JP2015130075A - Information processing apparatus, information processing method, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an information processing apparatus which executes processing at a high speed with a plurality of filters, while stably acquiring log.SOLUTION: An information processing apparatus performs pipeline processing by executing a filter corresponding to a function. A control section actuates a plurality of filters in the same process in a first mode, and does not output log data (S1103). When processing does not end normally in the first mode (an error or crash occurs in S1104), the processing is executed again in a second mode which outputs log data. When an error occurs, filters for outputting log data are limited (S1105, S1106). When a crash occurs, the control section causes the filter and a log output function to operate in different processes for retry, to output log data (S1108).

Description

  The present invention relates to an information processing apparatus, an information processing method, and a program for converting data.

  Patent Document 1 discloses a technique for converting image data using a plurality of filters such as “monochrome printer” and “resolution of 300 dpi or less”.

JP 2013-25415 A

  Depending on the filter shown in Patent Document 1 and image data input to the filter, there is a possibility that the processing is not normally completed by the filter. When the process is not completed normally, the administrator or developer needs to identify the cause. Therefore, a method of leaving log data (hereinafter simply referred to as a log) is examined in preparation for a case where some problem (data conversion failure or the like) occurs. In the method of outputting logs for all jobs processed by the system that executes the filter when leaving a log, there is a problem that processing speed decreases when a large number of jobs are processed. That is, the log output amount and the processing speed are in a trade-off relationship.

Furthermore, in order to realize a high-speed system, a method of operating a plurality of system filters in one process (single process) and a method of operating in multiple processes will be studied. Processing by a single process can be expected to obtain higher performance than processing by a multi-process. The reason for this is that multi-process communication requires inter-process communication when data is exchanged between processes, whereas single process requires only function calls and memory exchanges between components. This is because of high speed.
However, in a single process, if any component crashes or goes into an infinite loop, the entire system crashes or goes into an infinite loop. In that case, even if it is configured to leave a log as described above, the log may not be output normally. As a result, the log cannot be acquired normally, and the failure analysis of the system may be difficult.

  In this way, processing by a single process is fast, but there is a possibility that a log cannot be taken at the time of a crash. On the other hand, multi-process processing is slower than single-process, but there is a high possibility that logs can be acquired normally because of high availability.

  Therefore, an object of the present invention is to perform processing at high speed while stably acquiring logs in an information processing apparatus that performs processing using a plurality of filters.

  In order to solve the above-described problem, an apparatus according to the present invention is an information processing apparatus that performs processing by executing a plurality of filters according to a function, and the plurality of filters are processed in the same process and log data. If the processing using the plurality of filters in the first mode does not end normally in the first mode, the plurality of filters are processed in a plurality of processes and log data is output. Re-execution means for re-execution in the output second mode.

  ADVANTAGE OF THE INVENTION According to this invention, in the information processing apparatus which performs a process using a some filter, a process can be performed at high speed, acquiring a log stably.

It is a figure which shows the example of a system configuration | structure in order to demonstrate embodiment of this invention combined with FIGS. 2 is a diagram illustrating a hardware configuration of a server on which a print data conversion service operates. FIG. It is a figure explaining the software which operate | moves on the server of FIG. It is a block diagram which shows the component group relevant to a print data conversion service. It is a figure which illustrates the setting file which defined the kind and arrangement order of a filter. It is a figure which shows the example of a data conversion process. It is a figure which shows another example of a data conversion process. It is the schematic diagram which showed transmission / reception of the message between components. It is a schematic diagram which shows the structural example of a service catalog. It is a schematic diagram which shows the structural example of a service inventory. It is a flowchart which shows the flow of a process about a query. It is a flowchart which shows the flow of the whole process in this embodiment.

Embodiments of the present invention will be described below with reference to the accompanying drawings.
FIG. 1 is a system configuration diagram of a cloud print service according to the present embodiment.
When requesting conversion of print data, the requester 101 uses devices such as a PC (personal computer) 102, a tablet 103, and a portable terminal 104. These devices accept an operation from the requester 101 and request a print data conversion service existing on the network 105 to convert the print data. The print data conversion service is a program operating as a so-called Web service that publishes a known URL (Uniform Resource Locator) on the server 106. On the server 106, a Web server is operating that transfers a request for the URL to the print data conversion service. When there are a large number of requests, a single server 106 does not process, and a server system including a plurality of servers 106 processes. Accordingly, the load balancer 107 is installed in front of the plurality of servers 106, and the requests transferred to the servers 106 are uniformly distributed.

