Related
I have a ASP.NET project with my own Development, Staging and Production servers.
In all environements, I move code manually. So everytime I have to promote a change, I perform the following steps:
Get my latest code from SVN.
Merge the code between lower and to be promoted environment using tools like Beyond Compare.
Then I move the respective ASPX and DLL files and any Stored Procedures or table data manually to Production.
This is a very time consuming process and I would like to get some automated methods for code moves.
Is there a way I can get the code moved from SVN to my Servers using some automated tools or in a automated packages.
I am using ASP.NET 2.0 with IIS 7 and SQL Server 2008.
msbuild can help you with getting the code from svn and building it. You will need to create simple batch files to run it, alternatively you can use Cruisecontrol for that.
Manual merges should be avoided. If you are using VS2010, you can use xsd to transform your config files to production version
I am not a big fan of storedprocs. If you can encapsulate your stored procs with the code there is less room for errors and rolling back changes etc as well as making the deployment easier. Database schema updates should be done in a batch file and applied automatically.
There are multiple ways to deploy: webdeploy or msi file. It depends on how much work is required during the deployment process
I would look into continuous integration. My favorite because it is simplest to use is TeamCity.
You will still have to do some work with MSBuild.
You can set up the builds to be a button push from the site.
Have the code pushed when you check into svn.
Just about any way you can think of.
I would strongly urge you to use it to ALWAYS build you code and run unit tests on each SVN check it. It does not have to deploy but TeamCity will provide to you constant feedback on the state of your solution.
The process of creating a new build and releasing it to production is a critical step in the SDLC but it is often left as an afterthought and varies greatly from one company to the next.
I'm hoping people will share improvements they have made to this process in their organisation so we can all takes steps to 'reduce the pain'.
So the question is, specify one painful/time consuming part of your release process and what did you do to improve it?
My example: at a previous employer all developers made database changes on one common development database. Then when it came to release time, we used Redgate's SQL Compare to generate a huge script from the differences between the Dev and QA databases.
This works reasonably well but the problems with this approach are:-
ALL changes in the Dev database are included, some of which may still be 'works in progress'.
Sometimes developers made conflicting changes (that were not noticed until the release was in production)
It was a time consuming and manual process to create and validate the script (by validate I mean, try to weed out issues like problem 1 and 2).
When there were problems with the script (eg the order in which things were run such as creating a record which relies on a foreign key record which is in the script but not yet run) it took time to 'tweak' it so it ran smoothly.
It's not an ideal scenario for Continuous Integration.
So the solution was:-
Enforce a policy of all changes to the database must be scripted.
A naming convention was important for ensuring the correct running order of the scripts.
Create/Use a tool to run the scripts at release time.
Developers had their own copy of the database do develop against (so there was no more 'stepping on each others toes')
The next release after we started this process was much faster with fewer problems, indeed the only problems found were due to people 'breaking the rules', eg not creating a script.
Once the issues with releasing to QA were fixed, when it came time to release to production it was very smooth.
We applied a few other changes (like introducing CI) but this was the most significant, overall we reduced release time from around 3 hours down to a max of 10-15 minutes.
We've done a few things over the past year or so to improve our build process.
Fully automated and complete build. We've always had a nightly "build" but we found that there are different definitions for what constitutes a build. Some would consider it compiling, usually people include unit tests, and sometimes other things. We clarified internally that our automated build literally does everything required to go from source control to what we deliver to the customer. The more we automated various parts, the better the process is and less we have to do manually when it's time to release (and less worries about forgetting something). For example, our build version stamps everything with svn revision number, compiles the various application parts done in a few different languages, runs unit tests, copies the compile outputs to appropriate directories for creating our installer, creates the actual installer, copies the installer to our test network, runs the installer on the test machines, and verifies the new version was properly installed.
Delay between code complete and release. Over time we've gradually increased the amount of delay between when we finish coding for a particular release and when that release gets to customers. This provides more dedicated time for testers to test a product that isn't changing much and produces more stable production releases. Source control branch/merge is very important here so the dev team can work on the next version while testers are still working on the last release.
Branch owner. Once we've branched our code to create a release branch and then continued working on trunk for the following release, we assign a single rotating release branch owner that is responsible for verifying all fixes applied to the branch. Every single check-in, regardless of size, must be reviewed by two devs.
