[Original] loadruner FAQs

[Original] loadruner FAQs

Why does the IE browser not pop up when 1loadrunner records the script?
A: Start the browser, open the Internet option dialog box, switch to the Advanced tab, remove the check box for "enable third-party browser extension (restart required)", and run vugen again to solve the problem;

2 when LoadRunner records the script, why does the system prompt that the default browser does not support the solution?
A: In recording options-> browser-> modify browser settings-> change to IE browser access. Restart the LoadRunner recording script to OK;

3. What should I do if garbled characters exist in the generated web script during recording?
A: Before recording the script, open the recording option configuration dialog box record-options, go to the Advanced tab, Check support charset, and select support UTF-8. If you record the file, no Chinese garbled characters will occur.

4 what is the difference between HTML-Based Script and URL-Based Script?
A: The recording script in "HTML-based script" mode is used. vugen generates separate steps for each HTML operation. This script looks intuitive; when a script is recorded in URL-Based Script Mode, vugen can capture all HTTP requests sent to the server as user operation results, and then generate corresponding methods for each user request.
Generally, browser-based Web applications use the "HTML-based script" mode to record scripts; without browser-based Web applications, web applications include Java applets that interact with servers, browser-based applications include JavaScript/VBScript code that communicates with servers, and browser-based the application uses the HTTPS security protocol, the "url-based script" mode is used for recording.

5. Why is the check method web-find added to the script but not executed during script playback?
A: LoadRunner disables text and image check by default. Go to the "run-time setting" dialog box, go to "Internet Protocol → Preferences", and select the "enable image and text check" option under checks.

Note: The web_reg_find function is recommended for text and image check.

6. What is the main impact of the pacing settings during running?
A: pacing is mainly used to set the interval of repeated iteration scripts. There are three methods: start immediately after the end of the last iteration, wait for a fixed time after the end of the last iteration, and start executing the New iteration at a fixed or random interval. Set iteration as needed. Generally, there is no time interval, which will lead to greater pressure.

7. If "Enable Logging" is not checked in the log label set during runtime, can a manual message be sent?
A: The Enable logging option only affects Automatic logging and messages sent through lr_log_message. Even if it is not checked, if lr_message, lr_output_message, and lr_error_message are used in the virtual user script, the messages sent by the virtual user script are still recorded.

8. How does LoadRunner record scripts in IE7 + win2003 environments?

Answer: Method 1: Use the IE6 kernel recording script
In the system installation directory c: \ windows \ IE7 \ iexplore.exe, and in recording options-> browser, specify the IE6 kernel. After the setting is confirmed, you can use LoadRunner to record the script;

Method 2: Check the solution provided by LoadRunner:
Install LoadRunner 8.1 Feature Pack 4-> then install: Internet Explorer 7 (ie 7) Support for LoadRunner 8.1 Feature Pack 4;
9 LoadRunner vugen 8.0 is normal when recording web services protocol scripts, but an error "error: server returned an incorrectly formatted soap response" is reported during playback ".?
A: The reason is that the LoadRunner 8.0 vugen has a defect in recording the Web service protocol Script: if the server's operating system is Chinese, vugen will automatically change the header of the WSDL file to <? XML version = "1.0" encoding = "zh_cn"?>, Therefore, the above error message is displayed.
Therefore, you need to install the following patches: “lr80webservicesfpi_setup.exe”and “lrunner_web_sevices_path_1.exe ".

10 does vugen support Netscape client certificates?
A: No. Currently, vugen 8.0 only supports client certificates of Internet Explorer. When recording a script, you can first export the required certificates from Netscape, then import them to Internet Explorer, and ensure that these certificates are exported and imported in the same order. In addition, the previous process must be repeated on each computer that will record or run the Web vuser script that requires the certificate.

11 when the LoadRunner scenario is executed, the following error occurs: Missing newline in D: \ test \ test1.dat. When the LoadRunner scenario is executed, no error is reported?
A: When the loadruner parameter sets the test1.dat text, you need to press enter to confirm the last parameter.