  The service provided by the print data conversion service is to convert the data received from the client 101 into a format that can be printed by the printer 108. The data received from the requester 101 includes formats created by various application programs such as documents and image files. On the other hand, the data that the printer 108 accepts as print data is data described in PDL (Page Description Language) and data for controlling printing such as the number of copies, unlike the above-described format.

  Thus, since there is a difference in the data format handled between the client 101 and the printer 108, a program for eliminating the difference by the data conversion process is required. Normally, a program called a printer driver performs this data conversion processing. The printer driver operates on a PC on which an application program operated by the client 101 is operating. However, the process of the printer driver is a process with a relatively large processing load that requires resources such as a CPU (Central Processing Unit) and a memory. For this reason, if the PC resources operated by the client 101 are not sufficient, the printing process takes time. As the demand for printing from terminals with relatively few resources, such as the tablet 103 and the portable terminal 104, increases, it is important to have the printer driver process performed on the Internet. The PC 102, the tablet 103, and the mobile terminal 104 receive an instruction from the requester 101 and requests the above-described URL to convert the print data into a format that can be printed by the printer 108. The converted data generated by the print data conversion service operating on the server 106 in response to the request is returned to the requesting terminal device. Depending on the instruction received from the client 101, the converted data may be stored in the file server 106 accessible by the print data conversion service.

Next, the hardware configuration of the server according to the present embodiment will be described. FIG. 2 is a hardware configuration diagram of the server 106 on which the software group of the print data conversion service operates. FIG. 2 illustrates only the basic configuration of the information processing apparatus, and various devices are added according to functions.
The print data conversion service is an application program that operates on an operating system (hereinafter referred to as OS), and is stored in an HDD (Hard Disk Drive Device) 202 or a ROM (Read Only Memory) 203. The CPU 205 reads the OS and application programs from the HDD 202 or the ROM 203, loads them into the RAM 204, and executes them. As a result, various processes proceed, and the function of the print data conversion service is realized. The processing result is stored in the HDD 202 as a file or stored in a RAM (Random Access Memory) 204 as data. The application program acquires user input operation signals and reading values of various sensors from an input device 207 connected to the computer. The application program further outputs information to the output device 206 and displays the processing result on the display device. Further, communication processing is performed with other computers and devices connected to the network via the communication device 208. These hardware components are connected to each other via a bus 201 and can be operated from an application program.

Next, the software configuration will be described with reference to FIG. FIG. 3 is a block diagram illustrating a configuration example when a software group including the print data conversion service 301 operates on the server 106.
The Web server 302 receives a request transmitted by an HTTP (Hypertext Transfer Protocol) protocol from an application on the PC 102 operated by the client 101. The load balancer 107 is positioned in front of the Web server 302 and distributes requests to the Web servers 302 operating on the plurality of servers 106. The details of this distribution method are not relevant to the present invention, and thus the description thereof is omitted.

Note that in this embodiment, the server 106 does not necessarily mean a single physical machine. A virtualization technique that allows a plurality of virtual computers to appear on a single computer has become common. From the viewpoint of software, it is impossible to determine whether or not the OS on which it is operating is virtually created, or it is not necessary to determine. Therefore, the server 106 of this embodiment may mean a “virtualized computer”. Hereinafter, “computer” is treated as meaning both a computer as a physical machine and a virtualized computer in a broad sense.
The application server 303 analyzes the URL designated as the request destination, and determines the associated print data conversion service 301. Normally, the Web server 302 and the application server 303 can handle a plurality of types of services, and can identify requests corresponding to URLs and distribute requests. In this embodiment, one of a plurality of types of services is assumed to be a print data conversion service 301. In addition to this service, a service is operating, but the description is omitted because it is not directly related to the present invention.

