812 of this analyzer 110 as well as creation revision dates and other background information.
At [0148] step 1206, the designer 812 is presented with a list of all the collectors, organized by managed element class, and is invited to select which collectors will provide input information to this analyzer. In general, a single analyzer may accept input information from one or any number of collectors 104 and may then process that information in any number of complex ways. Dozens of collector-generated node configuration files may be fed into an analyzer that may be a huge computer program written in “c” or “c++” and compiled from numerous “c” or “c++” program files under the control of a “make” file. On the other hand, an analyzer may be a very simple program written in “java” or “perl.” To keep the example being developed here simple, the analyzer will collect information only from the collector identified as “disk usage” and assigned the identification number “100024”.
Next, at [0149] step 1208, the designer creates one or more analyzer issue templates, which may contain placeholders for the insertion of variable information. Such a template is illustrated in Appendix D, where three templates are preceded by template numbers. These templates contain, within their text, delimiters marking where variable information may be inserted. In this case, the variable information identifies the file system directory that has given rise to an issue.
Next, at [0150] step 1210, the source code (with comments) for the analyzer is created. A typical source code file is illustrated in Appendix B. This source code file is either compiled or interpreted or otherwise executed after having been harnessed and wrapped in a framework that provides it with the proper input information to do its analysis of a particular node. The source code file does not know the identity of the node whose configuration information it is analyzing. It simply knows that a list of file name arguments has been passed in and, from their order, it knows from which collectors those files originated. In this simple example, the argument is a single file name, the name of a file that contains the report which appears in Appendix E and which contains information defining the identity of the attached file systems and how close their volumes are to overflowing, as has been explained. In a more complex case, a series of source files developed by a series of different collectors would be passed as arguments into the analyzer code for execution. The analyzer code is written independently of any particular node, in most cases, so that the same analyzer may be used repeatedly to analyze hundreds of different nodes having thousands of different configurations.
In this embodiment of the invention, the [0151] analyzer harness 806 is written in “java,” and accordingly, “java” may be used as one of the languages in which an analyzer may be written. Byte code compiled versions of such “java” analyzers are simply loaded and executed directly by the analyzer harness 806 and are passed the list of file names as an input.
In some cases, it may be desirable to write an analyzer in the “c” or “C++” programming languages, or it may be desirable to adapt an existing program written in “C” or “C++” that is already being used for field testing to serve as an analyzer. Such executable files are compiled and linked under make file control in the normal manner and are placed into operation by the [0152] analyzer harness 806 which calls a conventional operating system “shell” utility to launch such executable analyzers. Such “C” or “C++” programs may be highly complex and may include a source module as well as additional code modules and a make utility for controlling compiling and linking. Provisions in the analyzer creation utilities 802 are made to include all of these analyzer file components in the analyzer database 804 and to re-compile them as needed for execution whenever changes are made to the source code, just as in a standard program development environment.
In other cases, the analyzer may be written in an interpretable language such as perl or “kshell” rather than in “java” or “C” or “C++”. In these cases, the [0153] analyzer harness 806 launches an interpreter program and passes to it the argument list containing the names of the files containing the collector configuration information that is to be processed.
As [0154] step 1210 indicates, the present invention accepts analyzers written in “java,” “C” or “C++,” “perl,” or “kshell” and stores them in the analyzer database 804 complied (if necessary) and ready to be executed.
With reference to FIG. 13, [0155] step 1212 calls upon the designer 812 of an analyzer 110 to specify which nodes 302 it is to analyze, if necessary. The default is that an analyzer 110 processes all nodes 302, and then the analyzer 110 is called upon by the analyzer harness 806 to process files generated by the collectors 104 for all of the nodes 302, 304, 306, (etc.) in sequence, with the analyzer 110 being executed one time for each node with different input file argument lists provided each time the analyzer 110 runs.
In some cases, a given analyzer may operate on a subset of managed elements, which may be specified as a collection of nodes included in a list of managed elements. Alternatively, in a system where nodes are clustered together to back each other up and to provide alternate platforms for executing programs when one or another of the nodes in a cluster is out of service, it may be that a given analyzer processes input data received from all nodes in a cluster simultaneously. [0156]
The [0157] content expert 812 having completed the process of defining the analyzer 110, it now remains to store the information defining an analyzer 110 in the analyzer database 804. The template text (Appendix D) and the source code (Appendix B) along with the textual description (Appendix A) of the analyzer are simply stored as separate files in the analyzer database 804 and are linked to the name of the analyzer in a directory. There remains the other information provided by the user defining the conditions under which the analyzer is to run, the names of the files containing the templates and source code, the author of the code, the particular language in which the template is written, and other things. This information, in the form of an XML analyzer descriptor file, is also saved in the analyzer database at step 1214. An exemplary XML file appears in Appendix C.
With reference to Appendix C, this XML file portion of the analyzer definition includes the following: [0158]
A standard XML header; [0159]
The analyzer name and version number; [0160]
The author's e-mail address; [0161]
A one line description of the analyzer; [0162]
A link or pointer to the ASCII template that defines the issues (see Appendix D); [0163]
A link or pointer to the analyzer code file; [0164]
The particular programming language (“kshell”, “java”, “C”, “perl”) used in writing the analyzer; [0165]
The type of analyzer source—text, binary (“C” or “C++”), or byte code (“java”); [0166]
The “purpose” of this analyzer file—main, auxiliary, or make, and whether it is executable; [0167]
(Repetitions of the above four lines for each separate analyzer program file.) [0168]
The number of nodes that can be processed simultaneously; and [0169]
The sources of input in the form of a list of collector names or numbers. [0170]
As can be seen, the XML file set forth here defines completely how the [0171] analyzer harness 806 is to execute a particular analyzer after it wraps in a frame and sets it to examining the information relating to a particular node or series of nodes. This XML file may be thought of as a control program for the analyzer harness 806 that governs the execution of the corresponding analyzer 110.
Finally, at [0172] step 1216, all of the information just described is saved in the analyzer database 804 as part of the analyzer 110 definition.
