What is Performance Testing?
Performance Testing on a software application is done to determine how speed, stability and scalability the application is. Speed refers to how fast the software application will respond, and stability refers to how the application handles varying workloads. Scalability is the total amount of workload the software application can handle.
What is Load Testing?
It is done to understand how varying expected loads would make the system behave. For instance, its behavior if they were many users concurrently using the same application to carry out critical transactions for a certain period of time. It is able to give the application's response time for all the important transactions in a business. Load testing can be done for a business's database and servers which will be very helpful in identifying any existent bottlenecks.
What is Stress Testing?
This is done to determine whether the software application can still perform well even in cases of extreme load. It helps the application administrators know to what extent the application can be stretched without creating bottlenecks.
What is Soak Testing?
Soak testing monitors the system's performance over time. It is also referred to as endurance testing. It aims at determining the system's capability to sustain the normal load over time. The test involves looking out for potential leaks by monitoring memory utilization, as well as performance degradation. Performance degradation refers to the system's response time over a prolonged period of sustained use. Does it remain constant, improve or decline? Results of this test are derived from application of a significant amount of load to the system for a particular period of time.
What is Spike Testing?
This test involves a sudden increase in the amount of load, for instance a sudden increase in the number of users making transactions using the application. It is meant to determine whether the system will crash, slow down or whether it will be able to handle the surge.
What is Isolation Testing?
This is usually a repeat of the test execution that had caused a problem in the system. It is done to confirm the domain in which the fault had occurred.
What is Volume Testing?
It is a non-functional test. A certain amount of data is used to test the software application. The data can be a database or an interface file like .dat or. xml. To volume test the database, expand it to the desired size and then test the application for it. If you need the application to interact with a specific interface file, get a sample of the size you desire and then run the application.
What is Failover Testing?
Failover testing is done in order to determine if a server fails during the heavy load: Is load balancer able to react quickly and is able to drive the traffic to another web server/application server. Is the other web server/application server able to handle the unexpected load. This test helps to determine how the server behaves during these type of scenarios.
Explain the performance test life cycle that you follow?
- Gathering the requirements from client
- Understanding the architecture/environment of the application
- Planning and Designing the tests
- Preparing and enhancing the scripts
- Executing the tests in controller
- Analyzing the results
- Fine tuning the application
- Retesting the application after fine tuning the application and repeat this process until we meet the SLA (Not 100% up to 90% at least).
- The architecture of the application and the technologies used to develop the application?
- What is a web server, application server and database server?
- What are the browsers and Java versions supported by the application?
- What are the different load balance techniques used?
- What are the protocols used between client and server? Ex: HTTP, HTTPS, FTP etc.
- What are the Business scenarios we need to test?
- What are the preferred tools for performance testing and monitoring? Ex: Loadrunner, Jmeter, Sitescope and Wily Introscope.
- What is the workload model and identifying the workload model?
- SLA of the application in terms of response times and hardware resource utilization.
Yes.
What are the contents in your Test Plan/Test Strategy?
It's always better to go through the test plan before the interview, it helps a lot. Some of the Test Strategy contents are Introduction, Purpose of The Document, Scope Of The Document, Application Architecture And Topology, Performance Testing Requirements And Dependencies, Objectives And Goals Of Performance Testing, Scope Of Performance Testing, In Scope, Out Of Scope, Application Performance Scenarios, Workload Characteristics, Application Test Environment, Performance Test Tools, Performance Test Dependencies, Hardware Dependency, Software Dependency etc.
What is little’s law?
Tell me some of the performance testing tools that you use in your company?
Loadrunner, Jmeter, Sitescope, HP Diagnostics, Wily Introscope, Dynatrace and Java Profiler.
What is performance tuning?
Performance tuning is done during the process of performance testing. Issues will be identified during the different types of performance testing. Performance tuning will be done based on the issues. It might be a software / hardware tuning. Usually Bottlenecks are identified in application and database servers. Proper tuning of the servers, code and SQL queries resolves the issue. Sometimes changes in the hardware resources will also help.