We were already using TeamCity (an excellent continuous integration tool) to do our builds, which included unit tests. There were three big improvements were mentioning:
1) Install kit and one-click UAT deployments
We packaged our app as an install kit using NSIS (not an MSI, which was so much more complicated and unnecessary for our needs). This install kit did everything necessary, like stop IIS, copy the files, put configuration files in the right places, restart IIS, etc. We then created a TeamCity build configuration which ran that install kit remotely on the test server using psexec.
This allowed our testers to do UAT deployments themselves, as long as they didn't contain database changes - but those were much rarer than code changes.
Production deployments were, of course, more involved and we couldn't automate them this much, but we still used the same install kit, which helped to ensure consistency between UAT and production. If anything was missing or not copied to the right place it was usually picked up in UAT.
2) Automating database deployments
Deploying database changes was a big problem as well. We were already scripting all DB changes, but there were still problems in knowing which scripts were already run and which still needed to be run and in what order. We looked at several tools for this, but ended up rolling our own.
DB scripts were organised in a directory structure by the release number. In addition to the scripts developers were required to add the filename of a script to a text file, one filename per line, which specified the correct order. We wrote a command-line tool which processed this file and executed the scripts against a given DB. It also recorded which scripts it had run (and when) in a special table in the DB and next time it did not run those again. This means that a developer could simply add a DB script, add its name to the text file and run the tool against the UAT DB without running around asking others what scripts they last ran. We used the same tool in production, but of course it was only run once per release.
The extra step that really made this work well is running the DB deployment as part of the build. Our unit tests ran against a real DB (a very small one, with minimal data). The build script would restore a backup of the DB from the previous release and then run all the scripts for the current release and take a new backup. (In practice it was a little more complicated, because we also had patch releases and the backup was only done for full releases, but the tool was smart enough to handle that.) This ensured that the DB scripts were tested together at every build and if developers made conflicting schema changes it would be picked up quickly.
The only manual steps were at release time: we incremented the release number on the build server and copied the "current DB" backup to make it the "last release" backup. Apart from that we no longer had to worry about the DB used by the build. The UAT database still occasionally had to be restored from backup (eg. since the system couldn't undo the changes for a deleted DB script), but that was fairly rare.
3) Branching for a release
It sounds basic and almost not worth mentioning, yet we weren't doing this to begin with. Merging back changes can certainly be a pain, but not as much of a pain as having a single codebase for today's release and next month's! We also got the person who made the most changes on the release branches to do the merge, which served to remind everyone to keep their release branch commits to an absolute minimum.
Automate your release process whereever possible.
As others have hinted, use different levels of build "depth". For instance a developer build could make all binaries for runnning your product on the dev machine, directly from the repository while an installer build could assemble everything for installation on a new machine.
This could include
binaries,
JAR/WAR archives,
default configuration files,
database scheme installation scripts,
database migration scripts,
OS configuration scripts,
man/hlp pages,
HTML documentation,
PDF documentation
and so on. The installer build can stuff all this into an installable package (InstallShield, ZIP, RPM or whatever) and even build the CD ISOs for physical distribution.
The output of the installer build is what is typically handed over to the test department. Whatever is not included in the installation package (patch on top of the installation...) is a bug. Challenge your devs to deliver a fault free installation procedure.
Automated single step build. The ant build script edits all the installer configuration files, program files that need changed ( versioning) and then builds. No intervention required.
There is still a script run to generate the installers when it's done, but we will eliminate that.
The CD artwork is versioned manually; that needs fixed too.
Agree with previous comments.
Here is what has evolved where I work. This current process has eliminated the 'gotchas' that you've described in your question.
We use ant to pull code from svn (by tag version) and pull in dependencies and build the project (and at times, also to deploy).
Same ant script (passing params) is used for each env (dev, integration, test, prod).