This embodiment of the invention provides a number of optional but useful tools to the analyzer designer in the form of user interfaces and utilities. These are listed in FIG. 14: [0173]
A general utility facilitates the creation of new analyzers and the modification or deletion of existing analyzers at [0174] 1402 and is similar to such tools provided in any program development environment.
When one is done entering the information concerning an analyzer, an XML [0175] descriptor creation utility 1404 generates the XML file, such as that illustrated in Appendix C, automatically.
Managers are provided at [0176] 1408 for program source code, auxiliary source code, and make files that are similar to those to be found in any general purpose program development environment. Additionally, a manager is provided to assist in creating template text files, such as that shown in Appendix D, which includes provision for embedding variable space holders into such a text file. For example, an analyzer might pull out of a collector document a piece of information and then arrange for the information to be inserted at the variable insertion point within such a text-file template, so that the template contains node-specific information in addition to background boilerplate information defining a given issue.
To assist in managing the many collectors that may be available, a [0177] collector search facility 1410 is provided that facilitates searching through collectors both by the Unix or other command assigned to a collector (for example, “bdf”) and also by the logical names and numbers assigned to a collector.
To facilitate the testing of analyzers, at [0178] 1412 there is provided a simple analyzer development environment which encapsulates an analyzer in a framework and provides it with live sample information obtained from editable stored samples of collector output files input so that a prototype analyzer may be tested in a safe environment to see how it performs.
A particularly useful tool is one that facilitates the creation of the lists of collector output configuration information files that are fed into the analyzers as their information input and that helps the designer to arrange the names of those collector output files in the proper order expected by the analyzers. This [0179] facility 1414 presents, in tabular form, a vertical list of analyzers on the left and a vertical list of collector reports on the right, and it permits one, by clicking the mouse, to connect each analyzer to one or more collector reports, and to indicate by the ordering of those connections the order in which collector reports are to be fed into each analyzer as arguments that are passed to an analyzer each time an analyzer is executed.
Another useful utility simply provides the designer with views of sample collector reports at [0180] 1416. Since the analyzers must go into those reports and extract text from the middle of the reports using programming commands, it is essential that the designers of the analyzers have access to sample reports so they can see how the text is positioned and then come up with the proper text extraction code for cutting excerpts out of the reports for analysis and for possible inclusion in the issue statements. In the future, the framework will parse the information and the analyzers will just receive the piece of information they need to do their analysis.
Finally, at [0181] step 1418, a manager is provided that groups analyzers into logical groups and thereby permits analyzers to be assigned to a particular assessment type. This tool facilitates the arrangement of analyzers such that when an auditor 813 performs task definition using the system 810, the auditor 813 may make an intelligent decision at step 1002 (FIG. 10) as to what task is to be performed and come up with an appropriate set of analyzers, which may be modified at step 1004 as has previously been explained.
Creating Template Reports [0182]
FIGS. [0183] 16-18 illustrate in flow chart manner the process of creating template reports and also the rules that govern the assembly of templates and that perform a high-level analysis of the issues identified by the individual analyzers. The ability of the system to generate readable, meaningful, summary management reports largely depends upon this portion of the system.
In this embodiment of the invention, a program called “Crystal Reports” obtainable from Krystal Decisions, Palo Alto, Calif. is used to transform template documents into finished reports under database control. Crystal Reports includes analysis tools that can examine trends, expose relationships, and zero in on important facts. It thus combines rules and report templates into a single “batch” oriented report generator that can be driven by the issues database transferred from XML file form into a Microsoft Access database system. Other rule and template based report generators can also be used, such as “Caps Author” and “Hot docs” (from Capsoft Development Corporation, Provo, Utah). [0184]
The setting up of these systems involves two aspects. First of all, the incoming information is presumed to be in some form of a database. In this embodiment of the invention, the database is Microsoft's Access database program. The XML issues file is simply read and its contents are fed directly into Access. [0185]
The template documents, designed in accordance with the report generator specifications, contain large amounts of boilerplate or background text. They may also contain embedded commands analogous to the programming language commands “if”, “else”, and “end if” which are used in programs to select which sections of code are to be executed or to be skipped over or performed in the alternative. In the case of document templates, these same types of commands determine which text ends up in a report and which boilerplate text is cut out and left on the cutting room floor. [0186]
Such report generators also provide for the generation of repetitive passages through the use of commands analogous to the “repeat” and “end repeat” language commands found in programming languages. Finally, templates may contain various types of placeholders for expressions which are evaluated and then inserted into the boilerplate to become part of the finished report. [0187]
In addition, most such systems have the facility to provide for separate rules which compute not only values to be inserted into the reports but also values that control the inclusion or exclusion of boilerplate text as well as the inclusion of repetitive insertions. These rule systems can be quite elaborate, constituting “backward chaining” expert systems which may include hundreds of rules and which may perform very detailed and elaborate analyses upon the incoming analyzer information which is presumed to be stored in a Microsoft Access database (any other database system on any platform could be used here). A final provision of these systems provides for the insertion of tables into the text of a report, where the tables are extracted from the Access database. For example, such a detailed technical issue report might include a table that summarizes each and every one of the issue reports generated through analysis, as is illustrated at the end of Appendix G. [0188]
