JP2011086083A - Test device and test method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To perform a quick and accurate degradation test using real data with respect to a service to be provided on a daily basis or the huge patch of an OS. <P>SOLUTION: This test method in a test device is configured to create test data information from request input history information and access history information to the system, and to update the test data information so that the request input time of the second test data can be delayed by a prescribed time when the access time of first test data as update access and second test data as update access are overlapped, and to input test data included in the updated test data information to a system for a test. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a test apparatus and a test method, and more particularly to creation of test data for testing whether or not an updated service or OS operates normally.

  In the recent Internet and intranet, services that can be used without stopping for 24 hours are increasing. However, there are cases where there are security holes and bugs that can be used as a trigger for attacks on services and OSs that provide them, so patches are provided on a daily basis. Therefore, it is indispensable to conduct an immediate and reliable regression test.

  The regression test can be performed by comparing the execution result of the service before applying the patch and the execution result of the service after applying the patch. For example, in Patent Document 1, an input parallel device is placed on a production system before application of a patch in which a service to be tested operates, and a duplicated packet is transmitted to a verification system after application of the patch. A regression test is carried out by comparing the results of the verification system with the verification system.

JP 2007-257588 A

  In Patent Document 1, since a packet input to the production system is duplicated and transmitted to the verification system as it is, when access to the same table is caused by a plurality of packets, a lock collision or a dirty read occurs, There may be a difference in the order of access to the DB of the production system and the verification system. If there is a difference in the access order, the result may be different, so that there is a problem that an accurate regression test cannot be performed. The lock collision is a problem that occurs when a certain DB access locks the information and another DB access tries to lock the information. Dirty read is a problem that occurs when one DB access updates information while another DB access refers to the information.

  Further, in the method of Patent Document 1, a test time equivalent to the time required when the real data is processed by the production system is necessary for the execution of the regression test. Therefore, there is a problem that a huge amount of time must be spent on the test time.

  In the test method by a test apparatus for solving the above-described problem, the test apparatus includes request input history information indicating input times of a plurality of requests to a production system including a first database and a first server, Test data information including request input times, access times, and access types of a plurality of test data is created based on access history information indicating access types and access times of the plurality of accesses to the first database corresponding to the request. The test apparatus refers to the test data information, is included in the plurality of test data, and the access start time is later than the first test data, the access type being update access, and the first test data, The access time for the second test data whose access type is update access overlaps If so, the test data information is updated so as to delay a request input time of the second test data by a predetermined time, and the test apparatus is provided for a test system including a second database and a second server. The test method is characterized in that the plurality of test data included in the updated test data information is input.

  A test apparatus for solving the above-described problem includes an interface connected to a production system including a first database and a first server, a test system including a second database and a second server, Created by request input history information indicating input times of a plurality of requests to the production system and access history information indicating access types and access times of a plurality of accesses to the first database corresponding to the requests A memory for storing test data information including a request input time, an access time, and an access type of a plurality of test data; a reference to the test data information; the access type being an update access included in the plurality of test data; Access time is later than the first test data and the first test data. When the access time overlaps with the second test data whose access type is update access, the test data information is updated so as to delay the request input time of the second test data by a predetermined time, and the test And a processor for inputting the plurality of test data included in the updated test data information to a system for use.

  According to the present invention, it is possible to perform a quick and reliable regression test using real data for a vast number of patches and services provided daily.