Project process
Capturing requirements as user 'stories' (helps avoid quibbling over an interpretation of a requirement, when phrased as a meaningful user interaction with the product)
following an Agile principles so that each iteration of the project (2 wks) results in demo of current functionality and a releasable, if limited, product
manage release stories throughout the project to understand what is in and out of scope (and prevent confusion abut last minute fixes)
(repeat of previous response) Code freeze, then only test (no added features)
Dev process
unit tests
code checkins
scheduled automated builds (cruise control, for example)
complete a build/deploy to an integration environment, and runs smoke test
tag the code and communicate to team (for testing and release planning)
Test process
functional testing (selenium, for example)
executing test plans and functional scenarios
One person manages the release process, and ensures everyone complies. Additionally all releases are reviewed a week before launch. Releases are only approved if there are:
Release Process
Approve release for a specific date/time
Review release/rollback plan
run ant with 'production deployment' parameter
execute DB tasks (if any) (also, these scripts can be version and tagged for production)
execute other system changes / configs
communicate changes
I don't know or practice SDLC, but for me, these tools have been indispensible in achieving smooth releases:
Maven for build, with Nexus local repository manager
Hudson for continuous integration, release builds, SCM tagging and build promotion
Sonar for quality metrics.
Tracking database changes to development db schema and managing updates to qa and release via DbMaintain and LiquiBase
On a project where I work we were using Doctrine's (PHP ORM) migrations to upgrade and downgrade the database. We had all manner of problems as the generated models no longer matched with the database schema causing the migrations to completely fail half way.
In the end we decided to write our own super basic version of the same thing - nothing fancy, just up's and down's that execute SQL. Anyway it worked out great (so far - touch wood). Although we were reinventing the wheel slightly by writing our own, the fact that the focus was on keeping it simple meant that we have far less problems. Now a release is a cinch.
I guess the moral of the story here is that it is sometimes OK to reinvent the wheel some times as long as you are doing so for a good reason.
I've always deployed my web applications via FTP (sometimes even xcopy), and then manually run database scripts myself.
I started deploying this way in the 90's, but lately, I've seen a few web apps with installers. I'm starting to question, if I'm locked into an out dated process. I'm a consultant, my apps are usually internal, so I don't worry about distributing and having others installing them.
But I'm curious; does anybody create installers to deploy internal asp.net web applications?
If so, why? (Voluntarily, mandated, or part of an automation process)
And have you had any problems doing it this way?
absolutely. We use it to do all of our apps. That way we create the installer and run it on the qa and uat environments to test and we know exactly what is going to happen in production. There are no guesses as to what order someone might do something in, or if they miss a step. It makes things a lot easier.
Ooh I forgot about the automated process too. We have systems in place (Ant Hill Pro) which automatically deploy it to the proper environments. The qa people don't have to wait for something to be done, because it's all done at 2 am. If they need to rerun the build with updates, the devs check the code in and we push a button, and it's automatically deployed. No waiting for the build engineer, because he's in a meeting or sick or whatever.
You always want to have an automated way to build and deploy - it greatly reduces the chances of a one-off error if you forget a certain step. Also, it allows you to offload the deploy to someone else easily without having to teach them 100 customized steps. Whether the project is internal or not, all applications should follow best practices.
Personally I'm a bit like the OP; generally I just deploy using FTP, but in saying that typically my applications are internal, or in the case of other projects, 100% managed by me.
I've also been thinking about this lately however, and have started to think about how using proper deployment may improve the process - having to document a detailed install process can be a real pain.
I use Powershell and found really easy to automate lots of tasks. You will probably find a bit different at the very begining but at the end you will see that it's all about the power of the .NET libraries !!!
I have use the "Web Setup Project" to create an MSI that installed the output of a "Web Deployment Project" for an internal app. Our server admin wasn't up to the task to doing a 50 step manual install. For my current app, my server admin doesn't like the 'black box' feel of MSI installers and prefers getting a pile of files and a 50 step deployment manual. (See a pattern here? Ask your server admin what he wants.)
The Web Setup Project doesn't make it immediately obvious how to install to anything other than the "Default Website", other than that, it made the installation process repeatable and created a built in way to rollback (by just running the installer from 1 version ago).
This of course assumes that your virtual directory doesn't hold any user modified content-- I wouldn't trust an MSI to properly merge user created and new files.
We use the "XCopy" deploy model here, since the Ops folks have their own method of setting up security on a new web application on the server.
However, we did need to use an installer when we had to install a web application that was using a newer version of Crystal Reports since it had to do something special with a key and we didn't have a full blown version of CR on the server itself. So keep that in mind when working with third party apps, they may need to do some kind of merge module that the MSI handles easily.