At the beginning of the process of report template creation, the first step at [0189] 1602 is to make available to the author of the report the symbolic names of the analyzer output information that is to be used both embedded into the report templates and also to control report template boilerplate selection. These information values necessarily come out of the analysis system. With reference to Appendix F, Part 2 of Appendix F presents the XML information output that comes from issues detected through the execution of analysis programs on configuration information gathered by collectors. This information structure is transferred into a Microsoft Access database from which the values can be retrieved using identifiers the same as or similar to those which appear in the XML information structure. Accordingly, the author needs to be provided with sample issue reports of the kind shown in Part 2 of Appendix F for each and every one of the available analyzers and for each and every one of the issues that each such analyzer is capable of reporting. This information enables the template and rule author to come up with a set of rules which can analyze and summarize the issues database 112 information as well as pick up detailed information from there and incorporate it into any arbitrary type of report that may be necessary or desirable.
Using this information concerning the names of variables, at [0190] step 1604 the author develops a set of rules that can review the overall performance of an enterprise 300. The author, clearly, must be an experienced enterprise technical expert who knows what to look for in the issue information generated by a large computer enterprise installation and who knows how to write rules that will perform the same types of analyses that the expert himself or herself would have performed in person had they been able to go to the enterprise site.
Having developed rules to control report generation, at [0191] step 1606, the expert then moves on to develop a management summary report template that states, in a readable form, the general state of the system. This report ideally should be readable by lay individuals, such as the chief operating officers of the company that owns and manages the enterprise. It should point out general problems, and it should then give examples of why these problems merit the attention of management. It may then provide general overview suggestions as to how those problems should be managed. Built into this report is clearly the experience of an expert in enterprise system design and management. It may present cost estimates for configuration changes that can improve enterprise performance.
For the technical manager, the expert develops a different template that lists all the critical technical points in summary form, states their degree of criticality, and then presents brief explanations and costs estimates for their correction. This report will not include the background explanations needed by the lay managers. It will be much more concise and to the point, and it will also include a level of technical detail that is omitted from the previous report. [0192]
At [0193] step 1610, an additional template can be developed which produces a completely detailed presentation of all the specific issues for each of the nodes or groups of nodes that were analyzed. In addition to the statement of issues obtained directly from the analyzers, this report may include reasons why certain issues need to be addressed, costs of implementing changes, and possibly a detailed explanation of these issues, which may be through a hypertext linkage to other documents. Snippets of the actual collector configuration information may be included, as needed, to clarify an issue. As is illustrated in both Parts 1 and 2 of Appendix F, such snippets can be readily incorporated into the XML output of an analyzer and may then be passed out and included in the technical report to assist the technical experts in understanding any of the issues not addressed by the report generation rules.
Finally, at [0194] step 1612, it is desirable to develop a variety of master report templates intended for differing audiences and business needs that include varying combinations of the reports just described, as well as possible additional introductory and explanatory reports hyperlinked to background information. Thus, a single report may contain multiple parts destined for different individuals and may be printed and delivered as a unit to the owners and managers of an enterprise.
The report generation process, triggered by the [0195] auditor 813, is carried out by the report generator 206 which, in this embodiment of the invention, is the program “Crystal Reports”. Referring now to FIGS. 19 and 20, the report generator 206 receives a list of reports from the assessment task 814 and begins by calling forth from the report template and rules database 204 the templates and rules needed for generating the desired reports, at step 1902. Next, at step 1904, the report generator 206 goes to the (XML) issues database 112 and obtains the necessary XML issues information that was generated by the analyzers 110, and it transfers this information content into a Microsoft Access database for convenient access and retrieval by the report generator 206. The report generator 206 also obtains from the task definition system 810 additional information about this audit, its author, and so on and transfers this information into the Access database for convenient retrieval by the report generator 206. Next, at step 1908, the report generator 206, again at the direction of the task definition system 810, may obtain directly from the tracker database 106 background information on the specific enterprise and nodes being audited for inclusion in the reports, placing this information into the Access database along with the rest. In addition, the list of enterprises, the list of nodes and the list of analyzers in the assessment task 814 can be added to the Access database and included in a report at an appropriate place.
Finally, at [0196] step 1910, the Crystal Reports program operates in batch mode as a template and rule processor controlled by the templates and rules and by the content of the Access database to generate all the necessary and required reports. This could also be a non-paper-based report such as a workflow management system.
The Analyzer Harness [0197]
The analyzer harness is described from three different perspectives. First, FIG. 15 presents, in program flowchart form, the steps executed by the [0198] analyzer harness 806 and the analyzers 110 to illustrate in general terms how the individual analyzers are called upon to process information provided by the individual nodes. (FIG. 15 does not address the task of describing the complex interactions that occur between the individual analyzers and analyzer harness 806.) Secondly, FIG. 21 presents a software element block diagram that illustrates the precise relationship which exists between the three subcomponents 2102, 2104, 2106 of the analyzer harness 806 and the various types of analyzers 2110, 2112, and 2114 supported in the present embodiment of the invention. Finally, FIGS. 20-24 present a flow diagram of the steps that are carried out by the analyzer harness 806 components illustrated in FIG. 21, addressing the issue not addressed by FIG. 15.
Referring now to FIG. 15, the information processing operations carried out by the [0199] analyzer harness 806 are shown in a conventional software flow diagram. The process begins at 1502 when the task definition system 810 passes to the analyzer harness 806 the assessment task 814 that lists the enterprises, nodes, and analyses which are to be performed. This enables the analyzer harness 806 to set up a triple set of nested loops, as shown, branching back at step 1522 after each node is processed, after each analyzer is processed against every node, and after each enterprise is processed.
The overall process begins at [0200] step 1504 when a first enterprise 300 is selected. Then at step 1506, a first analyzer 110 is selected. Next, at step 1508, the definitions of the analyzer and information concerning how it is to be executed (see Appendices B and C) are retrieved from the analyzer database 804 along with the template (Appendix D) containing issue messages for the analyzer. Next, at step 1510, the names of the collectors designated by that analyzer in the analyzer's XML descriptor (Appendix C) are collected as an indication of where the input information is to come from. At step 1512, if the analyzer is computing a trend over time, then a designated set of times and dates or range of times and dates are also picked up.