It is a block diagram of the production environment system in a present Example. It is a block diagram of the test environment system in a present Example. It is a hardware block diagram of the computer in a present Example. It is a hardware block diagram of the test data control apparatus in a present Example. It is a hardware block diagram of the packet replication apparatus in a present Example. It is a hardware block diagram of the output log collection device in a present Example. This is real data input timing information in this embodiment. It is a figure which shows DB access log information in a present Example. It is a figure which shows the test data information in a present Example. It is a figure which shows the flowchart which shows the real data collection process in a present Example. It is a figure which shows the flowchart explaining the test data creation method in a present Example. It is a figure which shows the flowchart explaining the deletion candidate selection process of the lock collision and dirty read occurrence location in a present Example. It is a figure which shows the flowchart explaining the deletion implementation process of the lock collision and dirty read occurrence | occurrence | production location in a present Example. It is a figure which shows the flowchart explaining the input timing shortening process in a present Example. It is a figure which shows the flowchart explaining the input timing adjustment process in a present Example. It is a figure which shows the flowchart explaining the regression test process in a present Example.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 is a diagram illustrating the configuration of a production environment system according to the present embodiment. The production environment system includes a load balancer 101, AP servers (for production) 102 and 103, DB servers 104 and DB105, a packet replication device 106, a test data repository 107, an output log collection device 108, and an output log repository 109. Is included. Each repository and DB may be a hard disk drive (HDD) or a storage device that is a set of HDDs. The load balancer 101 is connected to an AP server (for production) 102, an AP server (for production) 103, a packet replication device 106, and an output log collection device 108 via a LAN 111. The DB server (for production) 104 is connected to the AP server (for production) 102, the AP server (for production) 103, and the DB (for production) 105 via the LAN 111. The packet replication device 106 is connected to the load distribution device 101, the test data repository 107 and the LAN 111. The output log collection device 108 is connected to the output log repository 109 via the LAN 111.

  FIG. 2 is a diagram showing the configuration of the test environment system in the present embodiment. The test environment system includes a test and other control device (test device) 205, an AP server (for test) 201 and 102, a DB server (for test) 203, a DB (for test) 204, an output result repository 206, and test data. Repository 107, output log collection device 108, and output log repository 109 are included. The test data repository 107, the output log collection device 108, and the output log repository 109 are the same as in FIG. Each repository and DB may be a hard disk drive (HDD) or a storage device that is a set of HDDs. The test data control device 205 is connected to the test data repository 107, the output result repository 206, the AP server (for testing) 201, the AP server (for testing) 202, and the output log collection device 108 via the LAN 111. The DB server (for test) 203 is connected to the AP server (for test) 202, the AP server (for test) 201, and the DB (for test) 204 via the LAN 111. The output log collection device 108 is connected to the output log repository 109 and the test data control device 205 via the LAN 111.

  FIG. 3 shows an AP server (for production) 102, an AP server (for production) 103, a DB server (for production) 104, an AP server (for testing) 201, an AP server (for testing) 202, and a DB server in this embodiment. 2 is a diagram illustrating a hardware configuration of (for testing) 203. FIG. The computer 301 includes a CPU 302, a memory 303, an input / output I / F 304, an HDD 305, a ROM 306, and a network card 307 that are connected to each other, and is connected to the LAN 311 through the network card 307. A display 308, a keyboard 309, and a mouse 310, which are user interfaces, are connected to a computer program through an input / output I / F 304. A program for monitoring a service stored in the HDD 305 or the ROM 306 is read into the memory 303 and the CPU 302 executes the monitoring program.

  FIG. 4 is a diagram illustrating a hardware configuration of the test data control apparatus 205 in the present embodiment. In the present embodiment, the test data control device 205, the packet duplicating device 106, and the output log collecting device 108 are described as physically separate devices, but are not limited to this. The test data control device 205 and the packet duplicating device 106 may be mounted on the same physical device. The test data control device 205, the packet replication device 106, and the output log collection device 108 may be mounted on the same physical device. The test data control device 205 includes a memory 4001 that stores programs and the like, a CPU 4002 that reads and executes programs stored in the memory 4001, and an I / F 4003 that is connected to the LAN 111. The memory 4001 stores a test data extraction program 401, an input timing control program 402, a test data transmission program 403, an output log comparison program 404, an output result comparison program 405, an output result reception program 406, and an output log extraction program 407. Hereinafter, when a process is described using a program as a subject, it indicates that the CPU executes the process by reading and executing the program stored in the memory.