What is the difference type of tests you conducted during testing?
Load Testing, Stress Testing, Failover Testing etc.
Loadrunner General Questions:
What is Loadrunner?
Loadrunner is a performance testing tool. It is developed by Mercury and acquired HP. It has versions Loadrunner 7 to Loadrunner 12.
What are the Loadrunner components?
Vugen, Controller, Load Generators, Analysis, Shunra.
What are the Loadrunner versions used?
Loadrunner 9.52, Loadrunner 11.3, and Loadrunner 11.5 and Loadrunner 11.52.
Vugen And Scripting Questions:
How do you start the script?
We usually start the scripting after understanding of Business Critical Scenarios.
What are the various functions you have used while creating the scripts?
You should able to tell at least 10 loadrunner functions and different functions that you have used in different protocols.
What is Step download timeout error?
If the application does not respond in 120 seconds, step download timeout occurs in loadrunner. We can increase the step download timeout error in Loadrunner (Runtime Settings -> Preferences -> Options -> General -> Step Download Timeout.
What are the different type of parameters available?
Custom, Date/Time,File, Group Name, Iternation Number, Load Generator Name, Random Number, Table, Unique Number, User Defined Function and Vuser ID.
What is advanced trace?
It logs all the vuser messages and function calls.
What are the Logs available in LR?
Code Generation, Replay Log
What are the protocols you worked on?
Tell the protocols that you know and explain how good you are in that protocols.
Difference between Think Time & Pacing?
- Think is time interval between two transactions.
- Pacing is the time interval between two iterations.
- Parameterization is giving the input through a parameter file in loadrunner.
- Correlation capturing a required value from the server response in loadrunner. Example: SessionID.
We can run the specific part of the script by using blocks in run time settings -> General -> Run Logic.
What are different recording modes available in LR?
We have mainly two recording modes HTML mode and URL mode.
What is the difference between HTML and URL mode?
You can go through this link for difference between HTML and URL mode.
What are the advanced features of URL recording mode in LR 9.5?
We have mainly two features:
- Create concurrent groups for resources after their source HTML page.
- Use web_custom_request only.
What is the difference between resource=0 and resource=1?
Here www.msn.com Resource is 0 as it is the main URL and the corresponding resources like image resource is 1.
What is the difference between lr_abort and lr_exit?
Lr_abort: it stops the execution of the action part and executed the vuser_end section.
Lr_exit: it exits from the particular action, script or iteration and there are different options with lr_exit;
Example:
lr_exit (LR_EXIT_VUSER, LR_FAIL);
lr_exit (LR_EXIT_ACTION_AND_CONTINUE, LR_FAIL);
What is the difference between lr_eval_string and lr_save_string?
Lr_eval_string: It extracts a value from a given parameter.
Lr_save_string: It saves c variable into loadrunner variable.
What is context less and context based?
Context less mean the request does not depend on the previous request. Context based means the request is dependent on the previous request.
What is the use of Save Count in web_reg_find?
Number of times the given text has found in a particular page.
What is Correlation and its function with attributes?
Capturing a value from the server response. The function name is web_reg_save_param. List of attributes are Step Name, ParameterName, LeftBoundary, RightBoundary, ORD, NotFound = Warning/Error, Savelen, Saveoffset.
How do you do auto correlation?
From Loadrunner 11.5 we have a beautiful option called Design Studio. The loadrunner will detect the dynamic values and we can do the correlation easily.
If source is not available, then how do you do correlation?
We have different option to do correlation, Generation Log, Replay Log.
Difference between “Save Offset” & “ Save Length” and what are test flags?
Save offset is used to ignore the number of letters and Save Length is used to save the length of the parameter.
When do you use Web_Custom_request?
The name of the function itself indicates the custom request. Web_custom_request is the last option that vugen prefer during creation of the script. It first tries to create the script using different web functions. This function is used by the vugen when it cannot interpret the requests with the other web functions.
Dynamically the post data is changing, what do you do?
In this situation, I will use web_custom_request and use the for loop to append the dynamically changing data to web_custom_request.
