devops

In my previous article on the server lifecycle I mentioned ConfigurationDrift, a term that I've either coined, or I've forgotten where I originally heard, although most likely I got it from the Puppet Labs folks.

Configuration Drift is the phenomenon where running servers in an infrastructure become more and more different as time goes on, due to manual ad-hoc changes and updates, and general entropy.

A nice automated server provisioning process as I've advocated helps ensure machines are consistent when they are created, but during a given machine's lifetime it will drift from the baseline, and from the other machines.

There are two main methods to combat configuration drift. One is to use automated configuration tools such as Puppet or Chef, and run them frequently and repeatedly to keep machines in line. The other is to rebuild machine instances frequently, so that they don't have much time to drift from the baseline.

The challenge with automated configuration tools is that they only manage a subset of a machine's state. Writing and maintaining manifests/recipes/scripts is time consuming, so most teams tend to focus their efforts on automating the most important areas of the system, leaving fairly large gaps.

There are diminishing returns for trying to close these gaps, where you end up spending inordinate amounts of effort to nail down parts of the system that don't change very often, and don't matter very much day to day.

On the other hand, if you rebuild machines frequently enough, you don't need to worry about running configuration updates after provisioning happens. However, this may increase the burden of fairly trivial changes, such as tweaking a web server configuration.

In practice, most infrastructures are probably best off using a combination of these methods. Use automated configuration, continuously updated, for the areas of machine configuration where it gives the most benefit, and also ensure that machines are rebuilt frequently.

The frequency of rebuilds will vary depending on the nature of the services provided and the infrastructure implementation, and may even vary for different types of machines. For example, machines that provide network services such as DNS may be rebuilt weekly, while those which handle batch processing tasks may be rebuilt on demand.

One of the cornerstones of a well-automated infrastructure is a system for provisioning individual servers. A system that lets us reliably, quickly, and repeatably create new server instances that are consistent across our infrastructure means we spend less time fiddling with individual servers. Instead, servers become disposable components that are easily swapped, replaced, and expanded as we focus our attention on the bigger picture of the services we're providing.

The first step in achieving this is making sure server instances are built using an automated process. This ensures every server is built the same way, that improvements can be easily folded into the server build process, and that it is a simple matter to spin up new instances and to scrap and replace broken ones. Automating this process also means your team of highly skilled, well-paid professionals don't need to spend large amounts of their time on the brainless rote-work of menu-clicking through OS installation work.

I first used automated installation by PXE-booting physical rack servers in 2002, following the advice I found on the then-current infrastructures.org site, and in later years applied the same concepts with virtualized servers and then IaaS cloud instances.

The machine lifecycle

I think of this as the machine lifecycle (which I tend to call the 'server lifecycle' because that's what I normally work with, although it's just as applicable to desktops). This involves a number of activities required to set up and manage a single machine instance, such as partitioning storage, installing software, and applying configuration.

These activities are applied during one or more phases of the machine lifecyle. There are three phases: "Package Image", "Provision Instance", and "Update Instance". There are a number of different strategies for deciding which activities to do in each phase.

The various activities may be applied during one or more phases, depending on the strategy used to manage the machine's lifecycle. Some strategies carry out more activities during the packaging phase, for instance, while other approaches might have a simpler packaging phase but do more in the provisioning and/or updating phase.

Machine lifecycle phases

Image packaging phase

In the image packaging phase, some or all elements of a machine instance are pre-packaged into a machine image in a way that can be reused to create multiple running instances.

This could be as simple as using a bare OS installation CD or ISO from the vendor. Alternately, it could be a snapshot of a fully installed, fully configured runnable system, such as a Symantec Ghost image, VMWare template, or EC2 AMI. Either way, these images are maintained in a Machine Image Library for use in the instance provisioning phase.

Different machine lifecycle strategies use different approaches to image packaging. ManualInstanceBuilding and ScriptedInstanceBuilding both tend to use stock OS images, which involves less up-front work and maintenance of the Machine Image Library, since the instances are take straight from the vendor. However, work is still needed to create, test, and maintain the checklists or scripts used to configure instances when provisioning.