12 an error occurs during execution in the LoadRunner scenario: "load generator is currently running the maximum number of vuser of this type"
A: By default, the maximum number of concurrent users in loadruuner scenarios is 1000. Therefore, you must set load generator> details> vuser limits> Other vusers to replace the parameter value, for example, 10000; of course, your serial number is supported. At present, the maximum serial number of 6.2w is supported. Sorry!
13 will vugen modify the proxy server settings in the recording browser?
A: Yes. When you start recording a web vuser script based on a browser, vugen starts the specified browser. Then, vugen instructs the browser to access the vugen proxy server. Therefore, vugen modifies the proxy server settings on the recording browser. By default, vugen immediately changes the proxy server settings to localhost: 7777. After recording, vugen restores the original proxy server settings to the recording browser. Therefore, you cannot change the proxy server settings during vugen recording.
14 how does the LoadRunner Script output the current system time?
A: LoadRunner provides the char * ctime (const time_t * Time) function. The call parameter is a long integer pointer, which is used to store the value of the returned time.
15 loadruner no problem occurs immediately after recording some Web virtual user scripts. However, when the number of iterations is set to greater than 1, only one iteration can be successful for playback. An error occurs from the second iteration?
A: In the "Browse emulation" setting of "Run-Time Setting, the "simulate a new user on each iteration" option and the following option "Clear cache on each iteration" are selected to simulate a new user and clear cache.
16 what is the difference between the running mode of threads and processes in "Run-Time Setting" in LoadRunner?
A: If "Run vuser as a process" is selected, a process is created for each virtual user during the scenario runtime. If "Run vuser as a threadpattern" is selected, each virtual user is run in one thread, and only one mmdrv.exe is displayed in the task manager. This method is more efficient, it can cause more pressure, which is the default option.
In addition, if IP spoofing is enabled, select "EXPERT mode" under the Tools menu in the controller ", then, set the IP Address Allocation Method in the Options> General tab under the Tools menu to the same as that in the vuser running mode, which is the same as in the thread or process mode.

17 when running web-related test scenarios in the controller, many timeout errors are often prompted. How can this problem be solved?
A: This is mainly caused by the default timeout settings of the script. When a web script is played back, a timeout error occurs sometimes because the server response time is too long. Modify the runtime configuration of the script.
After entering the "run-time setting" dialog box, go to "Internet Protocol → preference ". Then click "OPTIONS ..." Button to go to the Advanced Settings dialog box, you can modify the default values of various timeout settings.

18 why are there still plenty of resources in windows, such as CPU and memory, but the number of simulated users cannot go up?
A: In Windows computer Standard settings, the default Operating System limit can only use hundreds of vusers. This limit is irrelevant to the CPU or memory, this is mainly because the operating system specifies the maximum number of threads by default. To break the windows limitation, you must modify the Windows registry. Take Windows XP Professional as an example.
(1) Open the registry and enter the following keywords in the Registry key HKEY_LOCAL_MACHINE: System \ CurrentControlSet \ Control \ Session Manager \ subsystems.
(2) Find the Windows keyword. The windows keyword is as follows:
% SystemRoot % \ system32 \ csrss.exe bjectdirectory = \ WINDOWS
Sharedsection = 512, Windows = on subsystemtype = Windows serverdll = basesrv, 1
Serverdll = winsrv: userserverdllinitialization, 3 serverdll = winsrv: conserverdllinitialization, 2
Profilecontrol = off maxrequestthreads = 16
Sharedsection = 512, keyword format: xxxx, yyyy, zzz. XXXX defines the maximum value of the system range heap (in KB), and YYYY defines the size of each desktop heap.
(3) change the settings of yyyy from 3072 to 8192 (8 Mb) and add the sharedsection parameter value.
By changing the registry, the system allows more threads to run, so that more vusers can run on the computer. This means that the maximum number of concurrent users that can be simulated is not limited by the Windows operating system, but only limited by hardware and internal scalability.

19. The number of concurrent users is set in the controller, but why is the number of initialized users less than the actual number?
Answer: set the problem mainly. In tools → options → run-time setting, you can set the maximum number of virtual users that can be initialized each time. If you need 100 concurrent users, set this value to a value greater than 100. In addition, pay attention to the update of the license of the LoadRunner protocol to ensure that the license can be used to allow the number of concurrent users required.