Yep...we have an app that needs a lot of pre-requisites set up....web service, windows service, user accounts, security, folder creation, GAC bits etc....I rolled it all up into a nice MSI with custom actions that can install and uninstall cleanly. Saved about one hours worth of work to deploy on a new box.
A lot of the other smaller apps are just deployed by doing Publish Website to a local folder then ftp'ing the contents to the target.
It greatly depends upon the scale of your project, your enviornment and your internal user base. I rarely deploy with an msi because we are too small an operation to have multiple environments (except for SharePoint, that's different all together) . We develop and use VS to deploy web apps to a development box, assuming they are approved then we use VS again to deploy to the live box.
The only proviso is that we have multiple copies of the web.config (appended with test, dev and live) and we then delete the suffix off the relevant file depending upon where its been deployed.
It's probably not the best methodology (I know it's not), but it works and it aids rapid deployment of small to medium sized solutions in a small-scale user environment.
F5ToDebug...
Your saying its OK to take short cuts if you dont have time to do it properly?
"who's going to test the code on the test environment?" You said it yourself that you have config files for _test - why would that not be a suitable test?
Seems like there are so many different ways of automating one's build/deployment that it becomes difficult to parse through all the different scenarios that people support in tutorials on the web. So I wanted to present the question to the stackoverflow crowd ... what would be the best way to set up an automated build and deployment system using the following configuration:
Visual Studio 2008
Web Application Project
CruiseControl.NET
One of the first things I tried was to have CCnet automatically zip the output and copy it to the server, but then that requires manual work to unzip at the destination. However, if we try to copy all the files individually, then it could potentially take a long time if it's a large application (build server lives outside of the datacenter in our office ... I know).
Also of particular interest is how we would support multiple environments as we have dev, qa, uat, and then of course prod.
MSDeploy seems really interesting, but unless I'm interpreting the literature incorrectly, doesn't help in the scenario of deploying from the output of a build server. If anything, it seems like it'll be useful in deploying one build across a build farm ... but even for deploying from one environment to another, one would have to manually change config settings and web service URLs, etc.
I recently spent a few days working on automating deployments at my company.
We use a combination of CruiseControl, NAnt, MSBuild to generate a release version of the app. Then a separate script uses MSDeploy and XCopy to backup the live site and transfer the new files over.
Our solution is briefly described in an answer to this question Automate Deployment for Web Applications?
You might be interested in MSDeploy. Here's a Scott Hanselman post on this. It's only available as a technical preview at the moment (September 2008) but is worth evaluating against your requirements.
There is another new build tool (a very intelligent wrapper) called NUBuild. Its lightweight, open source and extremely easy to setup and provides almost no-touch maintenance. I really like this new tool and we have made it standard tool for our continuous build and integration process of our projects (we have about 400 projects across 75 developers). Try it out.
http://nubuild.codeplex.com/
Easy to use command line interface
Ability to target all .Net framework
version i.e. 1.1, 2.0, 3.0 and 3.5
Supports XML based configuration
Supports both project and file
references
Automatically generates the “complete
ordered build list” for a given
project – No touch maintenance.
Ability to detect and display
circular dependencies
Perform parallel build -
automatically decides which of the
projects in the generated build list
can be built independently.
Ability to handle proxy assemblies
Provides visual clue to the build
process e.g. showing “% completed”,
“current status” etc.
Generates detailed execution log both
in XML and text format
Easily integrated with
Cruise-Control.Net continuous
integration system
Can use custom logger like XMLLogger
when targeting 2.0 + version
Ability to parse error logs
Ability to deploy built assemblies to
user specified location
Ability to synchronize source code
with source-control system
Version management capability
Do you have the ability to run commands remotely? The PsExec utility from Systinternals would let run a command line unzip program on the remote machine. If you have a script that copies the build as a .zip file to the remote site, you would just need one more line for the PsExec call to unzip the files.
I had a related question about getting a deployable set of files from an automated build. I found Web Deployment Projects (links and all in the old question) did what I needed - they're a VS and MSBuild add-on.