On the other hand, CloningExistingMachineInstances and TemplatedMachineInstances both create pre-configured server images, which need only minor changes (e.g. hostnames and IP addresses) to provision new instances. This is appealing because less work is done to provision a new instance, but the drawback is that creating and updating images takes more work. Admins tend to make updates and fixes to running instances which may not make it into the templates, which contributes to ConfigurationDrift, especially if changes are made ad-hoc.

CloningExistingMachineInstances, which usually takes the shape of copying an existing server to create new ones as needed, tends to make ConfigurationDrift worse, as new servers inherit the runtime cruft and debris (log files, temporary files, etc.) of their parents, and it is difficult to bring various servers into line with a single, consistent configuration. TemplatedMachineInstances are a better way to keep an infrastructure consistent and easily managed.

The tradeoffs between scripted installs vs. packaged images depends partly on the tools used for scripting and / or packaging, which in turn often depends on the hosting platform. Amazon AWS requires the use of templates (AMIs), for example. In either case, exploiting automation more fully in the provisioning phase favours the case for keeping the packaging phase as lightweight as possible.

Instance Provisioning Phase

In the provisioning phase, a machine instance is created from an image and prepared for operational use.

Examples of activities in this phase include instantiating a VM or cloud instance, preparing storage (partitioning disks, etc.), installing the OS, installing relevant software packages and system updates, and configuring the system and applications for use.

There are two main strategies for deciding which activities belong in the packaging versus the provisioning phases. One is RoleSpecificTemplates, and the other is GenericTemplate.

With RoleSpecificTemplates, the machine image library includes images that have been pre-packaged for specific roles, such as web server, application server, mail server, etc. These have the necessary software and configuration created in the packaging phase, so that provisioning is a simple matter of booting a new instance and tweaking a few configuration options. There are two drawbacks of this approach. Firstly, you will have more images to maintain, which creates more work. When the OS used for multiple roles is updated, for example, the images for all of those roles must be repackaged. Secondly, this pattern gives you less flexibility, since you can't easily provision an instance that combines multiple roles, unless you create - and then maintain - images for every combination of roles that you might need.

With the GenericTemplate pattern, each image is kept generic, including only the software and configuration that is common to all roles. The role for each machine instance is assigned during the provisioning phase, and software and configuration are applied accordingly then. The goal is to minimise the number of images in the machine image library, to reduce the work needed to maintain them. Typically, a separate template is needed for each hardware and OS combination that can't be supported from a single OS install. The JeOS (Just Enough Operation System) concept takes this to the extreme, making the base template as small as possible.

The GenericTemplate pattern does require a more robust automated configuration during provisioning, and may mean provisioning an instance takes longer than using more fully-built images, since more packages will need to be installed during install.

Instance Updating Phase

Once a machine instance is running and in use, it is continuously updated. This includes activities such as applying system and software updates, new configuration settings, user accounts, etc.

Many teams carry out these updates manually, however it requires a high level of discipline and organization to maintain systems this way, especially as the number of systems grows. The number of machines that a team can be managed is closely dependent on the size of the team, so the ration of servers to sysadmins is low. In practice, teams using manual updates tend to be reactive, updating machines opportunistically when carrying out other tasks, or in order to fix problems that crop up. This leads to ConfigurationDrift, with machines becoming increasingly inconsistent with one another, creating various problems including unreliable operation (software that works on one machine but not another), and extra work to troubleshoot and maintain.

Automated management of infrastructure is vital for delivering highly effective IT services. But although there are plenty of tools available to help implement automation, it's still common to see operations teams manually installing and managing their servers, which leads to a high-maintenance infrastructure, which soaks up the team's time on firefighting and other reactive tasks.

I've met many smart and skilled systems administrators in this situation. These folks know automation can make their life easier, but they can't afford to take time away from turning cranks, greasing wheels, and unjamming the gears to keep their infrastructure puffing along in order to focus on improving their situation.

I'm convinced this is largely due to habit. Even though these teams understand that automation would be useful to them, when the pressure is on (and the pressure is always on), they roll up their sleeves, ssh into the servers and knock them into shape, because that's the fastest way to get stuff done. Manual infrastructure management is what they're used to. I find that most of these teams haven't had personal experience of well-automated infrastructures, and don't tend to believe it's something they can realistically implement for their own operations.