20. How can I enable the user in the scenario to continue running due to an error to ensure uninterrupted stress testing?
A: Use vugen to open the virtual user script. after entering the "run-time settings" dialog box, go to "General> miscellaneous ", you can see the "error handling" configuration in miscellaneous settings. Check "continue on error" to enable the virtual user to continue running due to an error.

17. Why can't. Net Virtual users be executed on a remote host sometimes?
A: The major issue is the version of LoadRunner. Based on my experience, if it is a virtual user developed by Microsoft Visual Studio 2005 and the LoadRunner Client Version is earlier than 8.1, an error will occur on the host that runs the controller.
Therefore, to correctly run a. net virtual user developed by Microsoft Visual Studio 2005, it is best to install a client version later than 8.1, while a controller host can install both versions 8.0 and 8.1. In addition, the. NET runtime environment should be installed on the LoadRunner client that generates pressure in advance. If Microsoft Visual Studio 2005 develops a. net virtual user, Microsoft. NET Framework SDK V2.0 should be installed.

18. In the test analysis result, the action time is counted. In fact, this data may not be needed. How can I only Display User transactions defined by myself?
A: Enter the runtime settings of the script, and choose General> miscellaneous. By default, there are two options under the automatic transaction configuration "Automatic transactions": the first is to define the action part of the script as a transaction; in the second case, each part of the script is defined as a transaction. After the two check items are removed, only user transactions defined by the user are displayed in the test result.

19. In the test results, why are the maximum, average, and minimum values of each transaction in the summary and average transaction response time distinct displayed?
A: It is mainly affected by the sampling time. The average transaction response time in summary is calculated based on the data obtained during the execution of the entire scenario. The maximum and minimum values are also obtained from the entire scenario. The average transaction response time graph mainly obtains the maximum and minimum values of the Transaction Response time according to the sampling frequency analyzed by LoadRunner, and then calculates the average value.
You can use "set granularity" to modify the sampling frequency of the average transaction response time graph. If "granularity" is set as the scenario execution time, the statistical results are consistent.

20. When the total clicks in the statistical results are total hits, does the user's mouse clicks?
A: From time to time, total hits is calculated based on the number of users' mouse clicks, but based on the total number of requests initiated by each virtual client to the backend. For example, if a page requested from the server contains five images, you can click the mouse to access the page, the number of clicks that a single virtual user accesses on LoadRunner is 1 + 5 = 6.

21. Some Web Test Result Analysis charts (such as the number of pages returned per second) cannot be seen in the test result analysis diagram. How can I configure them?
After opening the virtual user script with vugen, go to the "run-time settings" dialog box and go to "Internet Protocol> preference" in sequence. You can see some Web performance diagram configurations.
After the above options are selected, the Controller will generate data during the test execution, and then you can view the corresponding performance result analysis diagram in the analysis.
, Step download timeout (120 seconds)
This is a common problem. To solve this problem, follow these steps:
1. Modify the request timeout in run time setting to 600 s. Three parameters can be modified at one time, namely http-request connect timeout, HTTP-request receieve timeout, step download timeout, which is recommended to be modified to 600, 600, and 5000 respectively. After the Run Time Setting settings are complete, remember to set the corresponding parameters in the run time setting of the control component option;
2. If solution 1 fails, the solution is as follows:
In the advaced area of Internet protocol-preferences in runt time setting, there is a winlnet replay instead of sockets option, and the option is successfully played back. Remember that this method only applies to Windows systems. This method comes from Zee.

22 Problem description Connection reset by peer
This problem is not frequently encountered. Generally, the download speed is slow, resulting in timeout. Therefore, you need to adjust the timeout time.
Solution: in the run-time setting window, Set set advanced options (set advanced options) and reset "http-request connect timeout (SEC ), it can be slightly larger ";

23 Problem description connection refused
The cause of this error is complicated. It may be very simple and may need to be checked in several places. Different Operating System methods are also different;
1. First, check whether most of the services connected to WebLogic have been rejected. You need to monitor the connection Wait Status of weblogic. In this case, you need to add the acceptbacklog and increase it by 25% each time to see if the solution is solved, at the same time, you also need to increase the connection pool and adjust the number of execution threads. The value of (number of connection pools * Statement cache size) should be smaller than or equal to the maximum number of Oracle database connections;
2. If method 1 does not change, check whether the number of connections is limited in the server operating system. In Aix, you can directly modify the number of connections in the VI file limits, there are also TCP connection wait time intervals, similar to wiodows, but wendows modifies the Registry. For specific modification methods, check the manual. The registry contains the tcpdelaytime entry;