This is a common problem (and I wish I had read it sooner) for all development, not just ASP.NET. Being one of its developers, my team naturally uses BuildMaster internally for the entire release process, and for most scenarios it's free. Within the tool, we are able to perform all the standard CI builds to create artifacts and then set up an automation process to deploy these artifacts to any one of the 40+ servers we have internally or externally hosted, depending on the specific application or environment.
Since you specifically mentioned deployment to different testing environments, this is a fundamental aspect of the tool. The idea is to model the environment workflow (e.g. Integration -> QA -> Production) you already have in place and essentially promote a build all the way from source control to production. Most times, it's as simple as adding a deployment action that deploys an artifact to the environment, other times it can be much more complex.
You also casually mentioned configuration file changes are part of deployment, which is another built-in component to BuildMaster. The idea we had was to use the tool itself as the central hub for all configuration files and deployments, thus ensuring the latest changes are applied automatically with a simple "deploy configuration files" action in your deployment plan.
One thing you didn't mention with regard to this process is the database deployment aspect. Most ASP.NET applications require an associated database, otherwise they could just be static HTML files. It is crucial that the database schema gets updated to the appropriate database version with every deployment. There is, not surprisingly, a module within BuildMaster that handles this for you as well. The idea is to store DDL-DML scripts within the tool itself, and by executing scripts only once per environment, it ensures that all of your databases across each environment are up-to-date as your builds are deployed through them. Other scripts (e.g. stored procedures, views, triggers, etc.) are essentially code files and therefore belong in source control. These DROP-CREATE-CONFIGURE type scripts can be run each and every time in most cases with a simple deployment action.
Another piece of the deployment puzzle that most developers do not think about is process automation. Many developers need to perform sign-offs or fill out change request forms in order to manually perform these processes. Again, this is all available as part of the automated workflow setup within BuildMaster. You can setup blockers that do not allow promotion to say the QA environment unless all unit tests have passed, or block promotion to the Staging environment unless someone from the QA team approves the build and all issues in your issue tracking tool are resolved/closed for that particular release.
While I realize I left out CC.NET from the answer, our applications are all built and deployed through BuildMaster so we no longer need it, though we could however just as easily pickup the artifacts from a drop location and deploy them in later environments.
I see that many people use CC for their .NET projects, but why not use Jenkins, Sonarqube? They got all you need. I setup all this in 3 days. I have a Win 2008 server R2, MSSQL, Jenkins, VIsual SVN and Sonarqube.
It all works great and u get all metrics on your project. Sonarqube uses Gallio, Gendarme, FXcop, Stylecop, NDepths and PartCover to get your metrics and all this is pretty straight forward since SonarQube do this automatically without much configuration.
i post som pictures for u too get a feeling for it. Here is Jenkins witch builds and get Sonar metrics and a another job for deploying automatically to IIS
And Sonarqube, all metrics for my project. This is a simple MVC4 app, but it works great!:
If you want more information i can be more specific but i think you should at least consider jenkins. If CC suites you better, at least you looked at good alternative before you chose.
This whole setup uses MSBuild, too build and deploy the apps.
Or, actually establishing a build process when there isn't much of one in place to begin with.
Currently, that's pretty much the situation my group faces. We do web-app development primarily (but no desktop development at this time). Software deployments are ugly and unwieldy even with our modest apps, and we've had far too many issues crop up in the two years I have been a part of this team (and company). It's past time to do something about that, and the upshot is that we'll be able to kill two Joel Test birds with one stone (daily builds and one-step builds, neither of which exists in any form whatsoever).
What I'm after here is some general insight on the kinds of things I need to be doing or thinking about, from people who have been in software development for longer than I have and also have bigger brains. I'm confident that will be most of the people currently posting in the beta.
Relevant Tools:
Visual Build
Source Safe 6.0 (I know, but I can't do anything about whether or not we use Source Safe at this time. That might be the next battle I fight.)
Tentatively, I've got a Visual Build project that does this:
Get source and place in local directory, including necessary DLLs needed for project.
Get config files and rename as needed (we're storing them in a special sub directory that isn't part of the actual application, and they are named according to use).
Build using Visual Studio
Precompile using command line, copying into what will be a "build" directory
Copy to destination.