Sysadmins who have worked in teams with mature, comprehensive automation, on the other hand, can't go back. Sure, they might log into a box to diagnose and fix something that needs fixing right now, but they can't relax until they've baked the fix into their automated configuration, and made sure that their monitoring will alert them ahead of time if the problem happens again.

Breaking out of manual infrastructure management and setting up an effective automation regime is difficult. Although there are loads of tools out there to make it work, it helps to understand good strategies for implementing them. I recommend looking over the material on the infrastructures.org site. It hasn't been updated in a few years, so doesn't take much of the advances since then into account, including virtualization, cloud, and newer tools like Chef and Puppet, but there is still rich material there.

Another must-read which more up to date is Web Operations by John Allspaw, Jesse Robbins, and a bunch of other smart peeps.

I'm also planning to share a few of the practices I've seen and used for automation in upcoming posts.

In my previous post, I glibly said that SLAs represent waste that an organization has identified and formalized. ReaderKenfin commented on my post, rightly calling me out to provide alternatives.

If you believe that SLA's 'formalise waste' this way how would you approach my situation where communications are beyond poor (atrocious) and the org structure is silo'd and no one is accountable for their work?

Kenfin's example illustrates my point quite well - the organization's structure is an obstacle to effective delivery. Since he's not in a position to fix this problem, he's turned to SLA's as a way to manage the problem. They won't make the issues go away, but they may give him a handle to manage them, and importantly, make them more predictable.

But it's a fair question, what can someone in Kenfin's shoes do in the face of an IT organization which is inherently not aligned to effectively providing the services he needs to deliver software to his users effectively?

A common strategy, and one that I've helped teams inside these kinds of organizations do, is to completely bypass the existing IT organization. The goal is to put control of everything that the product team needs in order to deliver into its hands, rather than leaving it at the mercy of a group (or multiple groups) who have other priorities.

Outsource it!

One way to do this is outsourcing, finding another company that specializes in the functions that the IT group would provide, whether this means development, integration, hosting, or something else. This works best if the project is not seen as core to the business, so that it avoids fear of entrusting sensitive data or business critical functions to outsiders. It also helps if the project needs skills that can't be found in-house.

My friends at Cognifide have built their business on this, building technically complex content-focused websites for corporate clients, delivering far more quickly, and with greater expertise, than most corporate IT organizations can manage. This is also the premise that Software as a Service (SaaS) is based on. By choosing SalesForce for CRM a company completely bypasses the massive IT project that would be required to implement an off the shelf, self-hosted CRM package (integration with other applications aside).

There are pitfalls to outsourcing to bypass IT. Many outsourcers are no more responsive than an in-house IT department, using SLAs and change control processes to make their workload, risks, and profitability more manageable.

Do it yourself!

The strategy I've most often been involved with myself (although I didn't really think of it this way at the time) is product departments building their own IT capabilities. Again, this is about having control of the services and resources the group needs in order to deliver to its own customers.

The typical pattern is an "online" (or often, "digital") department of a company where online was originally on the fringes of the main business, but has in recent years grown into a major channel for sales, customer service, or even delivery of products (for example in publishing).

The online team leverages their growing importance, as well as the specialized needs they have compared with typical corporate IT custometrs, to get approval from top management to create their own "digital operations team" or similar. This team may outsource elements of infrastructure, such as hosting (with IaaS cloud providers as an increasingly appealing option), but they are able to respond immediately to the needs of the online product group, because a) they don't have to juggle requests from other departments and teams, and b) they report directly to the manager of that group.

But what if I have to use the crappy IT guys who don't care about my project?

Those strategies are not feasible for every team. I've certainly had to support projects where we had no alternative but to struggle along with unresponsive IT. In these cases, SLAs may well have to do, even though they represent waste and inefficiency.

There are a few other things you might at least try out in these cases. Your goal is still to have the resources you need in order to get things done at your disposal as far as possible. So see if you can identify those services which are especially critical, and particularly those which are likely to change frequently, and see if you can get some dedicated resource assigned to your project. You want someone who will sit with your team, be incentivized by the success of your project, and who has the skills, authority, and system privileges to carry out the tasks you need.