The print data conversion service 301 operates as a process. In order to execute a plurality of requests in parallel, the application server 303 activates a plurality of print data conversion services 301 and distributes the requests to each. The application server 303 further manages a session for specifying a series of requests from the requester 101 when the print data conversion service 301 operates. Specifically, session affinity is realized in which requests from the same client 101 are transferred to the same print data conversion service 301.
The print data conversion service 301 will be described with reference to FIG. FIG. 4 is a diagram for explaining the relationship between software component groups used when the print data conversion service 301 converts requested print data. Hereinafter, a component started as a process is referred to as a “component process”.

The print data conversion service 301 operates as a process, and delegates conversion processing to the job control unit 402. The print data conversion service 301 loads a job processing reception unit 401 as a substitute for the job control unit 402. Further, the job processing reception unit 401 calls a function group provided by Application Programming Interface (API) to request execution of the conversion processing, and further receives the result. When executing the print data conversion process, the print data conversion service 301 passes a set of data to be processed to the job process reception unit 401. This set of data to be processed is called a job. The job has a data structure called “ticket” in which conversion target data, setting values for conversion processing, and the like are bundled into a coherent format.
The job control unit 402 also operates as a process, and holds a communication interface for accepting conversion processing. A job request to the communication interface is not an HTTP protocol but a dedicated protocol. The process of calling the API of the job process reception unit 401 is performed as interprocess communication to the conversion process reception communication interface prepared by the job control unit 402. That is, it appears to the print data conversion service 301 that the conversion processing is being executed by the job processing reception unit 401. However, in actuality, this is realized by the interaction of the job control unit 402, a plurality of filter groups described later, and several components used by the filter groups.

In the following, the function of each component shown in FIG. 4 will be described, and the outline of the data conversion process will be described with reference to FIG. First, the process information management unit 403 will be described.
The process information management unit 403 is a special process that is activated only when the application server 303 is activated. A process of setting an OS running on the application server 303 and automatically starting the process of the process information management unit 403 when the OS is started is executed.

  The process information management unit 403 plays a role of starting a component as a process, and provides a function of returning an endpoint of another component in response to a request from another component. An endpoint refers to a TCP / IP (Transmission Control Protocol / Internet Protocol) listen port. One component obtains an endpoint published by another component, and uses the function provided by that component by accessing that endpoint. An operation in which a certain component inquires the process information management unit 403 about an endpoint of another component is referred to as a “query”.

  Among the components constituting the print data conversion service 301, the job processing receiving unit 401, the job control unit 402, and the filter control unit 404 can use a service for recording log data provided by the log output unit 408 as a log file. . For example, an administrator or developer can grasp what has occurred and take measures based on the analysis result of the log file. Therefore, they query the process information management unit 403 for the purpose of obtaining the end point of the log output unit 408. Similarly, the job processing reception unit 401 queries the job control unit 402, and the job control unit 402 includes, for example, a first filter (filter A) 405, a second filter (filter B) 406, and a third filter (filter C) 407. Make a query. The flow of query processing will be described later with reference to FIGS.

  Services provided by components are disclosed to other components as endpoints accessible from the network. Specifically, a set of TCP / IP address and port number is disclosed. A component acquires and uses a service I / F (interface) provided by another component. There are a plurality of types of components as processes, and a plurality of the same components exist on the same PC. Therefore, the port number of the service I / F disclosed by each component cannot be set to a fixed value, and must be dynamically generated. Dynamically generated endpoints cannot be referenced from others without some sort of management mechanism. The process information management unit 403 is responsible for this management. In addition, the component needs a means to obtain endpoints dynamically generated by other components. This means is a query. The process information management unit 403 returns the end point of the requested component in response to a query from another component. The job control unit 402 requested to execute the job via the job processing receiving unit 401 selects a filter necessary for conversion based on “job type information” included in the job. For example, the job control unit 402 holds the necessary combinations of filters and the arrangement order thereof as a setting file.

FIG. 5 illustrates a setting file for describing a necessary combination of filters and their arrangement order. An element 501 for describing job type information is shown.
Job type information is specified in the id attribute. An element group 502 is an element for describing the combination and arrangement order of filters. The name attribute describes the type of filter, and the use of the version of the filter specified by the version attribute is described. The element group 502 defines the filter order in the element description order. In the case of the element group 502 illustrated in FIG. 5, processing is performed in the order of Filter A, Filter B, and Filter C. The job control unit 402 searches for an element 501 that matches the job type information included in the job, and adopts the combination and order of filters described under the matching element 501.
Note that the necessary combination of filters and their arrangement order may be determined by analyzing print data instead of job type information. In this case, a predetermined size is read from the top of the print data, and a process for determining the type of the print data from the characteristic content included therein is executed. For example, data in JPEG (Joint Photographic Experts Group) format, PDF (Portable Document Format) format, or the like is given as input data. Next, the conversion destination print data format included in the job is acquired, and when both the conversion source print data format and the conversion destination print data format are determined, a filter group necessary for this conversion is selected. Processing is performed.