Get any necessary additional resources - mostly things like documents, images, and reports that are associated with the project (and put into directory from step 5). There's a lot of this stuff, and I didn't want to include it previously. However, I'm going to only copy changed items, so maybe it's irrelevant. I wasn't sure whether I really wanted to include this stuff in earlier steps.
I still need to coax some logging out of Visual Build for all of this, but I'm not at a point where I need to do that yet.
Does anyone have any advice or suggestions to make? We're not currently using a Deployment Project, I'll note. It would remove some of the steps necessary in this build I presume (like web.config swapping).
When taking on a project that has never had an automated build process, it is easier to take it in steps. Do not try to swallow to much at one time, otherwise it can feel overwhelming.
First get your code compiling with one step using an automated build program (i.e. nant/msbuild). I am not going to debate which one is better. Find one that feels comfortable to you and use it. Have the build scripts live with the project in source control.
Figure out how you want your automated build to be triggered. Whether it is hooking it up to CruiseControl or running a nightly build task using Scheduled Tasks. CruiseControl or TeamCity is probably the best choice for this, because they include a lot of tools you can use to make this step easier. CruiseControl is free and TeamCity is free to a point, where you might have to pay for it depending on how big the project is.
Ok, by this point you will be pretty comfortable with the tools. Now you are ready to add more tasks based on what you want to do for testing, deployment, and etc...
Hope this helps.
I have a set of Powershell scripts that do all of this for me.
Script 1: Build - this one is simple, it is mostly handled by a call to msbuild, and also it creates my database scripts.
Script 2: Package - This one takes various arguments to package a release for various environments, such as test, and subsets of the production environment, which consists of many machines.
Script 3: Deploy - This is run on each individual machine from within the folder created by the Package script (the Deploy script is copied in as a part of packaging)
From the deploy script, I do sanity checks on things like the machine name so things don't accidentally get deployed to the wrong place.
For web.config files, I use the
<appSettings file="Local.config">
feature to have overrides that are already on the production machines, and they are read-only so they don't accidentally get written over. The Local.config files are not checked in, and I don't have to do any file switching at build time.
[Edit] The equivalent of appSettings file= for a config section is configSource="Local.config"
We switched from using a perl script to MSBuild two years ago and haven't looked back.
Building visual studio solutions can be done by just specifying them in the main xml file.
For anything more complicated (getting your source code, executing unit tests, building install packages, deploying web sites) you can just create a new class in .net deriving from Task that overrides the Execute function, and then reference this from your build xml file.
There is a pretty good introduction here:
introduction
I've only worked on a couple of .Net projects (I've done mostly Java) but one thing I would recommend is using a tool like NAnt. I have a real problem with coupling my build to the IDE, it ends up making it a real pain to set up build servers down the road since you have to go do a full VS install on any box that you want to build from in the future.
That being said, any automated build is better than no automated build.
Our build process is a bunch of homegrown Perl scripts that have evolved over a decade or so, nothing fancy but it gets the job done. One script gets the latest source code, another builds it, a third stages it to a network location. We do desktop application development so our staging process also builds install packages for testing and eventually shipping to customers.
I suggest you break it down to individual steps because there will be times when you want to rebuild but not get latest, or maybe just need to re-stage. Our scripts can also handle building from different branches so consider that also with whatever solution you develop.
Finally we have a dedicated build machine that rebuilds the trunk and maintenance branches every night and sends out an email with any problems or if it completed successfully.
One thing I would suggest ensure your build script (and installer project, if relevant in your case) is in source control. I tend to have a very simple script that just checks out\gets latest the "main" build script then launches it.
I say this b/c I see teams just running the latest version of the build script on the server but either never putting it in source control or when they do they only check it in on a random basis. If you make the build process to "get" from source control it will force you to keep the latest and greatest build script in there.
Our build system is a makefile (or two). It has been rather fun getting it working as it needs to run on both windows (as a build task under VS) and under Linux (as a normal "make bla" task). The really fun thing is that the build gets the actual file list from a .csproj file, builds (another) makefile from that, and run that. In the processes the make file actually calls it's self.
If that thought doesn't scare the reader, then (either they are crazy or) they can probably get make + "your favorite string mangler" to work for them.
We use UppercuT.
UppercuT uses NAnt to build and it is extremely easy to use.
http://code.google.com/p/uppercut/
Some good explanations here: UppercuT