If full time secondment of people to your team is not quite feasible due to budget, lack of available resource, etc., see if you can at least get commitments of time from the right people. Can someone come to daily standups? Weekly meetings? Regular release management meetings? Ask for as much as possible to start with, then see what you can get.

Also, maybe you can hire someone into your own team with qualifications and background that will help them effecitely liaise with difficult IT teams. Your own DBA, security consultant, etc. can engage with the IT groups using their own language, couching things in terms that address their concerns. They may be able take certain tasks off the IT group's plates, which ends up giving you the ability to get things done more quickly, while at the same time making IT grateful that their workload is lighter.

But the right thing to do is ...

These are all ways to work around the core problems. The best solution is of course for the organization to restructure itself in a way that aligns its resources with its goals. Most companies, especially large ones, insist on organizing themselves in ways that are self-defeating. It's a shame that many people who work in large companies accept this as normal, often even as desirable.

Grouping everyone with a given function into a single group forces them to focus on juggling the competing needs of many stakeholders, managing their own risks (especially the risk of getting blamed when projects fail). They will inevitably favor the abstract principles of their own technical practices over what is most effective in making the business succeed. Much better to group people into units that have complete ownership for delivering business value, and find ways to connect staff of given function with each other so they can develop their skills as working practices.

Unfortunately most of us are rarely in a position to influence this, so I hope that my suggestions will be helpful to some people in making things a little less painful.

A development team's infrastructure - development and QA environments, CI servers, SCM servers, etc. - are indisputably business critical, but rarely given the kind of monitoring attention that production environments are. This is a missed opportunity, not only to ensure the continuity of development work, but also to gain valuable insight.

Reasons to monitor your application in every environment it's deployed to:

#1 Keep your development team moving

This is the obvious one. You need to know before you run out of disk space, RAM, etc.

#2 Optimize your CI / deployment pipeline

Do you know what the limiting factors are on the time it takes your automated tests to run? The shorter you make your dev/test/fix feedback loop, the more productive your team will be, so why not analyze and optimize it as you would any other key software system? If checkout from SCM takes 20% of the time to test results, what can you do to reduce it? Are your unit tests constrained by CPU, RAM, or disk I/O?

#3 Understand your applications

We're conditioned to think that measuring performance and resource consumption is only useful in an environment that mirrors our production hardware. But if we build up an awareness of how our application uses memory and other resources as a part of every execution and every environment, we'll have a deep and intuitive understanding of what makes it tick.

#4 Develop and test your monitoring

By having monitoring running against applications while they are still in development, you will find ways to improve how you monitor ("let's measure the activity on our queues"), catch holes in your monitoring ("why didn't our monitoring tell us when the dev server queue broker went down?"), and test changes to your monitoring.

Once you put monitoring in place in development and testing and make a habit of using it, it becomes a ubiquitous and indispensable part of your team's working processes, similar to the shift to using CI well.

Don't leave it as a low priority task, something to get around to at some point after you get around to setting up a perfect performance testing environment. Put it at the center of your team's toolset for understanding your work.

A common pattern in software development teams is to have a person who owns the build system. This may be a deliberate decision, or it may evolve organically as a particular team member gravitates towards dealing with the build scripts, automated testing and deployment, etc. While it's normal for some team members to have a deeper understanding of these things than others, it's not a good idea for the knowledge and responsibility for the build to become overly concentrated in one person.

The build system should be looked at as a module or component of the software application or platform being developed, so the philosophy taken towards code ownership apply.

If a single person owns the build system, everyone else becomes dependent on them to fix issues with it, and to extend it to meet new needs. There is also a risk, especially for projects which are big enough that maintaining the build system becomes a full time job, that a bit of a siloed mentality can develop.

If developers have a poor understanding of how their software is built and deployed, their software is likely to be difficult and costly to deploy. On the flip side, if build and test tools are implemented and maintained entirely by people who don't develop or test the software, it isn't likely to make the life of those who do as easy as it could be.

In the past few months I've taken on a role which is largely focused on this area, and have been helping a development team get their build and delivery system in place. Pairing with developers to implement aspects of the system has worked well, as has letting them take on the setup of particular areas of the build and test tooling. This follows what Martin Fowler calls "Weak Code Ownership", allowing everyone to take part in working on the build and test system.