  The test data extraction program 401 extracts test data to be input next from the test data repository 107. The input timing control program 402 refers to the test data information and controls when and to which server the test data extracted by the test data extraction program 401 is transmitted. The test data transmission program 403 transmits test data to each server. The output log extraction program 407 extracts various logs stored in the output log repository 109. The output log comparison program 404 compares the log before the test environment system update extracted by the output log extraction program 407 with the log after the test environment system update. The output result receiving program 406 receives the results sent from the AP server (for testing) 201 and the AP server (for testing) 202 and stores them in the output result repository 206. The output result comparison program 405 compares the output result before updating the test environment system stored in the output result repository 206 with the output result after updating the test environment system. Detailed processing executed by each program will be described later with reference to flowcharts.

  FIG. 5 is a diagram illustrating a hardware configuration of the packet replication device 106 according to the present embodiment. The packet replication device 106 includes a memory 5001 that stores programs and the like, a CPU 5002 that reads and executes a program stored in the memory 5001, and an I / F 5003 that is connected to the LAN 111. The memory 5001 stores a packet replication program 501 and a packet control program 502. The packet duplication program 501 duplicates the request transmitted from the user to the production environment system. The packet control program 502 controls the operation on the packet duplicating device 106.

  FIG. 6 is a diagram illustrating a hardware configuration of the output log collection device 108 in the present embodiment. The output log collection device 108 includes a memory 6001 that stores programs and the like, a CPU 6002 that reads and executes a program stored in the memory 6001, and an I / F 6003 that is connected to the LAN 111. The memory 6001 stores an output log collection program 601 and an output log management program 602. The output log collection program 601 collects output logs from the AP server (for testing) 201, the AP server (for testing) 202, and the DB server (for testing) 203. The output log management program 602 stores the output log collected by the output log collection program 601 in the output log repository 109.

  FIG. 7 is a diagram showing input timing information of real data in the present embodiment. The input timing information is created by the packet replication program 501 operating on the packet replication device 106 and stored in the test data repository 107. The input timing information includes a session ID 701, a request ID 702, a transmission source IP 703, and a transmission time 704.

  For example, item 1 in FIG. 7 indicates real data in which the session ID 701 is 123 ABC, the request ID 702 is AAABBB, the transmission source IP 703 is 192.168.1.10, and the transmission time 704 is 20090901012754. The session ID is a management ID issued by the AP server to manage a session, which is a series of processes from establishing a connection to disconnecting it. One session includes one or more requests. The request ID is a management ID issued by the AP server in order to manage a plurality of SQLs included in a request transmitted from the AP server to the DB server for each request.

  FIG. 8 is a diagram showing DB access log information in the present embodiment. The DB access log information is created by the DB server (for production) 104, collected by the output log collection program 601 of the output log collection device 108, and stored in the output log repository 109 by the output log management program 602 of the output log collection device 108. Saved. The DB access log information includes a request ID 801, contents 802, a table 803, a type 804, and an access time 805. A content 802 indicates the processing content of the request. A table 803 shows a request access destination table. A type 804 indicates the type of request processing. Update indicates, for example, a write process, and reference indicates, for example, a read process. The mapping between the request ID 801 and the content 802 of the DB access information is collected by cooperation between the DB server and the AP server.

  For example, item 1 in FIG. 8 indicates a DB access log in which the request ID 901 is AAABBB, the content 902 is DB access start, the table 803 is A, the type 804 is updated, and the access time 805 is 200909012754.

  FIG. 9 is a diagram showing test data information in the present embodiment. Real data input timing information extracted from the test data repository 107 by the test data extraction program 401 operating on the test data control device 205, and an output log repository by the output log extraction program 407 operating on the test data control device 205 The DB access log information extracted from 109 is created by organizing the information in units of input data by the input timing control program 402 operating on the test data control device 205 and stored in the test data repository 107. This database includes a session ID 901, a request ID 902, an input timing 903, a type 904, information 905, an access start 906, an access end 907, and a transmission destination 908. The time indicated by the input timing 1003, access start 1006, and access end 1007 is the elapsed time from the input timing (0 seconds) of item 1. Also, the access start 906 is delayed from the input timing because it takes time until the access is actually started after the request is input.

  For example, in item 1 in FIG. 9, the session ID 901 is 123ABC, the request ID 902 is AAABBB, the input timing 903 is 0 second, the type 904 is updated, the information 905 is A, the access start 906 is 2 seconds, and the access end 907 is 4 After 2 seconds, the transmission destination 908 indicates the test data of the AP server.