Program control then commences at the [0201] step 1514, and the steps 1514 through 1522 are repeated for each node 302, etc. that is to be analyzed by this particular analyzer.
At [0202] step 1516, the names of the designated collector report files for this particular node are retrieved from the tracker database 106 and are listed as input arguments to the analyzer's execution. Then on step 1518, the analyzer code is actually executed in the context of the collector reports provided in this input argument list. At step 1520, if any issues or failures arise, the analyzer generates small XML reports and returns them to the analyzer harness 806 which augments them with information such as the name of the node and the template issue text (See Appendix D). Looping then proceeds from step 1522 until all of the nodes are processed by this analyzer.
Next, the [0203] analyzer harness 806 moves on to the next analyzer and repeats steps 1506 through 1522 until all of the analyzers have been processed. It moves on to the next enterprise and repeats the steps 1504 through 1522 until all of the enterprises have been processed, if there is more than one.
When processing is complete, at [0204] step 1524, the analyzer XML output report is stored as a file in the (XML) issues database 112.
Referring now to FIG. 21, a detailed illustration of the structure of the [0205] analyzer harness 806 is presented that illustrates its relationship to the analyzers and also illustrates how the various different types of analyzers are processed. This figure will be discussed in the context of the flowchart of analyzer harness operation presented in FIGS. 22-24.
In FIG. 21, the [0206] analyzer harness 806 is shown to have three major components. An analyzer loader 2102 is the primary component that manages the loading and execution of the analyzers during the processing of configuration information. This analyzer loader 2102 calls upon an analyzer descriptor 2104 to obtain from the analyzer database 804 the XML information for a particular analyzer 110 and to put that information into a database from which the analyzer loader can have easy access to it as it determines how to harness and encapsulate into a framework a particular analyzer 110 that is to be executed. The analyzer loader 2102 also calls upon an argument manager 2106 to reach out into the tracker database 106 to find collector reports 2108 and to arrange for those reports, or the file names for those reports, to be passed as incoming arguments to the analyzers 110 when they are executed by the analyzer loader 2102. The loader 2102, being written in “java”, executes “java” executable analyzers 2110 directly. It calls upon a standard operating system “shell” utility to execute analyzers 2112 written in the “C” or “C++” programming languages. It calls upon an operating system interpreter to interpret and execute analyzer written in “perl” or “kshell” or other interpretive languages 2114. The output of an analyzer is shown to be an XML output at 2124 which is generated by special XML output sub-routines 2116 that are called by the analyzer 110 regardless of its form to generate XML “issues” output messages and also error output messages. The output 2124 does not include identification of the relevant node, nor does it include textual information about an issue, since that information is not available to the analyzers 110 when they are executing. Accordingly, the analyzer loader 2102 accepts the XML output of the analyzers 2124 and generates an expanded XML report 2126 that include the issue defining language taken from the analyzer issue templates 2122 as well as the identification of the node being processed and also the name of the assessment task 814. This may be seen in Appendix F, where the XML output of an analyzer 110 is shown in Part 1, and the XML report 2126 of the analyzer loader 2102 is shown in Part 2 for the simple analyzer that has been used as an exemplary analyzer throughout this patent application.
Referring now to FIGS. [0207] 22-24, the process of the analyzer harness 806 executing an analyzer 110 will be described. At step 2202, the analyzer loader 2102 (FIG. 21) receives the name of an analyzer 110 that it is to load and process. This corresponds generally to the step 1506 shown in FIG. 15. The loader passes the name to an analyzer descriptor program 2104.
At [0208] step 2204, the analyzer descriptor program 2104 locates and reads in the corresponding XML database for the analyzer (see Appendix C) and thereby creates in its own internal information structure an analyzer descriptor object which is passed back to the analyzer loader 2102.
At [0209] step 2206, the loader 2102 next calls upon the argument manager 2106 to retrieve from the tracker database 106 the appropriate collector reports for the collectors designated in the descriptor object for a particular node (designated in step 1514 of FIG. 15). The collector reports 2108 are identified by file name, and these file names are organized in order as an input argument list for the analyzer 110 to read these files in order and process them. The file names are passed back to the loader 2102.
Finally, at [0210] Step 2208, the analyzer loader 2102 calls upon an appropriate entity to execute the analyzer 110 in accordance with the nature of the analyzer, passing the list of file names as an input argument.
Referring to the top of FIG. 23, the XML analyzer information which now resides within the descriptor object is now referred to in an effort to determine what kind of analyzer is at hand. At [0211] step 2210, if the analyzer is written in “java,” then it is simply called as a “java” subroutine, and it is passed the input list of file names as an input argument at 2212. If the analyzer is not written in “java”, then step 2214 checks to see if it is interpretable. If so, then an appropriate interpreter for “perl” or “kshell” is run at step 2218, and again the input argument list of file names is passed to the interpreter as input for the execution of the analyzer 110. Finally, if the analyzer is not interpretable but is executable, as in the case of a compiled “C” or “C++” program, then at 2216 an operating system “shell” utility is called upon to launch the executable analyzer program, again providing it with the list of file names as an input argument to the executing program.
Thus wrapped by the [0212] harness 806 in a framework that includes input collector configuration information files obtained from the appropriate node 302, the analyzer 110 performs its tests and thereby checks to see if any issue reports need to be generated. If so, or if any errors occur, the analyzer calls upon special routines 2116 that normalize the analyzer output as XML structures such as that shown in Part 1 of Appendix F.
The arguments passed through the [0213] output routine 2116 include:
The issue ID number; [0214]
Zero or more arguments for later insertion into [0215] analyzer issue templates 2122; and
An optional snippet taken from a collector report, or an explanatory text string, or a null string. [0216]
At [0217] step 2222, the issue list or failure list XML output (See Sample 1 in Appendix F) is captured and is returned to the analyzer loader 2102 in the form of an XML output 2124. At step 2222, the analyzer loader 2102 expands the size of this XML output by adding to it information defining the name of the node 302 from which the collector configuration information came, the name of the assessment task 814 launched by the auditor 813, and other such background information. The analyzer loader 2102 also retrieves the analyzer issue templates 2122 (Appendix D) for this analyzer from the analyzer database 804 and from this set of templates retrieves, by number, the particular text template that identifies and corresponds to the particular issue which triggered the XML output. If there are any variable space holders in that text template, the analyzer loader 2102 looks to the XML output 2124 to see what value is to be inserted at that point, and it inserts the value into the template, transforming it into an issue message. The analyzer loader 2102 then generates a new segment for a growing output XML report 2126 that includes all of the above added information added to the analyzer XML output 2124. A sample of such a report can be seen in Part 2 of Appendix F.