FIG. 4A shows a state in which each filter is loaded in the filter control unit 404 after a plurality of necessary filters are selected. In the example of FIG. 4A, a filter A405, a filter B406, and a filter C407 are illustrated as filters necessary for conversion. Each filter is loaded in the filter control unit 404.
The job control unit 402 generates a data transfer channel called a pipeline 409 between the filter A 405, the filter B 406, and the filter C 407. For example, TCP / IP is used as the protocol. Pipeline 409 basically transfers data only in one direction. A pipeline 409 is configured to make a round with the job control unit 402, the filter A 405, the filter B 406, the filter C 407, and the job control unit 402. The job control unit 402 also generates a pipeline 409 with the job processing reception unit 401. As described above, the print data from the job processing receiving unit 401 is returned to the plurality of components by the generation of the pipeline 409, and finally returns to the job processing receiving unit 401 after the conversion is completed.

  Further, the job control unit 402 can query the process information management unit 403 for “filter control unit 404 loaded with a plurality of filters”. FIG. 4B shows an example in which filter A 405, filter B 406, and filter C 407 are all loaded in the same filter control unit 404. In this case, the filter control unit 404 can also hold a local log output unit 411 as shown in FIG. Each filter (405, 406, 407) can output log data (logging data) via the local log output unit 411.

The filter is prepared in the form of a library module that implements only data conversion processing. The filter control unit 404 exists as a process, and loads a filter designated at the time of execution. The filter control unit 404 is responsible for connection to the control bus 410 and message transmission / reception, communication with the job control unit 402, processing for transferring the log output of the filter to the log output unit 408, and the like.
The local log output unit 411 can operate in the process of the filter control unit 404. Each filter passes the log data to the local log output unit 411 via the API. The local log output unit 411 outputs a log file when a certain amount of log data is accumulated. The log data may be sequentially output to the log file without accumulating. The setting of the log file output path and the like is specified by a configuration file delivered when the local log output unit 411 is activated.

At least one log output unit 408 exists on the control bus 410 as a process. Each component transmits log data to the log output unit 408 via the network. The log output unit 408 outputs a log file when a certain amount of log data is accumulated. The log data may be sequentially output to the log file without accumulating. The setting of the log file output path and the like is specified by a configuration file delivered when the log output unit 408 is activated.
The component on the control bus 410 performs log output during job execution. If the local log output unit 411 performs output individually, the log file is divided for each filter control unit 404, making log analysis difficult. Therefore, in this embodiment, the log output unit 408 is configured to be able to output to log files in a lump.

Next, the outline of the data conversion process will be described with reference to FIGS.
FIG. 6 shows an example of data conversion processing after the job processing receiving unit 401 receives job processing from the print data conversion service 301.
In step S <b> 601, the job processing reception unit 401 makes a query request to the job control unit 402. The job processing reception unit 401 makes a query request to the process information management unit 403 and acquires the end point of the job control unit 402. In step S <b> 602, the job processing reception unit 401 makes a job processing request to the job control unit 402. Here, job data is transferred from the job processing receiving unit 401 to the job control unit 402. In that case, it may be passed by a job data path, or may be passed by memory or inter-process communication.

In step S <b> 603, the job control unit 402 determines a filter type and an arrangement order (see FIG. 5) based on the job contents, and makes a query request for the filters. The job control unit 402 makes a query request to the process information management unit 403 for the filter control unit 404 corresponding to each filter, and acquires each endpoint. Here, description will be made assuming that processing is performed in the order of filter A405, filter B406, and filter C407.
S604 is processing for generating a pipeline. The job control unit 402 instructs each filter control unit 404 to generate a pipeline. Each filter control unit 404 generates a pipeline with each adjacent component process. Each filter control unit 404 waits for a job to flow from the input pipeline of each filter control unit 404 and notifies the job control unit 402 of the completion of the pipeline generation processing. When the job control unit 402 instructs each filter control unit 404 to generate a pipeline, the job control unit 402 notifies each filter control unit 404 of the service I / F endpoint of the job control unit 402. When an error occurs, the filter control unit 404 notifies the job control unit 402 of the error through the service I / F of the job control unit 402.