What are check point available in LR?
web_find, web_reg_find, web_image_check.
Do you write any user defined functions?
Make sure you prepare well for this question.
What are the run time settings you have used in your script?
We use the run time settings based on our requirements. Iteration, Run Logic, pacing thinktime etc.
What are the challenges you have faced during script generation & execution?
Prepare well for this question, you should be able to answer the question clearly.
What is the difference between get and post request?
Get Request
|
Post Request
|
We can use up to 2048 character length.
|
It has no restrictions
|
Get request is less secure.
|
Post request is secure.
|
It can be cached
|
It cannot be cached.
|
It can be bookmarked
|
It can not be bookmarked.
|
What is the difference between web_url and web_link?
web_url:
- It's a GET request
- The web_url function does not require a context
- web_url is recorded only when VuGen is in either URL–based recording or in HTML–based recording mode.
web_link:
- web_link function is an action function that emulates a mouse click on the link that is defined by the attributes.
- The web_link function can be executed only in the context of a previous operation.
- web_link is recorded only when VuGen is in HTML-based recording mode.
What is the difference between web_submit_data and web_submit_form?
Web_submit_form
|
Web_submit_data
|
It is a context based submission
|
It is a context less submission
|
It is recorded only html based
|
It is recorded both in html and url mode.
|
What is the difference between web_submit_data and web_custom_request?
Web_submit_data
|
Web_custom_request
|
It is a context less submission
|
It is also context less submission.
|
We cannot write a custom request.
|
We can write a custom request using this function.
|
We can send through get or post request
|
We can send only post request
|
What is the difference between web_find and web_reg_find?
Web_find deprecated and we have only web_reg_find now.
What is the difference between lr_output_message and lr_error_message?
Lr_output_message
|
Lr_error_message
|
The lr_output_message sends message to the output window, log files (such as the Vaughan log file and the Application Management Web site and agent log file), and other test report summaries. When a script is run in VuGen, the output file is output.txt.
|
The lr_error_message function sends an error message to product output windows (such as the LoadRunner output window), log files (such as Vugen and the Application Management agent log file), and other test report summaries. For details regarding the output for each product, see the product's user guides.
|
If Run-time settings > General > Miscellaneous > Fail open transactions on lr_error_message is selected, calling lr_error_message will fail the transaction in addition to sending the message
| |
To send a message to the output file, you must enable logging in the run-time settings, and select Always send messages. If you select Send messages only when an error occurs, there is no output from this function.
|
VuGen displays the message text of the lr_error_message function in the Execution log in red, using Error code 17999. Note that this function sends a message to the output even when logging is disabled in the run-time settings.
|
Controller And Analysis Interview Questions And Answers:
Difference between Scenario & Group, Real world Schedule & Basic Schedule?
Difference between Performance Center & Controller?
HP Loadrunner
|
HP Performance Center
|
It is a standalone application
|
It is a web based application.
|
Only single load test is executed at a time.
|
Here we can execute multiple load tests.
|
We cannot trace the requirements and defects
|
Here we can trace the requirements and defects
|
We cannot reserve time slots for machines
|
Here we can use timeslots to reserve the machines
|
Loadrunner is good if we are using from one location.
|
Performance Center is good if we are operating from different locations.
|
How do you come to know which Vuser failed?
We can check in the vuser status message.
What is Rendezvous point?
When lr_rendezvous function is called, all the users wait at that point and executes the next step.
What is IP Spoofer in loadrunner?
We can create the IP addresses dynamically using IP Spoofer.
Can you configure load generators in Windows & Unix?
Yes, we can configure the load generators in both Windows and Unix.
Can we run different scripts by using Performance Center and the Controller?
Yes, we can run.
Analysis
What are the default counters that loadrunner analysis file shows?
Vusers, Hits Per Second, Throughput, Average Transaction Response Time, Transactions per second.
What are the Operating Systems counters you monitor?
CPU, Memory, Disk, Network
Give me 5 server counters that you regularly check?
Response times, throughput, hits per second, CPU, Memory, Heap Memory.