"Operation when collecting real data"
FIG. 10 is a flowchart illustrating real data collection processing in the present embodiment. This process is executed by the packet duplicating device 106 and the output log collecting device 108.

  In step 1001, the packet control program 502 starts collecting real data.

  In step 1002, the DB state of the DB (for production) 105 is copied to the DB (for test) 204. Specifically, the DB server (for production) 104 is connected to the DB server (for testing) 203, and the controller of the DB server (for production) 104 is stored in the storage device of the DB (for production) 105. Is transmitted to the DB server (for testing) 203. Then, the controller of the DB server (for test) 203 receives the data sent from the DB server (for production) 104 and stores it in the storage device of the DB (for test) 204.

  In step 1003, when the packet duplicating apparatus 106 receives a request from the user, the packet control program 502 confirms a real data collection end flag. As the method for setting the end flag, there are a method for determining whether to set the real data according to the collected time of the real data, the number of types of collected requests, the number of received data, a method for setting by the user (administrator), and the like. In this embodiment, the method for setting the end flag is not limited. If the end flag is not set, the packet control program 502 proceeds to step 1004 to continue collecting real data. If the end flag is set, the packet control program 502 proceeds to step 1010 to end the collection of real data. The request received from the user is, for example, an HTTP request, but is not limited to this.

  In step 1004, the packet duplicating device 106 receives a request from a user.

In step 1005, the packet replication program 501 of the packet replication device 106 replicates the request received in step 1004. Here, the request received from the user is called an original request, and the duplicated request is called a replica request.
In step 1006, the packet replication device 106 transmits the original request to the load distribution device 101.

  In step 1007, the packet control program 502 of the packet replication device 106 stores the replica request copied in step 1005 in the test data repository 107. In the present invention, test data is created using the replica request stored in step 1007.

  In step 1008, the packet replication device receives the responses returned from the AP server (for production) 102 and the AP server (for production) 103, and the packet control program 502 uses the session ID included in the response as a method for handling the session. Get by. However, the method for acquiring the session ID is not limited to this method.

In step 1009, the packet control program 502 updates real data input timing information stored in the test data repository 107.

  In step 1010, real data collection is terminated.

  In step 1011, the output log collection program 601 of the output log collection device 108 outputs the AP server (for production) 102, the AP server (for production) 103, and the DB server (for production) 104 when the original data is processed. The log is collected, the output log collected by the output log management program 602 is stored in the output log repository 109, and the DB access log information is updated.

"Operation when creating test data"
FIG. 11 is a flowchart showing the procedure of the test data creation process in the present embodiment. This process is executed by the test data control device and the output log collection device.
In step 1100, the test data input timing control program 402 of the test data control device 205 creates test data information by matching real data input timing information with DB access log information.

  In step 1101, the test data control device 205 deletes a DB access that causes an “access failure” such as dirty read or lock collision that occurs on the DB (for production) 105. The detailed deletion procedure will be described with reference to a flowchart showing a procedure for selecting a deletion candidate for a location where a lock collision or dirty read occurs in FIG. 12 and a flowchart showing a procedure for deleting a location where a lock collision or dirty read occurs in FIG.

  In step 1102, the test data input timing is shortened. The detailed shortening procedure will be described with reference to the flowchart of FIG. 14 showing the input timing shortening procedure.

  In step 1103, the controller of the test data control device 205 refers to the test data information and transmits test data to the AP server (for testing) 201 and the AP server (for testing) 202. Test data updated in 1101 and 1102 is used as test data.

  In step 1104, the test data input timing control program 402 of the test data control device 205 checks whether a dirty read or lock collision has occurred in step 1103. If a dirty read or lock collision occurs, the process proceeds to step 1105. If there is no dirty read or lock collision, the process proceeds to step 1107.

  In step 1105, the test data input timing control program 402 of the test data control device 205 updates the test data information so as to delay the input timing of the test data related to the DB access in which the dirty read or lock collision has occurred. The detailed procedure will be described with reference to the flowchart of FIG. 15 showing the input timing adjustment procedure.