S605 is a data conversion process. The job control unit 402 passes data to the pipeline connected to the first-stage filter control unit 404. The filter A 405 receives data through the filter control unit 404 and performs data conversion processing. Next, the filter A 405 passes the data to the pipeline connected to the filter B 406 through the filter control unit 404. Similarly, the filter B406 and the filter C407 sequentially perform data conversion processing. Finally, the converted data is transmitted from the filter C 407 to the job control unit 402 through the filter control unit 404. In step S <b> 606, the job control unit 402 returns the converted data to the job processing reception unit 401.
FIG. 6 illustrates a case where each filter is loaded into the filter control unit 404.

  Next, an example of loading a plurality of filters to the same filter control unit 404 will be described with reference to FIG. The difference from FIG. 6 is the contents of each process of S603 and S604. In step S <b> 603-01, the job control unit 402 determines the type and order of filters based on the job contents. Then, the job control unit 402 makes a query request to one filter control unit 404 loaded with a plurality of filters to the process information management unit 403, and acquires an end point. Thereby, one filter control unit 404 loaded with the filter A 405, the filter B 406, and the filter C 407 can be acquired.

  S604-01 is processing for generating a pipeline. The job control unit 402 instructs one filter control unit 404 to generate a pipeline. In S604 of FIG. 6, as described above, each filter control unit 404 generates a pipeline with each adjacent component process. On the other hand, in S604-01 of FIG. 7, the filter control unit 404 generates a pipeline for the filter A 405, the filter B 406, and the filter C 407 in the same process.

Next, component generation will be described. FIG. 8 is a diagram illustrating a sequence for generating a component as a process.
FIG. 8A shows a situation where only the process information management unit 403 exists. This situation corresponds to a state immediately after the server 106 is activated in FIG. At this point, the print data conversion service 301 has not been activated yet. The process information management unit 403 is connected to the control bus 410 by itself.
The process information management unit 403 holds a service catalog 701. For example, the service catalog 701 is stored as a file in the HDD 202 of the server 106 and is read by the process information management unit 403. FIG. 9 shows a conceptual diagram of the service catalog. The service catalog 701 includes a list of components installed on the server 106.

9A and 9B, an entry 801 is a service catalog entry and exists for each component. A component name 802 is a component name of the entry. The service version 803 indicates the component version. Even components with the same name may have different versions. In this case, in order to provide different functions, they are treated as different entries on the service catalog 701.
A hosting process path 804 indicates the hosting process path of the component. An execution file path 805 indicates an execution file path loaded by the hosting process path 804. The configuration file path 806 is a path of a configuration file passed from the process information management unit 403 to the component when the component is activated as a process. The start mode 807 indicates whether the component process is started on demand (Ondemand) or is in hot standby (HotStandby). When the activation mode is on-demand, the component process is activated when there is a query for a component that specifies the component. In the case of hot standby, when the process information management unit 403 is activated, the component process is activated.

  FIG. 9A shows a catalog of the filter control unit 404 loaded with “filter A 405”, that is, the filter control unit 404 loaded with the filter A 405 described with reference to FIG. On the other hand, FIG. 9B shows a catalog of the filter control unit 404 that loads the filter A405, the filter B406, and the filter C407 on the same process. That is, it is a service catalog of one filter control unit 404 loaded with the filters A405, B406, and C407 described with reference to FIG.

  FIG. 8B illustrates a case where the startup mode of a certain component A 702 is designated as hot standby. When the process information management unit 403 is activated, the process information management unit 403 refers to the service catalog 701 and searches for the component A 702 in which the activation mode is designated as hot standby. Next, the process information management unit 403 gives the configuration file path 806 from the service catalog 701 as an activation argument, and activates the component A 702 as a process. When the component A 702 is activated and connected to the control bus 410, a unique ID (identification information) is given. For example, a GUID (Globally Unique Identifier) is used as this ID. The component A 702 refers to the configuration file 703 from the configuration file path 806 to acquire necessary initial values and the like, and initializes itself using the initial values.