What AWR Reports? From which version of the Oracle – AWR feature is added?
AWR stands for Automatic workload Repository. It is introduced from Oracle 10g.
What is Granularity?
Granularity is the time interval of sampling the data in loadrunner analysis.
Difference between Thread & Process?
Thread
|
Process
|
Threads are dependent on each other in a process
|
Process are independent of each other
|
Threads uses very less resources
|
Process uses very high resources
|
Threads shares the address space
|
Process requires a separate address space
|
Threads carry very less information when compared with process.
|
Process carry more information
|
What is JVM & Unix?
JVM stands for Java Virtual Machine which is a platform independent and converts Java byte code to machine language. As it is platform independent we can execute the code on any operating systems.
Unix is a multiuser operating system and has many flavors like Linux, Sun Solaris, AIX, HP Unix, Ubuntu etc.
What is Memory Leakage?
What is Garbage Collection?
What are the algorithms available in GC?
What is Heap memory?
How do you monitor server counters?
We use Sitescope or NMON tool and different commands available in Linux VMSTAT, IOSTAT, TOP etc.
What is Table Scanning?
When there are no Proper indexes given in the database tables, some queries will scan the entire data in the tables and collect the results which will increase the response time.
What are context switches?
What is a deadlock?
Deadlock occurs when two or more processes/threads waiting for the other to finish and it never happens and this leads to deadlock. This situation will apply to hardware/software, Java and databases.
What do you suggest to your client if there is a sudden hike in the CPU utilization while execution?
Sudden hike in the CPU utilization might have many reasons. We need to analyze what caused the high CPU utilization by using different tools and profilers.
What is Parsing, Soft Parse, and Hard Parse?
Parsing: It will check the syntax, misspelled SQL keywords and authentication.
Soft Parse: If the requested query is found in shared pool soft parsing occurs and it consumes less resources.
Hard Parse: If the requested query is not found in shared pool hard parsing occurs and it consumes very high resources.
What is the difference between Web Server, Application Server and Database Server?
Web Server
|
Application Server
|
Database Server
|
Ex: Apache, IIS and Nginx
|
Weblogic and Websphere
|
Oracle, My SQL, and SQL server
|
It handles HTTP requests.
|
It handles HTTP requests and it contains business logic.
|
It handles different queries, and records.
|
Resource utilization is low.
|
Resource utilization is high.
|
Resource utilization is high.
|
What is standard deviation in loadrunner?
Standard deviation in loadrunner is done with the total response times of a particular transaction.
For example, if you ran a transaction for 10 iterations. It will take the 10 transaction response times and calculates the Standard Deviation. If the standard deviation between these 10 values is high. Then it is a bottleneck, it should always low.
Examples:
Response times for login page with 5 iterations ->1, 1, 1, 1, 1=Standard Deviation is 0
Response times for a logout page with 5 iterations 1, 2, 3, 4, 5, 6=Standard Deviation is 2.5 which is high.
Is there any impact if we reduce “Think Time” up to 50%?
Yes, there will be impact as the load on the server increases if we decrease think time. Mainly we have an impact on response times, throughput and system resources etc.
At Last......
How to check network bandwidth is not sufficient?
Data Transfer rate between application server and database server or between client and the server is more than the bandwidth, then we can conclude the bandwidth is not sufficient.
Weblogic Top Counters for Performance Testing:
Thread Pool, DB Connection Pool, Garbage Collection, Java Heap memory, Load Balancing, Application logs, Hard Page Faults.
Oracle DBA Top Counters for Performance Testing:
In AWR Report: Shared Pool Memory, Cache Hit Ratios, Top 5 Events, Top SQL by CPU, Top SQL by I/O, Top wait events, Buffer pool cache hit ratios.
Operating System Counters are common for both Weblogic and Oracle DBA: CPU, Memory, Disk I/O, and Network.
At Last......
Every option in the Vugen, Controller and Analysis are important. The interviewer might ask any question on the tool. We should prepare for the interview in such a manner. Please go through the following settings once before you go for the interview.
- Run time settings
- Recording Options
- Controller Options