24. Problem description: Open Program Files
The problem usually occurs when the pressure is high. Because the server or Application Middleware has a maximum limit on the number of opened files, the solution is as follows:
1. modify the number of files in the operating system. Modify the nofiles restrictions under limits under Aix, increase the number or set it to no limit, and try to modify all the involved servers;
2. If method 1 cannot be solved, check the application server WebLogic commonenv. sh file, modify the number of max-nofiles files in the nofiles file to increase, you should be able to pass, specifically find the nofiles method, modify the execution body of the else condition, and increase the number of opened files; remember to back up this file before modification to prevent modification errors;

25. Problem description: has shut down the connection prematurely
Generally, when you access the application server, both the large number of users and the small number of users will appear;
Explanations from the Internet:
1> application access dead
Small users: Program problems. Database problems in the program
2> application service is not dead
Application Service parameter settings
For example:
Many clients are rejected to connect to the WebLogic application server. If no error is displayed on the server side, the value of the acceptbacklog attribute of the server element in WebLogic may be too low. If the connection refused message is received during the connection, the value should be increased by 25% each time.
Java connection pool size settings, JVM settings, etc.
3> database connection
The performance parameter of the application service may be too small.
Maximum number of connections started by the database (related to the Hardware Memory)
The above information has some reference value. You can refer to this type of debugging in actual situations.
If it is a small user: A program problem. If there is a database problem in the program, you must use more professional tools to capture the problematic program, mainly the SQL statements with low execution efficiency in the program. WebLogic can use introscope to locate the problem, during this period, you can check whether JVM garbage collection is normal. In practice, when I have concurrent users of 500 and 600, I have experienced changes in the JVM jagged model, and the increase and decrease are very fast, this should be abnormal;

26. Problem description: failed to connect to server
This problem is generally caused by a failure of connecting the client to the service due to two client connection restrictions (that is, the stress-loaded machine). If one network delay is serious, the solution is as follows:
1. Modify the tcpdelaytime Registry Key of the Server Load balancer to a smaller value;
2. Check the network latency to see what the problem is;
In order to reduce this situation, it is recommended that you complete the test before the test to ensure a clean network environment. The number of stress testing users for each server load is not too large, and the number of users for each server Load balancer should be averaged as much as possible, in this way, the probability of the above problems is very small.

27. Problem description: overlapped transmission of request to... wsa_io_pending
Solution to this problem:
1. Method 1: Add web_set_sockets_option ("overlapped_send", "0") to the script to disable ttfb subdivision. The problem can be solved, but the ttfb subdivision graph cannot be used any more.
2. Method 2: Increase the connection pool and application system memory by 25% each time.

28. Problem description: deleted the current transaction... since response time is not accurate
This problem is rare. Generally, the ping value on the stress machine is negative (AMD dual-core CPU). You can restart the PC or patch it, as shown in the figure below.
 
29. Description: HTTP status-code = 500 (internal server error)
1. When the application service fails, restart the application service.
2. When the available memory of the application system is below the threshold, the probability of HTTP status-code = 500 is very high. In this case, you only need to increase the memory of the application system.

30. Problem description: failed to transmit data to network: [10057] socket is not connected
This error is caused by network reasons. The same LoadRunner 9.0 is installed on both pC1 and PC2, if the same number of virtual users are used to run the same service (the other conditions on the machine are the same), a small number of users on pC1 report errors, and all users on pc2.
 
31. Problem description: Error-27257: Pending web_reg_save_param/reg_find/create_html_param [_ ex] request (s) detected and reset at the end of iteration number 1
Solution: The web_reg_save_param location is incorrect. It should be placed in front of the Request page.

32. Problem description: an error is reported when the remote proxy is called through controler. Error: CCI security error: You are running under secure mode and the function system is not allowed in this mode.
Solution: remove the firewall check box when the proxy is enabled.