  FIG. 8C shows a state in which the component A 702 acquires a TCP / IP port and starts waiting for a service I / F 704 in order to provide a service. Since the component A 702 is ready to provide services to other components, the component A 702 creates and transmits its advertisement message 705 to the control bus 410. The advertisement message 705 describes the name, ID, version, and end point (address, port number) of the service I / F 704 of the component A 702. This message is transmitted to the process information management unit 403. The advertisement message 705 is transmitted in order for the component to declare its own activation completion and service provision start. The process information management unit 403 that has received the advertisement message 705 can confirm the activation of the component A 702 and can acquire the end point of the service I / F 704.

FIG. 8D shows a state in which the process information management unit 403 receives the advertisement message 705 from the component A 702 and registers it in the service inventory 706 that is a management table for subsequent management.
FIGS. 10A and 10B are diagrams schematically showing the contents of the service inventory 706. FIG. 10A shows the service inventory 706 of the filter control unit 404 described with reference to FIG. FIG. 10B shows the service inventory 706 of the filter control unit 404 described with reference to FIG. An entry 901 is an entry of the service inventory 706. An entry 901 exists for each component process. When the process information management unit 403 receives the advertisement message 705, the process information management unit 403 creates an entry of the component process that is the transmission source of the message in the service inventory 706. Then, the process information management unit 403 describes the content of the message in the entry 901. Specifically, an ID 902, a name 903, a version 904, and an end point 905 (address and port number) are described. In the state 906, when the process information management unit 403 receives a state change message described later, the state of the component process is recorded. State 906 may take the values “idling” or “busy”. “Idling” is a state in which the component process can accept processing. “Busy” is a state in which the component process cannot accept processing. When each component process starts, it starts in an “idling” state. The “busy” state will be described later with reference to FIG.

  FIG. 8E shows a state in which the component A 702 transmits a query message 708 to the process information management unit 403 in an attempt to use the component B 709. The query message 708 describes the name of the component B 709 and the version to be used. A component can be created in multiple versions. Multiple components with the same name are created and released to the market as different versions due to differences in function. Accordingly, the component A 702 can make a query request by designating the version of the component B 709 that the component A 702 wants to use, such as version “2.0.0.1”. If there is no need to use a specific version, the version field is designated as blank. Since the component B 709 has not yet been generated at the time shown in FIG. 8E, the process information management unit 403 newly generates the component B 709. Similar to the generation procedure of the component A 702 shown in FIGS. 8B to 8D, the component B 709 is generated. Details of query processing in the process information management unit 403 will be described later with reference to FIG.

FIG. 8F shows a state in which the process information management unit 403 returns a response to the query as a query response message 710 to the component A 702. The query response message 710 includes an endpoint indicating the name, version, and service I / F of the component B 709.
FIG. 8G shows a state in which the service I / F 711 of the component B 709 acquired by the component A 702 that has received the query response message 710 is used. Some components can handle service requests from other services that access the service I / F at the same time, and others cannot. For example, the log output unit 408 has a capability of receiving log data from a plurality of services in parallel and recording them in a log file. On the other hand, a filter that converts a job by configuring a pipeline does not have the capability of simultaneous parallel processing. Therefore, when the component B 709 is a component that does not have the capability of simultaneous parallel processing, the component A 702 transitions to a busy state in which a request from another component cannot be accepted when accessed.

  The busy state continues until the component A 702 declares explicit access termination to the component B 709 through the service I / F. When the component B 709 transitions to the busy state, it is necessary to prevent itself from being detected by the query. As shown in FIG. 8H, the component B 709 transmits a state change message 712 to the process information management unit 403. The state change message 712 describes that the component B 709 has transitioned to the busy state. Upon receiving the state change message 712, the process information management unit 403 searches for an entry of the component B 709 in the service inventory 706, and changes the state 906 to a “busy” state. When the status 906 of the component B 709 is changed to the “busy” status, the process information management unit 403 will not select the process of the component B 709 even if it receives a query specifying the name. However, there may be a plurality of other component B processes in the service inventory 706. If any of them is not busy, it is returned to the component that issued the query as “available component B”. As to which process is selected as a result of a query from a plurality of component processes having the same name and version, there is a method of returning the first discovered process, for example. Alternatively, for each component process, a method may be employed in which the number of times returned to the query issuing component is counted and the process with the lowest count value is returned.