The [0218] analyzer loader 2102 then proceeds on to the next node at step 2224 and continues until all of the nodes have been analyzed by this particular analyzer 110. It then proceeds to the next analyzer 110 at step 2226 and thereby continues until all of the analyzers have been processed against all of the nodes. The XML report 2126 is then stored in the (XML) issues database 112 where it is made available to the report generator 206, as has been explained.
Hierarchical Organization of [0219] Analyzer XML Information 2118
An analyzer, as described above, consists of four parts: A textual description of the analyzer and its functions (See Appendix A), the source code of the analyzer [0220] 2120 (See Appendix B), the template 2122 text which defines the issues of an analyzer (See Appendix D), and an XML file that defines, for the analyzer harness 806, how the analyzer is to be processed and what information it requires.
In a different embodiment of the invention, the target of this application, which has been suggested above but has not been fully described, the [0221] analyzers 110 may be organized hierarchically to correspond to various broad and narrow arrangements or sets of managed elements within a given enterprise. For example, in FIG. 10, where the assessment tasks are defined, the assessment types are broken down into overall, hardware/operating system, networking, file management and storage, and so on. These assessments each correspond to a particular set or collection of managed elements. These sets can be still further subdivided. For example, as shown in FIG. 25 at 2502, the file system can be broken down into logical volumes, file system usage analysis, and logical volume management. Similarly, hardware may be broken down into memory, CPU, and other categories. Networking can be similarly subdivided.
To facilitate these types of assessment audits of the subparts of a large system, an XML control file can be created which includes, as its base or lowest elements, all of the links to the analyzer XML definitions of other sub-components of an audit contained in the [0222] analyzer database 804. This XML file might include, as its main heading, “Enterprise”, or some such general title; and it might then have links to the XML files for secondary headings such as “Operating System”, “File System”, and so on. Then each of these subheadings could be linked further to sub-subheadings, and so on such that many different possibilities for audits through various collections of managed elements of the enterprise can be contemplated. This XML file, indicated symbolically in Appendix G, then becomes a major tool for controlling the operations of the analyzer harness 806, causing it to do a particular auditing on a particular collection of managed elements. No changes are needed by the framework to handle these multiple combinations of assessments. The only changes that are needed involve the set of analyzer/heading definitions contained in the XML files which can be swapped in and out without affecting the framework.
The actual hierarchy would actually be created by links in the assessment type XML file to the Operating System, File System, . . . components which would in turn have links to their sub-components, until the link to the actual analyzer is reached—See Appendix H. [0223]
Accordingly, one achieves the conceptual assessment XML hierarchy shown at [0224] 2502 in FIG. 25 with little change to the current harness. In FIG. 26, the task definition system 814 is modified such that at step 2602 (which replaces step 1002 in FIG. 10) one merely specifies a desired assessment type in the outline structure defined by the large XML structure 2502 shown in FIG. 25. This task assessment type corresponds to a set of managed elements which correspond to a subpart of the XML hierarchy shown in FIG. 25 and which contains all the analyzers 110 for the set of managed elements. Note that the XML data 2118 (see Appendix C) for the “Disk Usage” analyzer appears beneath “File System”, “Logical Volumes”, and “File System Usage Analyses” in the exemplary XML structure shown symbolically at 2502 in FIG. 25. At step 2604 (which replaces step 1502 in FIG. 15), one simply passes to the analyzer harness 806 the lists of enterprises and nodes, along with the larger XML structure 2502 that includes, as a substructure, the list of the analyzers that are to perform in accordance with FIG. 25, including the portion of the XML structure in FIG. 25 which the auditor has designated for execution. The analyzer harness 806 then carries out these assessments automatically, covering as much or as little of the system hardware and software as the auditor desires to have covered in any given analysis operation, controlled entirely and continuously by the XML structure of FIG. 25 and by that portion of this structure which the auditor has designated for execution.
Description of the Analyzers [0225]
In addition to the “disk usage” analyzer which has been used as an exemplary analyzer throughout this specification, this embodiment of the invention includes a number of additional analyzers which are described below as examples of the many possible analyzers that can be created. Clearly, many other types of analyzers can and should be developed. [0226]
Single Points of Failure Analyzer [0227]
This analyzer checks to ensure that there are both a primary boot disk and an alternate boot disk. Further, this analyzer checks to ensure that the primary boot disk and the alternate boot disk are not installed on the same hardware path. Thus, an issue is flagged when there is a primary disk installed without an alternate disk. Also, an issue is flagged when the primary and alternate boot disks are installed on the same hardware path. [0228]
MC/ServiceGuard Timing Parameters Analyzer [0229]
This analyzer inspects the MC/ServiceGuard timing parameters for accuracy and supportability. The following issues are flagged: an issue is flagged when the value of NODE_TIMEOUT is set to the default value, which is typically two seconds; an issue is flagged when the NETWORK_POLLING parameter is configured to be greater than or equal to the NODE_TIMEOUT parameter; an issue is flagged when any of the values of HEARTBEAT_INTERVAL, NODE_TIMEOUT, AUTO_START_TIMEOUT, and NETWORK_POLLING_INTER-VAL is set below a predetermined value, which predetermined value is typically a time less than or equal to one second; and an issue is flagged when the value of the NODE_TIMEOUT is set to less than at least twice the value of the HEARTBEAT_INTERVAL parameter. [0230]
MC/ServiceGuard Release Analyzer [0231]
This analyzer checks the installed MC/ServiceGuard version on all nodes and ensures that the latest version is installed for a specific release of HP-UX (Hewlett-Packard, Inc.'s version of Unix) and that it is the same on all nodes in the cluster. An issue is flagged when MC/ServiceGuard does not appear to be installed, when there is a newer version of MC/ServiceGuard available for the operating system, or when any node in the cluster is running a different MC/ServiceGuard release than the others. [0232]
MC/ServiceGuard Package Configuration Analyzer [0233]
This analyzer checks the configuration of the packages within a MC/ServiceGuard cluster. An issue is flagged when the default package subdirectory does not exist, when the package switching is disabled, or when separate package control run and halt scripts exist for the package. [0234]
