After having just started using Microsoft Release Management, I am more and more convinced that it is not well suited to run integration tests. This might be a false feeling I'm having, and I'd love to get more input on this. When we first considered it, I had the intention to run the tests defined in our test plan through it's pipeline, but now I'm seeing that we should be running those as frequently as possible. We would like to run integration testing every night, but our release candidates are only defined at the end of sprints, so using Release Management for that seems conflicting.
With the tool out of the equation, we are considering exploring the Lab Template again. We did some very minor tests with it a few months ago in a legacy project but never went too far. My main concern now is that both stages need deployment:
the Release Management pipeline needs to deploy our projects to the QA and production environment
the Lab Template also needs to deploy the project on a few virtual machines to run integration tests on
The Release Management uses some very nice abstractions to achieve that. You can code machine scopes and define components based on the drop folder structure to define each part of the whole application to be deployed. On the other hand, the lab management workflow does not support this (or perhaps I'm just missing it). The standard way to make deployment work for lab testing, is to write a custom power shell script that moves the files from the build drop folder to the correct places, creates the application pools for web projects, and stuff like that, all by hand.
Ideally, I'd like to just share the entire deployment workflow between both tools and, since the Release Management way of doing it seems much simpler, I'd use that. This would make it easier to maintain both pipelines at the same time, which I assume is actually commonplace.
What is the correct approach to share the deployment code as much as possible between the two tools?
I would expect that better integration between RM and MTM/LM will be a future feature. In the interim, you could investigate using Desired State Configuration to handle having a single script that configures environments for you.
DSC support isn't really out-of-the-box in RM Update 2, but RM Update 3 will have built-in support for DSC to both Azure and on-prem VMs. Update 3 CTP 1 is out right now, but it's not production-ready.
You can still use DSC from RM in Update 2, it just requires a bit more work.
I have spent a couple of days going through a lot of sites and reading about Nant , Rake etc.
please forgive my Noob question but I still cannot find what is the difference between Nant and CruiseControl.
As far as I can see Nant can do automated builds , run tests .
so what extra does cruisecontrol do ?
Also there was mention of Teamcity . there too from the documents I can see it can do builds but it also can use Nant but I fail to understand why it needs to use Nant when it can do the builds itself
I am basically trying to follow proper software practices by introducing automated builds at my workplace
Appreciate all help
Nant by itself can do builds and test, but it needs to be launched by some other mechanism such as a windows scheduled job. There is not a capability of launching the build only when source code changes, at least without
an amount of additional scripting.
Nant itself is just a script runner, not a scheduler - it requires some other software (or a manual user action) to launch it.
Continuous integration (CI) tools such as CruiseControl or TeamCity provide monitoring of source control to launch a build process in addition to other things. The build itself could the be a single nant script which runs the build and tests as you suggest, or the build could be done using a series of tasks which are built in to the CI server. The difference is not how the builds are done, but how they are initiated and reported.
CI servers additionally usually provide web-based reporting of the details of the build runs and unit tests.
In summary, Ci tools provide monitoring, scheduling, and reportingin addition to scripting of the build process.
Nant is a scripting language
CruiseControl is a free continuse integration tools
Teamcity in another contiuse integration tools,
regards,
I'm currently looking to start setting up nightly builds with TFS and our company has never done this before. I'm looking for some pointers on maybe where to get started, what I should look out for as well as structure of solutions.
Background
Current TFS source location has 2 web projects, 5-10 windows services, 10-15 supporting dlls. These will continue to grow.
Currently there are solution files for each web project and each windows service. Each of these solutions contain the supporting projects (internal dlls) and also the correlating unit testing projects.
All of our external dependencies (log4net, nhibernate etc) are managed by NuGet and are in a folder within TFS called packages
Some of my questions include but are not limited to
Should I have a master solution file that contains all of these projects? Maybe this is easier when setting up nightly builds?
I'd also like to run the unit and integration tests as part of the nightly builds. Is this just additional configuration on the build server?
What tools are involved when setting up nightly builds with TFS?
I'm not necessarily looking for complete answers but it would be great if someone could point me to some good resources (books, websites, blogs)? Like I said I'm really green as far as nightly builds are concerned and I just want to make sure I start off on the right foot. Hopefully I can learn from others mistakes.