  In step 1106, the test is executed again from the beginning.

  In step 1107, the output log collection program 601 on the output log collection device 108 collects and outputs logs output from the AP server (for test) 201, AP server (for test) 202, and DB server (for test) 203. The output log collected by the log management program 602 is stored in the output log repository 109.

  In step 1108, the output result receiving program 406 operating on the test data control apparatus 205 receives the output results from the AP server (for test) 201 and the DB server (for test) 203, and uses the output result for the output result. Save in the repository 206.

  FIG. 12 is a flowchart illustrating the procedure of access failure deletion candidate selection processing according to the present embodiment.

  In step 1201, the input timing control program 402 of the test data control apparatus 205 sets a request for issuing a DB access, which is an access to the same DB table in the DB access information, because the access time to the DB is duplicated. Search for. DB access is access executed based on the above-described request.

  In step 1202, the input timing control program 402 of the test data control device 205 sequentially selects the pairs searched in step 1201. The test data created varies depending on the order in which the group searched in step 1201 is selected. However, regardless of the order of selection, lock collisions and dirty reads do not occur, and test data with reduced test time can be created. If there is a set to be selected, the process proceeds to step 1203, and if there is no set to be selected, the process ends.

  In step 1203, the input timing control program 402 of the test data control device 205 checks whether the combination selected in step 1202 is a combination of “W, W”. Here, W indicates a write process, and DB access including update. That is, the selected DB access pair is a write access / write access pair. If the combination is “W, W”, the process proceeds to step 1204. If it is not a combination of “W, W”, the process proceeds to step 1209.

  In step 1204, the input timing control program 402 of the test data control device 205 receives the test access timing of the DB access having the later DB access out of the set selected in step 1202 and the DB access time of the two DB accesses. Delay so as not to. That is, the input timing is delayed so as to change the start time of one DB access with a later access start time to be after the end time of the other DB access. The time for delaying the input timing is, for example, the difference between the start time of one DB access and the end time of the other DB access. Then, the test data information is updated. However, test data information before update is also temporarily stored as a backup file. In this way, the occurrence of an access failure can be avoided by delaying the other request input time.

  In step 1205, when the input timing control program 402 of the test data control device 205 performs a test using the updated test data information, problems such as dirty read, lock collision, and session ID cannot be inherited between test data of the same session. Check if this occurs. In this embodiment, these failures are collectively referred to as access failures. For example, the updated test data is confirmed, and the DB access whose input timing is changed is compared with the access time of another DB access to confirm whether there is a possibility of a dirty read or lock collision. If the problem occurs, go to step 1206. If the problem does not occur, the backup file of the test data information before update stored in step 1204 is discarded, and the process returns to step 1202.

  In step 1206, the input timing control program 402 of the test data control device 205 uses the backup file of the test data information before update saved in step 1204 as test data information. Then, the test data information updated in step 1204 is discarded.

In step 1207, the input timing control program 402 of the test data control apparatus 205 adds the DB access set selected in step 1202 to the first deletion candidate.
In step 1208, the input timing control program 402 of the test data control device 205 refers to the test data information, and second deletes the DB access added to the first deletion candidate and the DB access having the same session ID. Add to a candidate.

  In step 1209, the input timing control program 402 of the test data control device 205 checks whether the combination is “W, R”. If it is a combination of “W, R”, the process proceeds to step 1210. If it is not a combination of “W, R”, the process returns to step 1202. The above R means DB access only for reference (read).

  In step 1210, the input timing control program 402 of the test data control device 205 allows only the reference (read) access included in the set of “W, R” and the DB access with the same session ID as the reference (read) access only. Find out if there is. If all DB accesses with the same session ID as the reference access are reference accesses, the process proceeds to step 1211. If the DB access with the same session ID as the reference access is not a reference access, the process proceeds to step 1212.

  In step 1211, the input timing control program 402 of the test data control apparatus 205 performs access only for reference (read) included in the set of “W, R” and reference access having the same session ID as the reference access, DB access Is deleted from the test data information. Here, the reference access having the same access time as the update access (W) and the reference access of the same session as the reference access are deleted from the test data without affecting the subsequent update access and reference access. Access failures such as dirty reads can be avoided.