MC/ServiceGuard Daemon Checker [0235]
This analyzer checks to determine whether or not the MC/ServiceGuard daemon(s) are running. Using a daemon, a program can simply hand off data to a smaller program and go on to other work. Where the MC/ServiceGuard daemon ‘cmcld’ is not running on the node in the cluster system, an issue will be flagged. [0236]
MC/ServiceGuard Cluster Status [0237]
This analyzer checks the MC/ServiceGuard cluster status, node status, and package status. An issue will be flagged if: the cluster is down, the cluster is running on a subset of the nodes, one or more packages of the cluster are not running, global package switching is disabled, a package has node switching disabled for one or more nodes, a package is not running on its primary node or when no alternative node is available to which the package can be switched. [0238]
MC/ServiceGuard Cluster Configuration Analyzer [0239]
This analyzer checks the configuration of the MC/SwitchGuard cluster for correct configuration. An issue will be flagged when: the ASCII (American Standard Code for Information Interchange) configuration file is not found at the default location and/or with the default name, when the ASCII configuration file does not correspond to the binary cluster configuration file, when the ASCII cluster configuration file is not the same on all nodes in the cluster, when the binary cluster configuration does not exist on one or more nodes in the cluster, when the binary cluster configuration file is not the same on all nodes in the cluster, or when incorrect ownership and/or permissions exist on the binary cluster configuration file. [0240]
Root Mirror Analyzer [0241]
This analyzer verifies that the root volume is mirrored at least once, ensures that each mirror is bootable, ensures that an entry exists in /stand/bootconf for each mirror, and checks for single point of failure related to the boot disks. An issue is flagged when: there is no mirror found for the root logical volume, the boot mirror device is missing from /stand/bootconf, there is a non-bootable mirror of the root drive, there is a primary boot disk installed without an alternate disk, or the primary and alternate boot disks are installed on the same hardware path. [0242]
Processor Analyzer [0243]
This analyzer extracts information from the CPU (Central Processing Unit) hardware logs of each processor installed in the system. CPU specific information gathered includes the number of processors installed, PDC (Processor Dependent Code) revision level, the hardware model number (translated into English for the report), the processor chip revision, and the slot number in which the processor(s) are installed. In addition, the analyzer gathers HPMC (High Priority Machine Check) error information specific to the CPU, Memory, and I/O (Input/Output) error logs. Also, the analyzer identifies that the processor fast boot status flag has been set, which indicates that the memory self-test has been disabled during the system self-test portion of the PDC boot code. Processor chip revision is gathered, and may be used to proactively identify processors associated with class problems and/or known defects. An issue is flagged when: the processor PDC revision level is not the latest supported revision, a valid HPMC exists in the hardware logs, a valid memory error exists in the hardware logs, a valid fatal I/O error exists in the hardware logs, or if the fast boot flag has been set. [0244]
Memory Analyzer [0245]
This analyzer checks for the existence of single and double bit errors in the hardware logs, and also verifies the number of available entries in the page allocation table (PDT). An issue is typically flagged if: a single-bit memory error exists and its count is greater than or equal to 1000, a double-bit memory error exists and its count is greater than or equal to 1, or the number of PDT entries used exceeds 50% of the PDT total size, which by default is set to a maximum entry value of 50. [0246]
LVMGENERAL [0247]
This analyzer does basic tests on the LVM (Logical Volume Management) configuration such as comparing the content of /etc/lvmtab with the output of the lvm (logical volume management) commands in order to find volume groups, or physical volumes not activated or incoherencies. Also, it may run basic checks for each volume group, physical volume or logical volume. An issue will be flagged if: not all the volume groups listed in lvmtab (logical volume management tab) are activated, not all the physical volumes in a VG (Volume Groups) are currently active, a physical volume is missing in /etc/lvmtab, an activated VG is not defined in /etc/lvmtab, not all the volume groups listed in lvmtab are activated, or bad block relocation is not set to NONE for LV (Logical Volume) on EMC (Electric Machine Control, Inc.) drives. [0248]
LVLNBOOT Analyzer [0249]
This analyzer verifies the data reported by lvlnboot. “lvlnboot-v” displays the root, boot, swap and dump lvm configuration of an LVM enabled boot device. An issue is flagged if no dump device(s) has been configured. The analyzer may, where desired, verify Boot, Root, and Swap also. [0250]
KERNELINPLACE [0251]
This analyzer verifies that standard and backup kernel files are in place. An issue is flagged when: a standard kernel file is not in place, a standard backup kernel file is not in place, or a standard kernel could not be identified as a kernel. [0252]
KERNELCOMPARE [0253]
This analyzer compares and displays differences between kernel parameter settings for a list of hosts. An issue is flagged if some kernel parameters have different values. [0254]
General Errors Analyzer [0255]
This analyzer analyzes general errors and I/O error entries in the command line Support Tool Manager, i.e.: CSTM-logfile. (CSTM is an online support tool platform which operates under the control of the hp-ux Operating System. It provides automatic configuration mapping, a set of device verifiers, exercisers, information modules, expert tools, utilities, and firmware update tools for the system components and peripheral devices that are installed on hp-ux systems. The I/O error entries are part of the CSTM information module. The I/O error entries log specific device performance information for devices and system components which are exhibiting abnormal operational behavior.) The issue will typically be flagged if: the number of uncorrectable errors logged for a DLT or DDS tape drive device is greater than or equal to one, it has been less than 30 days since the data and time of the last entry in the formatted raw logfile (if not flagged, the analyzer is aborted), the number of overtemp entries is greater than or equal to one, the number of I/O error entries for the system components or peripheral devices is greater than or equal to ten, or the number of LPMC (Low Priority Machine Check) entries is greater than or equal to ten. In PA-RISC system architecture, there are recoverable and non-recoverable hardware errors. Detection of an HPMC (High Priority Machine Check) by PDC (Processor Dependent Code) is a fatal error which immediately stops the further execution of instructions by the CPU hardware. An LPMC is a recoverable error in the Main Memory of CACHE components of the CPU. When an LPMC occurs, a combined interaction between PDC and the hp-ux Operating System identify the type of recoverable error that occurred, triggers PDC code to execute a hardware retry to fetch critical data from the supplying hardware component and reinsert the date into the component at fault, to log the event that occurred, and then to re-initiate normal system operation. The LPMC is therefore a mechanism which provides a “hardware retry” of certain processor and memory errors which are not classified as fatal. [0256]
Currently, this analyzer is capable of analyzing, detecting, and reporting the presence of system hardware abnormalities and I/O errors. The functionality of this analyzer will be increased in future releases to include automatic analysis including criticality and threshold levels, of the detailed contents of the I/O error log contents as well. [0257]