Special attention is needed for stages of the path to production as they get further from the developer's workstation. Developers are keen to optimize their local build and deployment, but can often be fuzzy on what happens when things are deployed in server environments. This is exacerbated when the platforms are different (e.g. developers working on Windows, code deployed on Linux).

Even without platform differences, developers understandably focus on the needs of their own local build over those of production system deployment. This is natural when server deployment is not a part of their daily world. So the best way to compensate for this is to keep developers involved in implementing and maintaining server deployment.

Driving the implementation of the build and deployment system according to the needs of business stories has also been useful. So rather than setting up tooling to test parts of the system that haven't been developed yet, wait until the design of the code to be tested starts to be understood, and the code itself has actually started being developed. This helps ensure the tooling closely fits the testing and deployment needs, and avoids waste and re-work.

In the first post in this series, I explained why I think Maven Maven is a good idea. Most projects need pretty much the same thing from a build system, but using Ant normally results in complex, non-standard build system which becomes a headache to maintain.

In theory, Maven should be a better way to run a build. By offering a standardised build out of the box, you would massively reduce the setup and learning curve for new joiners to the development team and take advantage of a healthy ecosystem of plugins that can be simply dropped into your build, and save loads of setup and maintenance hassle.

Although its goes pretty far towards the first two of these advantages, in my second post I described how Maven's configuration is too complex even for simple things.

I note that the Maven site doesn't currently mention "convention over configuration", although I'm sure it used to in the past, and there are plenty of references to it around the web. The Wikipedia entry for convention over configuration lists Maven as a poster-child, and Sonatype, the commercial company supporting Maven, names a chapter of their reference book named after the concept.

But it's a load of bollocks.

Anyway.

My final point (for this series, anyway) is on flexibility. The tradeoff between configuration complexity is normally flexibility. This is certainly the case with Ant; the complexity which makes every Ant-based build system a painfully unique snowflake buys you the capability to do damn near anything you want with it.

But Maven's complexity does not buy us flexibility.

My team wants to divide up its testing into multiple phases, following the "Agile testing pyramid" concept as mentioned by Mike Cohn.

So we'd like to have four layers to our pyramid, unit tests running first; database integration tests running next; web service tests third; and only if all of these pass do we run web UI tests. These test groups run in order of increasing heaviness, so we get feedback on the simple stuff quickly.

Maven supports two levels of testing, unit tests and integration tests. The failsafe plugin which provides integration testing support seems fairly new, and is actually pretty good if you only need one phase of integration testing. It lets you configure setup and teardown activities, so you can fire up an app server before running tests, and make sure it gets shut down afterwards.

If we could get failsafe to run three times during the build, each time carrying out different setup and teardown activities, and running different groups of tests, my team would be fairly happy with Maven.

It is possible to use build profiles to set up different integration tests in this way, but to get them to run, you need to run the build three times, and each time the preceding steps will be re-run - compiling, unit tests, packaging, etc. So it's kind of nasty, brutish, and too long.

The right way to achieve what we're after is probably to customise the build lifecycle, or create a new build lifecyle. Either way, it involves creating custom plugins, or extensions, or both. I've taken a stab at working out how, but after burning a couple of evenings without getting anywhere, I've shelved it.

I have no doubt it can be done, but it's just easier to do it in Ant and move on to other tasks. And that's pretty much the bottom line for me. I still like the idea of Maven, I have hopes it will continue to improve (it's a thousand times more usable than it was a few years ago), and maybe even go through a shift to embrace convention over configuration for real.

In the meantime, I'm likely to reach for Maven when I need something quickly for a project that seems likely to fit its paradigm, but for heavier projects (most of the ones I'm involved with for client projects), Ant is the pragmatic choice.

In my previous post I explained why Maven is a good concept for Java project builds. In this post I'll delve into a key area where it falls down, the overcomplexity of its configuration.

In brief, we have a proliferation of home-brewed build systems created in many different ways using Ant, all of which do much the same thing. Since the vast majority of Java projects have very similar build requirements, an off the shelf build system should be able to do the job.

More to the point, a standard build system using convention over configuration should be simple to set up and maintain. This in turn helps the development team's velocity, since less time is spent fiddling with (or worse, fighting against) the build system, so more time can be spent delivering value for customers.