In step 1212, the input timing control program 402 of the test data control device 205 adds the DB access set selected in step 1202 to the first deletion candidate.
In step 1213, the input timing control program 402 of the test data control device 205 refers to the test data information, and second deletes the DB access added to the first deletion candidate and the DB access having the same session ID. Add to a candidate.

  FIG. 13 is a flowchart illustrating a procedure of an access failure occurrence location deletion process according to this embodiment.

  In step 1301, the input timing control program 402 of the test data control apparatus 205 uses the second deletion candidate to convert the DB access included in the deletion candidates of both the first deletion candidate and the second deletion candidate created in FIG. Remove from.

  In step 1302, the input timing control program 402 of the test data control apparatus 205 determines that two DB access sets containing the same DB access are included in the same group with respect to the DB access set included in the first deletion candidate. Grouping according to the condition that For example, the group “W1, W2” and the group “W1, W3” are grouped into the same group. The group “W1, W2” and the group “W4, W3” are not grouped into the same group.

  In step 1303, the input timing control program 402 of the test data control apparatus 205 sets each DB access included in the second deletion candidate as one group. For example, when “R1, W4” is included in the second deletion candidate, “R1” and “W4” are groups.

  In step 1304, the input timing control program 402 of the test data control apparatus 205 selects the groups grouped in step 1302 and step 1303 in order. As for the selection order, the result does not change no matter what algorithm is used. If there is a group to be selected, the process proceeds to step 1305, and if there is no group to be selected, the process is terminated.

  In step 1305, the input timing control program 402 of the test data control apparatus 205 generates SQL from the update access included in the group selected in step 1304. SQL is a language for accessing the DB.

  In step 1306, the input timing control program 402 of the test data control device 205 checks whether the DB state can be correctly restored with the generated SQL. An example of a method for investigating whether or not restoration is possible is as follows. First, the input timing control program 402 of the test data control device 205 creates a DB table for restoration investigation in which the state of the DB table immediately before issuing the SQL generated on the DB server (for testing) is restored. Then, the DB table state after issuing the SQL is compared with the DB table state after executing the original update access group for the restoration investigation DB table. The investigation method is not limited to the above investigation method as long as it can be investigated whether or not the generated SQL can correctly restore the DB state. When performing exclusive access control etc. in the DB, the DB state may be correctly restored by the generated SQL even if there are multiple accesses with overlapping access times, so the DB state is correctly restored. Execute the process to check whether it is possible. If the generated SQL can correctly restore the DB state after the DB access from the DB state before the DB access of the group selected in Step 1304, the process proceeds to Step 1307, and the generated SQL cannot be correctly restored. If yes, go to Step 1308. As described above, when the test data having the overlapping access time can be restored by converting to SQL, it can be confirmed in advance that no access failure occurs. Furthermore, since the request input time is not changed, the test can be performed without affecting other test data.

  In step 1307, the input timing control program 402 of the test data control device 205 deletes all DB access of the group selected in step 1304 from the test data information, and instead registers the SQL generated in step 1305 as test data. .

  In step 1308, the input timing control program 402 of the test data control device 205 deletes all test data before (or after) the group selected in step 1304 from the test data information. This is because an access failure occurs in the DB before and after the group selected in step 1304.

  FIG. 14 is a flowchart showing the procedure of the input timing shortening process in this embodiment.

  In step 1401, the input timing control program 402 of the test data control device 205 selects each DB access of test data in the order of access. If there is an unselected DB access, the DB access is selected and the process proceeds to Step 1402, and if there is no unselected DB access, the process ends.

  In step 1402, the input timing control program 402 of the test data control device 205 acquires the processing time in the AP server of the test data related to the DB access selected in step 1401.

  In step 1403, using the processing time in the AP server acquired in step 1402, the input timing control program 402 of the test data control apparatus 205 selects the DB access start time selected in 1401 and the previous DB access. The time interval with the end time is obtained and is set as T1.