FSTAB Analyzer [0258]
This analyzer checks the file systems in /etc/fstab for problems and checks whether they are correctly configured. An issue will be flagged if: a file system listed in /etc/fstab is not mounted, a file system is mounted that is not included in the /etc/fstab file, there is a conflict in the file system mount order, or a CDFS (CD file system) file system is mounted via /etc/fstab. [0259]
Firmware Analyzer [0260]
This analyzer verifies the firmware revision of all devices with write-able firmware that are installed in, and attached to, the system, including: internal and external SCSI (Small Computer Systems Interface) hard disks, disk array controller modules, internal array SCSI disk modules, and the status of fans, power supplies, batteries, DRAM (Dynamic Random Access Memory), and NVRAM (Non-Volatile Random Access Memory) on array controllers and array chassis. Also, where desired, this analyzer may be used to verify that each device is a supported Hewlett-Packard product. [0261]
The FIRMWARE ANALYZER typically flags an issue if: the product is not a supported Hewlett-Packard product, the firmware revision of the disk is not the current release, there are read, write, verify or non-medium errors logged on a SCSI disk and the number of errors is greater than or equal to one, if there is a bad status for an array fan, power supply, battery, DRAM, or NVRAM, or if the disk array is a Model 12 Hewlett Packard (“HP”) product. Unlike older models of HP arrays, the Model 12 can operate properly with internal SCSI disk modules that do not contain the current revision of device firmware. This is accomplished by the array controller firmware which is a separate hardware component from the internal disk modules. If the analyzer detects a Model 12, it performs only checks on the array controllers and aborts the firmware checks on internal disk modules. [0262]
File System Check Pass Numbers Analyzer [0263]
This analyzer checks that the file system pass number parameters are correctly configured and supportable. An issue is typically flagged if: the file system check pass number is set to zero, the file system check pass number for the root file system is set to zero, the file system check pass number for non-root file systems is set to one, the file system check pass number is missing, identical pass numbers for file systems are on the same disk drives, or file system check pass numbers are not optimized for parallel fsck (file system check). [0264]
DUMPFILESYSTEM [0265]
This analyzer checks that the file system used by savecrash is large enough. An issue is flagged when no dump device is configured or when the file system too small for a dump. The size of the file system that may be used to save the dump is checked, taking into account a compression of 75% if compression is enabled, and a kernel of 80 Mb. The calculation is made on the worse case based on total memory size. [0266]
DMESG Analyzer [0267]
This analyzer captures system error messages from dmesg (a UNIX command that displays the contents of a circular buffer that contains diagnostic information). Also, this analyzer verifies that there is support for capturing the dmseg output incrementally to ensure that none of the system buffer information is lost. If the support for capturing this information exists, the analyzer verifies that it is configured correctly. An issue is flagged if: there is no entry in the crontab file (a configuration file to schedule tasks in UNIX to run commands at specific times) to incrementally capture the dmesg output, if there is an entry in the crontab to capture the dmeswg, but it is not being written to the default location, there is an entry in the crontab to capture the dmesg, but the interval to collect it is greater than a predetermined recommended value, or there is an entry in the crontab to capture the dmesg output, but the execution of the dmesg command does not capture the incremental changes. [0268]
Diagnostics Configuration Analyzer [0269]
This analyzer checks whether or not online diagnostics have been installed and are running. An issue will be flagged if: the online diagnostics are not installed completely, one or both diagnostic daemons (Diagmond and Memlogd) are not running, or the version of diagnostic software is not the latest released revision. [0270]
CRASHCONF Analyzer [0271]
This analyzer is one of the required checks which, when combined with the other appropriate analyzers, will ensure that the system is configured correctly to perform a successful system dump. An issue will be flagged if: the system dump device(s) fail to have adequate space to hold the system dump, or the minimum necessary classes have not been enabled in fastdump. [0272]
CPU Chassis Analyzer [0273]
This analyzer checks the internal processor hardware logs and reports on the status of the CPU fans, power supplies, and CPU processor module de-configuration. The de-configuration state of a processor module is archived in the internal processor hardware logs. The possible states are “C” which indicates the processor module is configured and operating normally, and “D” which indicates the processor module has been de-configured by the PDC as a result of the occurrence of a PDC detected hardware error. There is an additional subset of the de-configuration status flags (RS) which monitors if the CPU processor module has been de-configured by the PDC as a result of a recoverable non-fatal hardware error, or if the CPU processor module has never been reconfigured by the PDC after physical replacement. If the RS flag is set, it indicates that the state of the processor is currently de-configured, but is pending reconfiguration and activation by physically restarting the system. Since it is possible for multiple CPU processor modules, fans, and power supplies to be installed in a CPU chassis, the physical location of each failed component is also reported. [0274]
An issue will be flagged if: the CPU processor modules that are installed have a “D” status bit set, if the CPU processor modules that are installed have both “D” and “RS” status bits set, if the “F” (failed) status bit for a cabinet fan failure is set, or if the “F” status bit for a cabinet power supply is set. The “D”, “RS”, and “F” status bits will only be set by the PDC if the CPU module, fan, or power supply is physically installed in the system processor chassis. [0275]
BOOTSCRIPT Analyzer [0276]
This analyzer verifies that the boot configuration scripts in /etc/rc.config.d have optimal configuration for an HA (High Availability) system. Where desired, this analyzer may only check the crashconf and savecrash configuration, but it is desirable to check all boot scripts. An issue will be flagged if: crashconf is not configured to run, savecrash is not configured to run, savecrash is not configured to run in the foreground, at least one dump device fails to be configured, the dump devices fail to be large enough for a full dump, the dump devices fail to be large enough for a dump, or the dump devices are unable to hold all physical memory. [0277]
The foregoing description of an embodiment of the invention has been presented for purposes of illustration and description. It is not intended to be exhaustive or to limit the invention to the embodiment disclosed, and modifications and variations are possible in light of the above teachings or may be acquired from practice of the invention. The embodiments were chosen to explain the principles of the invention and its practical application to enable one skilled in the art to utilize the invention in various embodiments and with various modifications as are suited to the particular use contemplated. It is intended that the scope of the invention be defined by the claims appended hereto, and their equivalents. [0278]