Next, the details of the query processing will be described with reference to the flowchart of FIG.
In step S1001, the process information management unit 403 receives a query request from another component. The query includes the name of the component and the version you want to use. In step S1002, the process information management unit 403 searches the service catalog 701 and determines whether there is an entry 801 corresponding to the component name and version received in step S1001. If it is determined that the entry 801 exists in the service catalog 701, the process proceeds to S1003. If it does not exist, the process proceeds to S1007.

  In step S1003, the process information management unit 403 determines whether there is a component process that can accept a job. The process information management unit 403 searches the service inventory 706 and searches for an entry 901 corresponding to the component name and version received in S1001. Then, processing for determining whether or not there is one or more component processes whose status 906 is “idling” is executed. If there is a component process whose status 906 is “idling”, the process proceeds to S1004, and if it does not exist, the process proceeds to S1005.

  In step S1004, the process information management unit 403 notifies the query request source of the endpoint of the requested component process. In S1004, a process of notifying the query request source of the endpoint for one of the detected component processes is executed. At that time, the first detected component process may be returned. Alternatively, for each component process, how many times the endpoint is returned to the query request source may be stored in the service inventory 706 and the component process with the smallest number of times may be returned.

  In step S <b> 1005, the process information management unit 403 activates the component for which the query has been requested as a process. In step S1006, the process information management unit 403 receives an advertisement message 705 from the component process activated in step S1005. After this processing, the process proceeds to S1004, and the component process activated in S1005 is notified to the query request source. Although not shown in FIG. 11, if the advertisement message 705 does not arrive at the process information management unit 403 within the specified time, it may be determined as an error and the process may proceed to S1007. In step S1007, the process information management unit 403 notifies an error to the query request source.

Based on the basic flow of data conversion service processing described above, processing for executing a job will be described with reference to the flowchart of FIG.
First, the job processing reception unit 401 receives input data and transmits the input data to the job control unit 402 (S1101). The job control unit 402 specifies a filter necessary for processing. In the present embodiment, processes activated by the process information management unit 403 will be described as filters A405, B406, and C407. The job control unit 402 makes a query to the process information management unit 403 regarding the filter control unit 404 in which the above three filters are operated in the same process. Upon receipt of the query, the process information management unit 403 activates the requested filter, and the job control unit 402 generates a pipeline (S1102). This is the pipeline described with reference to FIGS.

Next, using the constructed pipeline, the job control unit 402 instructs the filter control unit 404 to execute the job as “no log output”. As an instruction method, there is a method of describing in a ticket (a data group having a structure in which setting values for conversion processing and the like are bundled together), but a notification method is not particularly limited. The job control unit 402 executes the job using each filter without outputting the log (S1103).
The job control unit 402 determines whether the job has been executed normally (S1104), and branches to three depending on the determination result to switch the processing. If the job ends normally, the process proceeds to S1109. If a crash occurs, the process proceeds to S1107. If an error other than the crash occurs, the process proceeds to S1105.
In step S1105, the job control unit 402 identifies a filter in which an error has occurred. When an error occurs, each filter notifies the job control unit 402 that the error has occurred, so that the job control unit 402 can identify the filter in which the error has occurred.

Next, the job control unit 402 instructs the execution of the job as “local log output is present (instruction for causing the local log output unit 411 to output a log)” only for the filter specified in S1105. For this instruction, a method such as description in a ticket is used (S1106). Thereafter, the process proceeds to S1109.
The method of determining that a crash has occurred in S1104 can be realized by identifying a component in which a problem has occurred by detecting pipeline disconnection between components (between processes). Specifically, packets for survival confirmation are periodically sent between processes, for example, between the job control unit 402 and the filter control unit 404. It is possible to detect a crash by detecting that a packet has been interrupted by a timeout. For example, when the packet from the filter control unit 404 is interrupted, the job control unit 402 can determine that the filter control unit has crashed. In this case, the job control unit 402 queries the process information management unit 403 for each filter control unit 404 corresponding to a plurality of filters (three filters in this example). Upon receiving the query, the process information management unit 403 activates the requested filter, and the job control unit 402 generates a pipeline (S1107). This is the pipeline described with reference to FIGS.