Here are some simple "answers" to your 3 questions (though I agree with the comments above that this isn't the most answer-able SO question):
An good read on creating reliable builds in MSBuild : http://msdn.microsoft.com/en-us/magazine/dd483291.aspx
Yes running tests is just an option in a TFS Build Definition, you can configure a few options in addition to "on/off" : http://msdn.microsoft.com/en-us/library/ms253138.aspx
You can also use TFS Lab management and test agents to execute tests in a different manner: http://blogs.msdn.com/b/lab_management/archive/2009/05/18/vsts-2010-lab-management-basic-concepts.aspx
Configuring TFS builds : http://msdn.microsoft.com/en-us/library/dd647547.aspx
We are currently working on a Salesforce.com custom APEX project that involves a lot of apex classes, triggers and Visualforce pages. We also have numerous applications from AppExchange that are part of the system.
We develop all the Apex Classes, Visualforce pages, etc in test environment and then deploy it to the live environment using Eclipse IDE. What happens is that every time we deploy changes to the live environment, all the test methods of all the classes (including those from AppExchange Apps) seems to be executing. So deployment of a simple change could end up taking couple of minutes.
Is there a way in apex to "package" classes by namespace or something like that so that when we try to deploy a change, only the test methods relevant to that package are executed. If something like that exists, our deployment can happen much faster.
Unfortunately no, there is no partial testing for deployment of apex code, every change, no matter how minute or self-contained triggers a full test run. This among other things enforces code metrics (minimum total code coverage for instance)
IMHO, this is proving to be a two-sided coin when it comes to enforcing code reliability. When we started using apex all of our tests were very comprehensive performing actual testing of the code with lots of asserts and checks. Then we started having very very long deploy times so now our tests serve one and only function, satisfying minimum code coverage, and even with that simplification it takes almost 3 minutes to deploy anything and we only use 20% of our apex code allowance.
IMHO2, Apex is way too slow of a coding platform to be enforcing this kind of testing. I cant even imagine how long the tests would run if we reach 50% allowance, not to mention any more.
This is possible but you'll need to learn about Apache Ant and have a look at the Force.com Migration Toolkit. You can then use a Build file to determine which files are deployed as well as which tests are run.
I'm busy writing a whitepaper that'll touch on this and other related development strategies... I'll post to my blog when it's done.
If we use the apache ant migration tool we have many options for deployment
like
deployCodeFailingTest which will skip the test classes
and if you want to run only specific test classes
please use : something similar to this in ur build.xml
<target name="deployCode">
`<sf:deploy`
username="${sf.username}"
password="${sf.password}"
serverurl="${sf.serverurl}"
deployroot="codepkg">
<runTest>SampleDeployClass</runTest>
</sf:deploy>
</target>
for detailed reference please use this link
http://www.salesforce.com/us/developer/docs/daas/salesforce_migration_guide.pdf
I would recommend the following approach:
Git as repository for all your sf code
jenkins to deploy your code as CI/CD
PMD as the static code analyser
sfdx as the deployment method in jenkins for deployment.
Refer the trailhead link: https://trailhead.salesforce.com/users/strailhead/trailmixes/architect-dev-lifecycle-and-deployment
As it currently stands, this question is not a good fit for our Q&A format. We expect answers to be supported by facts, references, or expertise, but this question will likely solicit debate, arguments, polling, or extended discussion. If you feel that this question can be improved and possibly reopened, visit the help center for guidance.
Closed 10 years ago.
What tools do you use for Automated Builds / Automated Deployments? Why?
What tools do you recommend?
Hudson for automated builds. I chose it because it was the easiest to setup and demo. A system that's too complex and isn't slick-looking won't impress management enough to get them on-board for automated builds. Especially in a project that has a lot of inertia.
NAnt for builds (but MSBuild, Rake, almost anything would be fine) and CruiseControl.NET for deployments. I'm currently working with the new Cruise from ThoughtWorks studios as it provides a better way to stage the various pipelines and let's me deploy any version I want to a target environment.
We use TeamCity, from JetBrains. They also make Resharper And IntelliJ.