Claims

What is claimed is:

1. In a system that analyzes data gathered by collectors monitoring nodes, a hierarchical method of defining and organizing the analysis tasks to be performed comprising the steps of:

performing, for each low-level analysis task, analysis steps that comprise accepting data from one or more collector types, analyzing the data, and generating reports, guided by an analyzer descriptor written in XML or a comparable format identifying the collector types whose data the analyzer steps are to process;

for each higher-level analysis task, of which there is at least one, performing one or more sets of analysis steps for either low-level or high-level analysis tasks, guided by an analyzer descriptor written in XML or a comparable format that calls upon or that incorporates by reference other analyzer descriptors; and

guided by one or more analyzer descriptors and, in the case of higher level analyzer descriptors, by the analyzer descriptors they call upon or incorporate by reference directly or indirectly, repeatedly performing the set of analysis steps corresponding to one or more analyzer descriptors one or more times against collector data gathered from a different node or set of nodes during each such exercise of the set of analysis steps, and during this process, presenting for acceptance and analysis during the performance of each data acceptance step data gathered from collectors of the type indicated by the guiding analyzer descriptors.

2. A method in accordance with claim 1 wherein the step of obtaining an analyzer descriptor includes obtaining one or more such descriptors that contain an analyzer name, description, and in the case of at least one higher-level analyzer descriptor a list of the other analyzer names whose descriptors it calls upon or incorporates by reference.

3. A method in accordance with claim 1 wherein the step of obtaining an analyzer descriptor for each higher-level task includes obtaining at least one analyzer descriptor for a higher-level task that calls upon or that incorporates by reference at least one other analyzer descriptor for another higher-level task.

4. A method in accordance with claim 3 wherein the step of obtaining an analyzer descriptor includes obtaining one or more such descriptors that contain an analyzer name, description, and in the case of at least one higher-level analyzer descriptor a list of other analyzer names whose descriptors it calls upon or incorporates by reference.

5. A method in accordance with claim 1 wherein the step of obtaining an analyzer descriptor for each higher-level task includes obtaining at least one analyzer descriptor for a higher-level task that calls upon or that incorporates by reference at least one other analyzer descriptor for another higher-level task which in turn calls upon or which incorporates by reference at least one other analyzer descriptor for another higher-level task.

6. A method in accordance with claim 5 wherein the step of obtaining an analyzer descriptor includes obtaining one or more such descriptors that contain an analyzer name, description, and in the case of at least one higher-level analyzer descriptor, a list of other analyzer names whose descriptors it calls upon or incorporate by reference.

7. A method in accordance with claim 1 wherein the step of obtaining an analyzer descriptor for each higher-level task includes obtaining at least one analyzer descriptor for a higher-level task that calls upon or that incorporates by reference at least one other analyzer descriptor for another higher-level task which in turn calls upon or which incorporates by reference at least one other analyzer descriptor for another higher-level task that calls upon or that incorporates by reference at least one other analyzer descriptor for another higher-level task.

8. A method in accordance with claim 7 wherein the step of obtaining an analyzer descriptor includes obtaining one or more such descriptors that contain an analyzer name, description, and in the case of at least one higher-level analyzer descriptor, a list of other analyzer names whose descriptors it calls upon or incorporate by reference.

9. A system that analyzes data gathered by one or more collectors monitoring one or more nodes, said system comprising:

for each low-level analysis task that the system performs, an analyzer program and an analyzer descriptor, the program containing instructions that accept data from one or more collector types, analyze the data, and generates reports, and the analyzer descriptor written in XML or a comparable format and identifying the collector types whose data the program's instructions can accept;

for hierarchically higher-level analysis tasks, of which there is at least one, at least one analyzer program, and an analyzer descriptor for each of the higher-level tasks written in XML or a comparable format that calls upon, or that incorporates by reference, other analyzer descriptors; and

a framework that is guided by a set of one or more analyzer descriptors and, in the case of hierarchically higher-level analyzer descriptors, by the analyzer descriptors they call upon or incorporate by reference directly or indirectly, to execute repeatedly one or more times the set of analyzer programs corresponding to some of these analyzer descriptors against collector data gathered from a different node or set of nodes during each such execution of the set of programs, and during each execution of each program, to present the program with only the type of collector data the instructions of the executing program are designed to accept, as indicated by the analyzer descriptor corresponding to the executing program.

10. A system in accordance with claim 9 wherein each analyzer descriptor contains an analyzer name, an analyzer description, and in the case of analyzer descriptors for higher-level analyzer tasks, a list of other analyzer names whose descriptors it calls upon or incorporates by reference.

11. A system in accordance with claim 9 including at least one or more higher-level analyzer descriptor that calls upon, or that incorporates by reference, another high-level analyzer descriptor.

12. A system in accordance with claim 9 wherein each analyzer descriptor contains an analyzer name, an analyzer description, and in the case of analyzer descriptors for higher-level analyzer tasks, a list of other analyzer names whose descriptors it calls upon or incorporates by reference.

13. A system in accordance with claim 9 including at least one or more higher-level analyzer descriptor that calls upon, or that incorporates by reference, another high-level analyzer descriptor which, in turn, also calls upon, or that incorporates by reference, yet another high-level analyzer descriptor.

14. A system in accordance with claim 9 wherein each analyzer descriptor contains an analyzer name, an analyzer description, and in the case of analyzer descriptors for higher-level analyzer tasks, a list of other analyzer names whose descriptors it calls upon or incorporates by reference.

15. A system in accordance with claim 9 including at least one or more higher-level analyzer descriptor that calls upon, or that incorporates by reference, another high-level analyzer descriptor which, in turn, also calls upon, or that incorporates by reference, yet another high-level analyzer descriptor that, in turn, also calls upon, or that incorporates by reference, another high-level analyzer descriptor.

16. A system in accordance with claim 9 wherein each analyzer descriptor contains an analyzer name, an analyzer description, and in the case of analyzer descriptors for higher-level analyzer tasks, a list of other analyzer names whose descriptors it calls upon or incorporates by reference.

17. A system that analyzes data gathered by one or more collector means for monitoring one or more nodes, said system comprising:

for each low-level analysis task that the system performs, an analyzer means for accepting data from one or more collector means, analyzing the data, and generating reports, and an analyzer descriptor written in XML or a comparable format and identifying the types of collector means whose data the program's instructions can accept;

for hierarchically higher-level analysis tasks, of which there is at least one, at least one analyzer means, and an analyzer descriptor written in XML or a comparable format for each of the higher-level analysis task that calls upon, or that incorporates by reference, other analyzer descriptors; and

framework means guided by a set of one or more analyzer descriptors and, in the case of hierarchically higher-level analyzer descriptors, by the analyzer descriptors they call upon or incorporate by reference directly or indirectly, for executing repeatedly one or more times the set of analyzer means corresponding to some of these analyzer descriptors against collector means data gathered from a different node or set of nodes during each such execution of the set of analyzer means, and during each execution of each analyzer means, for presenting the analyzer means with only the type of collector means data the analyzer means are designed to accept, as indicated by the analyzer descriptor corresponding to the executing analyzer means.