In step S <b> 1108, the job control unit 402 instructs job execution only for the filter specified in step S <b> 1105 as “log output is present (instruction for causing the log output unit 408 to output a log)”. For this instruction, a method such as describing in a ticket is used. Thereafter, the process proceeds to S1109.
In step S1109, the job control unit 402 executes post-processing and ends the processing. Specifically, the post-processing is processing for discarding the pipeline or notifying the process information management unit 403 of the change to the “idling” state. The process information management unit 403 receives the notification of change to the “idling” state, and changes the state 906 of the component process to “idling” in the service inventory 706.

The information processing apparatus according to the present embodiment has at least a first mode and a second mode when pipeline processing is performed by operating a plurality of filters. That is, the first mode is a mode in which a plurality of filters are operated in the same process and log data is not output. The second mode is a mode in which a plurality of filters are operated by a plurality of processes and log data is output when processing using the plurality of filters is not normally completed in the first mode. The execution means for executing a plurality of filters in the first mode and the re-execution means for re-execution of the plurality of filters in the second mode may be the same means or different means.
As described above, the log data is not output in the first mode, but the processing speed can be maintained by operating the filter in the same process. When an error occurs, the job can be re-executed in the second mode to output log data. Log data can be stably acquired at the time of re-execution by using a log output mechanism that operates in a separate process and operates in a separate process at the time of a crash.
(Other embodiments)
The present invention can also be realized by executing the following processing. That is, software (program) that realizes the functions of the above-described embodiments is supplied to a system or apparatus via a network or various storage media, and a computer (or CPU, MPU, or the like) of the system or apparatus reads the program. It is a process to be executed.

402 Job control unit 403 Process information management unit 404 Filter control unit 405, 406, 407 Filter 409 Local log output unit

Claims (11)

An information processing apparatus that performs processing by executing a plurality of filters according to functions,
Execution means for executing the plurality of filters in the same process and in a first mode that does not output log data;
Re-execution means for re-executing the plurality of filters in a plurality of processes and in a second mode for outputting log data when the processing using the plurality of filters in the first mode is not normally completed; And an information processing apparatus.   The execution means executes pipeline processing by the plurality of filters in the first mode, and when the pipeline processing in the first mode does not end normally, the re-execution means executes the second mode The information processing apparatus according to claim 1, wherein the pipeline processing is re-executed to output log data.   3. The device according to claim 2, further comprising: a specifying unit that specifies which of the plurality of filters has caused an error when the pipeline processing in the first mode is not normally completed. 5. Information processing device.   The information processing apparatus according to claim 3, wherein the re-execution unit performs a process of outputting log data to the filter specified by the specifying unit.   The information processing apparatus according to claim 1, further comprising: a local log output unit that outputs log data acquired from the filter to a log file. A program executed in a computer that performs processing by executing a plurality of filters according to functions,
In the computer,
An execution step of executing the plurality of filters in the same process and in a first mode that does not output log data;
A re-execution step of re-executing the plurality of filters in a plurality of processes and in a second mode for outputting log data when processing using the plurality of filters in the first mode is not normally completed; , A program characterized by being executed. In the computer,
In the execution step, pipeline processing by the plurality of filters is executed in the first mode, and when the pipeline processing in the first mode does not end normally, in the re-execution step, the second mode The program according to claim 6, wherein the pipeline processing is re-executed and log data is output. In the computer,
8. The method according to claim 7, wherein when the pipeline processing in the first mode does not end normally, a specific step of specifying which filter among the plurality of filters has caused an error is executed. The listed program. In the computer,
The program according to claim 8, wherein in the re-execution step, a process of outputting log data to the filter specified in the specification step is executed. In the computer,
The program according to any one of claims 6 to 9, wherein a local log output step of outputting log data acquired from the filter to a log file is executed. A method of controlling an information processing apparatus that performs processing by executing a plurality of filters according to a function,
By the information processing apparatus,
An execution step of executing the plurality of filters in the same process and in a first mode that does not output log data;
A re-execution step of re-executing the plurality of filters in a plurality of processes and in a second mode for outputting log data when processing using the plurality of filters in the first mode is not normally completed; The control method characterized by having.

Images