Maven fulfils this if your build needs are ridiculously simple. Slap your source code into folders for application code and test code, name your tests right, and put in a very small pom.xml file. Run "mvn install" and your code is compiled, unit tested, and ready to deploy.

Things get ugly pretty quickly though, as soon as you need to make even a small tweak to the basic build.

For example, let's look at what's involved in changing the version of Java you want your project built to, for language compatibility and/or for compatibility with the runtime environment your project will be used in. Maven 3.0.x uses Java 1.5 by default, even if you have Java 1.6 installed, which is sensible enough.

So we have two parameters that need changing, language version for the source, and target version for generated classes. Here's what we have to add to our simple pom.xml to achieve this:

<project>
  [...]
  <build>
    [...]
    <plugins>
      <plugin>
        <groupId>org.apache.maven.plugins</groupId>
        <artifactId>maven-compiler-plugin</artifactId>
        <version>2.3.2</version>
        <configuration>
          <source>1.4</source>
          <target>1.4</target>
        </configuration>
      </plugin>
    </plugins>
    [...]
  </build>
  [...]
</project>

So the first thing that strikes you is, holy spaghetti, that's a lot of scaffolding for two parameters. Someone wading through a pom file trying to understand our build in order to modify or fix it will have a hard time working out the meat of this snippet, and it's not clear that it's all only there for those two parameters.

Much worse, the extra scaffolding isn't harmless. Before we had this, compilation was done by the maven-compiler-plugin, using the latest available version for the version of Maven we're using. Now this out of the box default has been surfaced in the pom file. Our project is locked into knowing about, and managing the plugin and its version. If we don't specify the version, Maven 3.x prints a scary warning suggesting that in the future it will no longer enable our careless disregard for managing the nuts and bolts of its innards, so expect our build to break horribly.

As soon as you set any configuration option on a built-in component of Maven, you are forced to take responsibility for that component that you wouldn't have otherwise, and clutter your configuration with stuff that would otherwise be assumed by Maven.

Now suppose you want to run integration tests on your code, that run heavier weight tests in a separate phase after unit testing. The testing model we're following is to have lightweight unit tests that run fast for quick feedback on basic errors, then heavier tests which probably need a container like Jetty, and perhaps a database, to test a bit more deeply. It's often referred to as integration testing, meaning you're testing the components of a single application integrated together.

(However, the naming leads to confusion since integration testing can also refer to testing integration with external applications, web services, etc.)

So Maven supports running a separate set of tests in an integration test phase that runs after the unit tests, using the failsafe plugin. How do we make use of this? Firstly, we write JUnit test case classes, and name them ITMyClass, and the "IT" beginning tells Maven to run them in this phase.

So does Maven just find and run these classes if they exist? Of course not, you have to add some configuration to your pom.xml so it knows. Fair enough, Maven shouldn't waste our precious build time searching all of our test classes for ones named with "IT" if we aren't using integration tests.

So, it should be a pretty simple configuration setting to tell Maven we're using the integration test phase. Right?

Right?

<project>
  [...]
  <build>
    <plugins>
      <plugin>
        <groupId>org.apache.maven.plugins</groupId>
        <artifactId>maven-failsafe-plugin</artifactId>
        <version>2.7.2</version>
        <executions>
          <execution>
            <id>integration-test</id>
            <goals>
              <goal>integration-test</goal>
            </goals>
          </execution>
          <execution>
            <id>verify</id>
            <goals>
              <goal>verify</goal>
            </goals>
          </execution>
        </executions>
      </plugin>
    </plugins>
  </build>
  [...]
</project>

(From the failsafe plugin docs.)

So, no.

Holy scaffolding! All we've really told Maven with all that configuration is to include the "failsafe plugin", and wire it into the build lifecycle to run at the integration-test phase. But again, we've locked the plugin details, including the version number, into our project's pom file.

In this case, we've also had to explicitly (and verbosely) specify the lifecycle phase. So this is a plugin whose whole purpose in life is to be used in the integration phase of Maven's build lifecycle, abut we have to take it by the hand and tell it exactly how and where to wire itself into the lifecycle, even though we're following its default use case.