We use it for building our .Net applications, and it has been quite easy to set up, connect to TFS, and run additional tools from. It is very polished, and actually kinda reminds me of this site. Found it much nicer than CruiseControl, and for our team size it is free. If you need lots of different builds, more per-user builds, and so on then it costs a bit (but still quite reasonable).
Funnily enough I just spent two weeks overhauling (read implementing from scratch) our nightly build process. Great fun (no, really). I toyed with the idea of installing Team Foundation Server, but we use Perforce for source control and I didn't think it was worth the hassle.
Our process is now a set of Powershell scripts that run on a dedicated build/test server that do the following on a scheduled task:
Wipe out the entire source tree (check that you didn't have anything checked out first!)
Bring down the entire source tree from Perforce (from the last labelled build)
Generate a change report (by syncing to HEAD and watching what comes down)
Build the App
Index the PDB files to the Perforce sources
Store the binaries and symbols in a dedicated symbol server
Run the test projects
Build the installer
Label
Send out emails to the group with status reports on all of the above
Works well.
make and bash on linux
make and cmd on windows
Visual Build Pro
We use a combination of build tools and continuous integration server:
Build tools:
Maven
SBT
Gradle
Rake
Continuous Integration Servers:
Jenkins
Hudson
Travis CI
Automated Build Studio.
Instead of letting you mes with scripts or xml files, it comes with predefined graphical macro operations that allows you to create tasks easily.
For our Windows-compilable stuff, we use FinalBuilder.
CruiseControl for automated builds. Works great.
For automated builds, I think the best tool going right now is JetBrain's Team City. The free version has all the features you'll need for most 5-10 person teams. Set up is easy, configuring new projects is painless (relatively), and most importantly, it's reliable.
For automated migrations, nothing beats PowerShell.
UppercuT uses NAnt to build and it is the insanely easy to use Build Framework.
Automated Builds as easy as (1) solution name, (2) source control path, (3) company name for most projects!
http://code.google.com/p/uppercut/
Some good explanations here: UppercuT
More information
UppercuT is a conventional automated build, which means you set up a config file and then you get a bunch of features for free. Arguably the most powerful feature is the ability to specify environment settings in ONE place and have them applied everywhere, including documentation when it builds the source.
Documentation available: https://github.com/chucknorris/uppercut/wiki
Features :
Simple setup
Simple upgrades
Custom extension points (pre, post, and replace) for each step of the build process http://uppercut.pbworks.com/CustomizeUsingExtensionPoints
Has documentation for integration w/Team City, CruiseControl.NET, and Jenkins (formerly Hudson) https://github.com/chucknorris/uppercut/tree/master/docs
Works on Linux w/Mono
Versioning DLLs based on build number and source control revisions (SVN, TFS, Git, HG)
Compile activities - F5 or Ctrl + Shift + B
Strong naming made as easy as true/false
Code Testing and Analysis
Testing
NUnit
MbUnit v2
Gallio
xUnit
NCover
NDepend
Nitriq
Mono Migration Analyzer
Obfuscation
ILMerge
Environment Templating and Building (ConfigBuilder, DocBuilder, SQLBuilder, DeploymentBuilder) https://github.com/chucknorris/uppercut/blob/master/docs/ConfigBuilder.doc?raw=true
Packaging output to prepare for deployment
Zips up output
At work we use good ol' Ant to build our Java servlets.
We used to use Visual Build from Kinook software, but recently with our new application we switched to MSBuild since it had better integration with TFS and the ability to create custom tasks.
The GNU Autotools definitely. The autoconf and automake are de-facto standard for unix systems.
I've had success using buildbot, triggered by a post-commit script on a subversion repository. This has been used for both automated builds and automated testing.
ANT for both build and deployment/installs.
Makes a great cross-platform installer.
We use Hericus Zed Builds And Bugs Management for our automated builds.
We have 4 branches of code, each with java, c++, C#, cross platform compiles and installers for 5 OS's.
Make for the builds.
Debian packages for deployments (since our production servers runs it).
TeamCity running NAnt scripts for building/packaging and PowerShell for deployment.
I've found that using NAnt, powered by TeamCity, instead of the native TeamCity runners allows us to have a much richer build process (eg. css minimiser, etc). It also means the full build/package process can be run on any developers PC instead of just the TeamCity servers making it much easier to customise and debug problems in the build process.