  In step 1404, the input timing control program 402 of the test data control device 205 selects all DB accesses having the same request ID as the DB access selected in step 1401.

  In step 1405, the input timing control program 402 of the test data control device 205 performs the DB access selected in step 1401 and the DB access selected in step 1404, the DB access to the same table, and the DB access accompanied by the DB update. Among them, the DB access selected in step 1401 and the DB access closest to each DB access included in the DB access selected in step 1403 are selected.

  In step 1406, the access start time of each DB access selected in steps 1401 and 1404 by using the processing time in the AP server acquired in step 1402 by the input timing control program 402 of the test data control apparatus 205. And the time interval (T2) between the last update access end time to the same table.

  In step 1407, the input timing control program 402 of the test data control device 205 sets the access start time of each DB access selected in steps 1401 and 1404 for the smallest time interval of the time intervals T1 and T2. Just move forward. By shifting the time interval in this way, the test time can be shortened without causing an access failure.

  In step 1408, the input timing control program 402 of the test data control apparatus 205 confirms the input timing of each test data having the same section ID as the test data related to the DB access selected in step 1401, and the DB access end time. It is checked whether or not a session error that the ID cannot be inherited occurs. If a session error occurs, the process proceeds to step 1409, and if no session error occurs, the process returns to step 1401.

  In step 1409, the input timing control program 402 of the test data control device 205 performs a test so as to delay the input timing of the test data related to the DB access selected in step 1401 by a predetermined time so that a session error does not occur. Data input timing information information is updated.

  FIG. 15 is a flowchart showing the procedure of the input timing adjustment process in this embodiment.

  In step 1501, the test data input timing control unit 402 of the test data control unit 205 operates the test data control unit 205 as a pseudo browser, inputs the test data created in the test environment system, It is examined whether a problem such as a lock collision or a session error in which a session ID cannot be inherited between input data having the same session ID occurs. If the problem occurs, the process proceeds to step 1502, and if the problem does not occur, the process ends.

  In step 1502, the test data input timing control device 402 of the test data control device 205 selects the test data with the later DB access or input timing out of the DB access set or the test data set in which the above problem has occurred. To do.

  In step 1503, the test data input timing control device 402 of the test data control device 205 sets the test data input timing information information so as to delay the input timing after the test data selected in step 1502 by the time when the above problem occurs. Update. In step 1504, the DB is restored to the state before starting the input timing adjustment. In step 1505, the input timing adjustment is restarted.

"Test behavior"
FIG. 16 is a flowchart showing the procedure of the regression test process in this embodiment.
In step 1601, the test environment system is updated with items to be subjected to regression tests, such as patch application and AP server update.

  In step 1602, the test data input timing control device 402 of the test data control device 205 operates the test data control device 205 as a pseudo browser, and inputs test data to the test environment system.

In step 1603, the output log comparison program 404 and the output result comparison program 405 of the test data control device 205 use the output log / output result saved in step 1107 and step 1108 and the output log / output result obtained in step 1602. Compare with. For the above comparison method, an existing method is used.

  As described above, the optimum embodiment of the test data creation method and the test method using real data according to the present invention has been described with reference to the accompanying drawings, but the embodiment of the present invention is not limited to such an example.

DESCRIPTION OF SYMBOLS 101 ... Load balancer, 102 ... AP server (for production), 103 ... AP server (for production), 104 ... DB server (for production), 105 ... DB (for production), 106 ... Packet replication device, 107 ... Test Data repository, 108 ... output log collection device, 109 ... output log repository, 111 ... LAN, 107 ... test data repository, 109 ... output log repository, 201 ... AP server (for testing), 202 ... AP server ( Test), 203 ... DB server (for test), 204 ... DB (for test), 205 ... Test data control device, 206 ... Output result repository, 301 ... Computer, 302 ... CPU, 303 ... Memory, 304 ... Input Output I / F, 305... HDD, 306... ROM, 307... Network card, 308. 309 ... Keyboard, 310 ... Mouse, 401 ... Test data extraction program, 402 ... Input timing control program, 403 ... Test data transmission program, 404 ... Output log comparison program, 405 ... Output result comparison program, 406 ... Output result reception 407 ... Output log extraction program, 501 ... Packet reception program, 502 ... Packet replication program, 503 ... Packet transmission program, 504 ... Packet control program, 601 ... Output log collection program, 602 ... Output log management program

