Press "Enter" to skip to content

Category: Testing

Kinesis Data Generation

Published 2017-05-19 by Kevin Feasel

Allan MacInnis shows off a new data generation tool for Amazon’s Kinesis:

Amazon Kinesis Streams and Amazon Kinesis Firehose enable you to continuously capture and store terabytes of data per hour from hundreds of thousands of sources. Amazon Kinesis Analytics gives you the ability to use standard SQL to analyze and aggregate this data in real-time. It’s easy to create an Amazon Kinesis stream or Firehose delivery stream with just a few clicks in the AWS Management Console (or a few commands using the AWS CLI or Amazon Kinesis API). However, to generate a continuous stream of test data, you must write a custom process or script that runs continuously, using the AWS SDK or CLI to send test records to Amazon Kinesis. Although this task is necessary to adequately test your solution, it means more complexity and longer development and testing times.

Wouldn’t it be great if there were a user-friendly tool to generate test data and send it to Amazon Kinesis? Well, now there is—the Amazon Kinesis Data Generator (KDG).

Check it out if you’re using Kinesis and need to do some load testing.

Comments closed

Designing A Data Warehouse Test Plan

Published 2017-05-18 by Kevin Feasel

Koos van Strien walks through some of the high-level concepts when automating data warehouse tests:

In my current project, I’ve got a database containing everything to perform these tests:

  • Tables with identical structure to the ones in the staging area (plus two columns “TestSuiteName” and “TestName”)
  • A table containing the mapping from test-input table to target database, schema and table
  • A stored procedure to purge the DWH (all layers) in the test environment
  • A stored procedure to insert the data for a specific testsuite / name

When preparing a specific test case (the “insert rows for test case” step from the diagram above), the rows needed for that case are copied into the DWH:

Testing warehouses is certainly not a trivial exercise but given how complex warehouse ETL tends to be, having good tests reduces the number of 3 AM pages.

Comments closed

Pester For Presentations

Published 2017-05-17 by Kevin Feasel

Rob Sewell takes Pester to the edge:

If you have PowerShell version 5 then you will have Pester already installed although you should update it to the latest version. If not you can get Pester from the PowerShell Gallery follow the instructions on that page to install it. This is a good post to start learning about Pester

What can you test? Everything. Well, specifically everything that you can write a PowerShell command to check. So when I am setting up for my presentation I check the following things. I add new things to my tests as I think of them or as I observe things that may break my presentations. Most recently that was ensuring that my Visual Studio Code session was running under the correct user. I did that like this

Rob’s scenario is around giving presentations, but while reading this, think about those services which should be running on your SQL Server instance—the same concept applies.

Comments closed

Load Testing Kafka

Published 2017-05-05 by Kevin Feasel

Satish Bhor shows off Pepper-Box, a load generator which can stress test Apache Kafka:

Pepper-Box is a Kafka load generator application that can be used as a plugin for JMeter or standalone utility. It allows sending plain text Kafka messages (JSON, XML, CSV, or any other custom format), as well as Java serialized objects. Pepper-Box includes a template engine and random data generation function which helps to design message in any format. If we use it with JMeter then we can use all JMeter features. Pepper-Box is very useful in streaming analytics and data pipelines implementation, where input data format is tightly coupled with business problems.

Pepper-Box includes four main components.

I’m going to keep an eye on this tool.

Comments closed

Ignoring LoadGeneratorLocationError

Published 2017-03-09 by Kevin Feasel

Melissa Connors shows how to ignore LoadGeneratorLocationError errors in Visual Studio load tests:

I use Visual Studio for performance testing and overhead analysis with the SentryOne products. Currently, I have Microsoft Visual Studio Enterprise 2015 Version 14.0.25431.01 Update 3 installed. Since the first edition of 2015 (possibly even Visual Studio 2013), I’ve received a LoadGeneratorLocationError during each Load Test execution.

Since I am running the test locally, this error is noise. Furthermore, no one wants to see an error in an otherwise successful test. It simply ruins the final results report. In addition, when the Load Test was created, “On-premise Load Test” was selected, which makes this frustrating. Possibly more frustrating is that it’s called “On-premise” when you get started in the New Load Test Wizard.

Read on for the answer.

Comments closed

R-Hub

Published 2016-10-27 by Kevin Feasel

David Smith discusses a new service to test packages on multiple platforms:

If you’re developing a package for R to share with others — on CRAN, say — you’ll want to make sure it works for others. That means testing it on various platforms (Windows, Mac, Linux, and all the versions thereof), and on various versions of R (current, past, and future). But it’s likely you only have access to one platform, and installing and managing multiple R versions can be a pain.

R-hub, the online package-building service now in public beta, aims to solve this problem by making it easy to build and test your package on a variety of platforms and R versions. Using the rhub R package, you can with a single command upload your package to the cloud-based R-hub service, and build and test your package on the current, prior, and in-development versions of R, using any or all of these platforms

This looks like an interesting service for package developers and companies with a broad distribution of R installations.

Comments closed

Testing Analysis Services Cubes

Published 2016-10-26 by Kevin Feasel

Jens Vestergaard shows how to test Analysis Services cubes using a Visual Studio test project:

Unit testing in Visual Studio is actually not that hard and can save you a lot pain down the road. The testing framework in Visual Studio offers extensive ways of executing batches of tests. You can group tests by Class, Duration, Outcome, Trait or Project.

When you right-click a test, you get the option to select how you want the tests in the Test Explorer to be grouped.

If you have an Analysis Services cube, definitely read this—testing is a vital part of software development, and automating tests can save you significant time later.

Comments closed

Tests Are Hard

Published 2016-10-21 by Kevin Feasel

Cody Konior explains how testing is harder than it seems in the demos:

We should have 4 backup jobs

Definitely 4, not 3, and 5 is right out.

Unless of course someone builds a 500GB database with SQL CLR UDTs which slow down DBCC to a crawl, so you move the backups and maintenance for that database into their own jobs or steps with different flags, and possibly start staggering some stuff so it finishes on the weekend without dripping into Monday.

Then you’ve got more. But just kind of sweep them under the carpet. We have file system backup checks so that covers most of it…

These are just administrative test examples; when you have to start testing queries and procedures, it’s a whole new level of fun, as it feels like you’re building a castle on sand.

Comments closed

Unit Testing With SSDT And tSQLt

Published 2016-10-13 by Kevin Feasel

Steph Locke looks at adding tSQLt to an existing SQL Server Data Tools database project:

Whilst I won’t be showing code in this, there is a companion sample database project. This is on GitHub and each key stage is shown by a branch of work. This means you can jump in at most stages and work from there. If you need some git advice, check out my quick git in 5 presentation.

The core DB is a super hero themed database.

Database unit testing involves a lot of compromises (because most of what we care about as database professionals tends to be integration tests rather than unit tests), but having imperfect tests generally beats having no tests, so check this out.

Comments closed

Bugfixes To SQLCover

Published 2016-10-03 by Kevin Feasel

Ed Elliott notes that there is a new version of SQLCover out:

There have been a couple of fixes in SQLCover this week, kindly submitted by John Mclusky (https://github.com/jmclusky):

Code coverage not reported correctly for CTEs at the end of a stored procedure if the ‘with’ is immediately preceded with a semicolon

and

DeclareTableVariableStatement statements cannot be covered, so report falsely as missing coverage

Go check out SQLCover.

Comments closed