roadmap
- 01
- 02
- 03
- 04
- 05
- 06
- 07
- 08
Prepare for Apache Maven interview questions on how a project is described, built and published. Covers the POM, the lifecycle, dependency scopes and mediation, plugins and multi-module builds.
250 of 250 questions in your plan: 0 known, 0 in progress, 0 due, 0 skipped, 250 new. 0 cut, outside your plan.
Maven seldom gets an interview round to itself. It comes up inside a Java backend loop, usually when the talk turns to how the service is built and shipped, and a common opening is a conflict: two libraries pull in different versions of the same dependency, the build is green, and production throws `NoSuchMethodError`. The interviewer wants the mediation rule that picked the losing version (the nearest declaration wins, and the first one declared breaks a tie), how `dependency:tree` shows it, and where a pin belongs: in `dependencyManagement` or an imported BOM, not in exclusions scattered across modules. Scopes follow, with `provided` and `test` the usual traps. SDET loops come at the same tool from the test side. Why do integration tests belong to Failsafe rather than Surefire, and why does a failing one still let `post-integration-test` tear the environment down before `verify` fails the build? What is the difference between skipping test execution and skipping test compilation? Build, release and platform engineers get the deep end: a repository manager mirroring Central, signed releases, reproducible artifacts, a plugin written in-house. Two role roadmaps set Maven among the other questions of the loop: the [Java Backend Developer roadmap](/roadmaps/java-developer) next to Spring and persistence, the [Java SDET roadmap](/roadmaps/java-sdet) next to test frameworks and automation. If the team builds with Gradle, drill the [Gradle roadmap](/roadmaps/gradle) instead.
Before drilling, read the [primer on the Maven hub](/topics/dev-maven) once; it explains how the POM, the lifecycle and the plugins fit together. Most wrong Maven answers come from treating a phase as a goal, or a parent POM as an aggregator, and the primer keeps those apart. Start with [the lifecycle](/topics/dev-maven-lifecycle-phases) until it survives follow-ups: what `mvn clean install` actually runs, why naming one phase runs every phase before it, and what `install` adds to the local repository that `package` does not. That covers most of what a junior is asked. A middle-level conversation asks why a build resolved the versions it did, which pulls [transitive resolution and mediation](/topics/dev-maven-dependencies-transitivity) forward, ahead of anything about plugins. Candidates who will own a large build should reach [the reactor](/topics/dev-maven-multimodule) early; senior conversations tend to settle on a slow build of many modules, how to rebuild one of them together with what it needs (`-pl` with `-am`) or with what depends on it (`-amd`), and how `-rf` resumes after a failure. Since Maven usually fills a few minutes of a wider Java interview rather than a round, importance is the control worth moving: raised to must know, it leaves the lifecycle and dependency questions those minutes go on. Whoever will own the build should leave it at its full range.
Maven seldom gets an interview round to itself. It comes up inside a Java backend loop, usually when the talk turns to how the service is built and shipped, and a common opening is a conflict: two libraries pull in different versions of the same dependency, the build is green, and production throws NoSuchMethodError. The interviewer wants the mediation rule that picked the losing version (the nearest declaration wins, and the first one declared breaks a tie), how dependency:tree shows it, and where a pin belongs: in dependencyManagement or an imported BOM, not in exclusions scattered across modules. Scopes follow, with provided and test the usual traps.
SDET loops come at the same tool from the test side. Why do integration tests belong to Failsafe rather than Surefire, and why does a failing one still let post-integration-test tear the environment down before verify fails the build? What is the difference between skipping test execution and skipping test compilation? Build, release and platform engineers get the deep end: a repository manager mirroring Central, signed releases, reproducible artifacts, a plugin written in-house.
Two role roadmaps set Maven among the other questions of the loop: the Java Backend Developer roadmap next to Spring and persistence, the Java SDET roadmap next to test frameworks and automation. If the team builds with Gradle, drill the Gradle roadmap instead.
Before drilling, read the primer on the Maven hub once; it explains how the POM, the lifecycle and the plugins fit together. Most wrong Maven answers come from treating a phase as a goal, or a parent POM as an aggregator, and the primer keeps those apart.
Start with the lifecycle until it survives follow-ups: what mvn clean install actually runs, why naming one phase runs every phase before it, and what install adds to the local repository that package does not. That covers most of what a junior is asked. A middle-level conversation asks why a build resolved the versions it did, which pulls transitive resolution and mediation forward, ahead of anything about plugins. Candidates who will own a large build should reach the reactor early; senior conversations tend to settle on a slow build of many modules, how to rebuild one of them together with what it needs (-pl with -am) or with what depends on it (-amd), and how -rf resumes after a failure.
Since Maven usually fills a few minutes of a wider Java interview rather than a round, importance is the control worth moving: raised to must know, it leaves the lifecycle and dependency questions those minutes go on. Whoever will own the build should leave it at its full range.