Claims (11)

A test method using a test device,
The test apparatus includes request input history information indicating input times of a plurality of requests to a production system including a first database and a first server, and a plurality of accesses to the first database corresponding to the requests. With the access history information indicating the access type and the access time, test data information including request input time, access time and access type of a plurality of test data is created,
The test apparatus refers to the test data information, includes first test data included in the plurality of test data and whose access type is update access, and an access start time later than the first test data. When the access time overlaps with the second test data whose type is update access, the test data information is updated so as to delay the request input time of the second test data by a predetermined time,
The test method, wherein the test apparatus inputs the plurality of test data included in the updated test data information to a test system including a second database and a second server. The test method according to claim 1,
The test method according to claim 1, wherein the predetermined time is a difference between an access start time of the second test data and an access end time of the first test data. A test method according to claim 2, comprising:
The test apparatus confirms whether an access failure to the second database occurs based on the updated test data information,
When the access failure occurs, the test apparatus restores the second database by an access language to the second database obtained by converting the first test data before the update and the second test data. Check if you can
When the second database can be restored, the test device replaces the first test data and the second test data included in the test data information with the access language,
The test method, wherein the test apparatus inputs the plurality of test data included in the replaced test data information to the test system. The test method according to claim 3, wherein
The test device refers to the test data information, is included in the plurality of test data, and third test data whose access type is update access, fourth test data whose access type is reference access, If the access times of the test data overlap, the access type of other test data included in the fourth test data session is confirmed,
When the access type of the other test data included in the fourth test data session is a reference access, the test apparatus is included in the fourth test data and the fourth test data session. The test method, wherein the other test data is deleted from the test data information. The test method according to claim 4, comprising:
When the access type of the other test data included in the fourth test data session includes update access, the test apparatus stores the third test data and the fourth test data before the update. Check whether the second database can be restored by the access language to the converted second database,
When the second database can be restored, the test apparatus replaces the third test data and the fourth test data included in the test data information with the access language,
The test method, wherein the test apparatus inputs the plurality of test data included in the replaced test data information to the test system. The test method according to claim 1,
The test apparatus selects test data in the order of input to the test system from a plurality of test data included in the updated test data information,
The test apparatus calculates a first time interval between the access time of the selected test data and the access time of the immediately preceding test data;
A test method, wherein the request input time of the selected test data is advanced by a first time interval. The test method according to claim 6, comprising:
The test apparatus selects other test data included in the selected test data request,
The test apparatus calculates a second time interval between the access time of the other test data and the access time of the test data immediately before accessing the storage area of the database accessed by the other test data,
The test apparatus characterized in that the request input time of the selected test data and the other test data is advanced by a first time interval or a second time interval. The test method according to claim 7, comprising:
The test apparatus characterized in that the request input time of the selected test data and the other test data is advanced by a short time interval of the first time interval or the second time interval. The test method according to claim 1,
The test apparatus inputs the plurality of test data included in the updated test data information to the test system, and when an access failure occurs, a test data request input in which the access failure occurs A test method characterized by delaying time. A test device,
An interface connected to a production system including a first database and a first server; and a test system including a second database and a second server;
Created by request input history information indicating input times of a plurality of requests to the production system, and access history information indicating access types and access times of a plurality of accesses to the first database corresponding to the requests. A memory for storing test data information including a request input time, an access time, and an access type of a plurality of test data;
Referring to the test data information, the first test data included in the plurality of test data and whose access type is update access, and the access start time is later than the first test data, and the access type is update access. If the access time overlaps with certain second test data, the test data information is updated so as to delay the request input time of the second test data by a predetermined time, and the update is performed on the test system. And a processor for inputting the plurality of test data included in the later test data information. The test apparatus according to claim 10,
The test apparatus according to claim 1, wherein the predetermined time is a difference between an access start time of the second test data and an access end time of the first test data.

Images