That isn't configuration over convention! That isn't even Mexico!

Oh, wait, this doesn't run our integration tests with a servlet container. Let's plug in Jetty. It's actually pretty easy, just copy, paste and tweak the example in this documenation from Codehaus (the second XML snippet on that page).

There is even a bit more meat in the configuration, which is fair enough, since there are a few things you might like to configure when running a servlet container. But you've still got loads more configuration, much of which is calling out details of Maven's execution lifecyle that are, when it comes down to it, the default for 90% of the people using this.

So you can see how your Maven project's pom.xml file quickly becomes packed with unneeded crap, which increases the barrier to understanding and modifying the build. Once you develop a pretty good understanding of Maven's internal structures and models you become adept at working out which parts of the pom.xml are scaffolding, and which you actually care about.

But most developers on a team won't invest the time to master Maven, and shouldn't need to. They have business value to deliver. This is the wall I'm hitting supporting a high performance development team, because there is a low tolerance for spending time on things other than building the application features customers need.

So is Ant any better than this? As I said in my previous post, Ant builds typically become over-complicated. But Maven projects not only suffer from complexity, they also suffer from Maven's inflexibility. Normally you'd consider trading off some flexibility for simplicity, as long as the end result was higher productivity.

So my next Maven post will focus on a specific area where Maven's inflexibility is hurting my team's productivity. I don't currently have more than these three posts in mind for this series, but then my team is still only in iteration 0.

Feedback on these posts: I'm more than happy to be told I'm wrong, particularly if there are easier, simpler, and better ways to implement the examples I've described. Since I've given up on blog comments (for the time being at least), the best way to give feedback is to mention me (@kief) in a Tweet, with a link to your response.

I've been building Java software using Ant for over 10 years. I've been giving Maven a try every few years since it first came out, and going back to Ant pretty quickly each time, until last year. Early last year I used Maven on a few smallish - pretty much one-person - projects, and found it to be pretty useful, so I decided it may be ready for prime time.

Recently I've begun working on my first build system for a ThoughtWorks project, which happens to be using Maven, and although I've been enthusiastic about the challenge, the weaknesses of the tool are showing pretty clearly.

The idea of Maven is a good one - it's a standardised build system that, off the shelf, does pretty much what you need for typical Java projects, using convention over configuration. Ant, on the other hand, is not a build system, it's a tool that you can use to create your own build system. It's almost a DSL for build systems, but far more flexible.

The thing is, at least 95% of Java projects have pretty much the same set of requirements from a build system. Get my dependencies. Compile my source. Run my unit tests. Run some analysis. Package my artefact. Run integration tests. Deploy. Build multiple modules into a single aggregated artefact. That's pretty much it.

So everyone who uses Ant is using it to do pretty much the same thing, but they create wildly different build systems to do it. As a result, new team members have a serious learning curve, and often require days or even weeks to get set up. Adding tools such as a new code analysis report is not a simple drop-in, and the build system's complexity tends to grow over time.

A typical project's Ant-based build system is as complex, and as time consuming to maintain, as any major component of the software it builds. It accumulates technical debt and adds drag on the project's velocity.

So rather than every team having it's own home-brewed special snowflake build system, Maven aims to be a build system that uses convention over configuration to build any Java project that follows a common pattern. It's become popular enough that other tools can just be dropped in, and a developer joining a Maven-based project could be up and running in less than a day, in ideal cases in under an hour.

Ideally, by using Maven on our project we eliminate the hassle of maintaining a complex build system, because Maven simply does what we need, by default.

This is a worthy goal. Unfortunately, Maven falls short. I'll share why I think this is in part 2 and part 3 of this series.

From Tom DeMarco's article Software Engineering:
An Idea Whose Time Has Come and Gone? [PDF]:

You say to your team leads, "I have a finish date in mind, and I'm not even going to share it with you. When I come in one day and tell you the project will end in one week, you have to be ready to package up and deliver what you've got as the final product. Your job is to go about the project incrementally, adding pieces to the whole in the order of their relative value, and doing integration and documentation and acceptance testing incrementally as you go."

DeMarco isn't recommending specific methodologies like Agile, but this is a pretty good business oriented description of Continuous Delivery without continuous (production) deployment.