druid/.travis.yml

680 lines
33 KiB
YAML
Raw Normal View History

# Licensed to the Apache Software Foundation (ASF) under one or more
# contributor license agreements. See the NOTICE file distributed with
# this work for additional information regarding copyright ownership.
# The ASF licenses this file to You under the Apache License, Version 2.0
# (the "License"); you may not use this file except in compliance with
# the License. You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
branches:
only:
- master
- /^\d+\.\d+\.\d+(-\S*)?$/ # release branches
2014-12-19 18:53:19 -05:00
language: java
dist: xenial
2014-12-19 18:53:19 -05:00
jdk:
- openjdk8
cache:
directories:
- $HOME/.m2
Speedup Travis CI jobs (#8240) Reorganize Travis CI jobs into smaller faster (and more) jobs. Add various maven options to skip unnecessary work and refactored Travis CI job definitions to follow DRY. Detailed changes: .travis.yml - Refactor build logic to get rid of copy-and-paste logic - Skip static checks and enable parallelism for maven install - Split static analysis into different jobs to ease triage - Use "name" attribute instead of NAME environment variable - Split "indexing" and "web console" out of "other modules test" - Split 2 integration test jobs into multiple smaller jobs build.sh - Enable parallelism - Disable more static checks travis_script_integration.sh travis_script_integration_part2.sh integration-tests/README.md - Use TestNG groups instead of shell scripts and move definition of jobs into Travis CI yaml integration-tests/pom.xml - Show elapsed time of individual tests to aid in future rebalancing of Travis CI integration test jobs run time TestNGGroup.java - Use TestNG groups to make it easy to have multiple Travis CI integration test jobs. TestNG groups also make it easier to have an "other" integration test group and make it less likely a test will accidentally not be included in a CI job. IT*Test.java AbstractITBatchIndexTest.java AbstractKafkaIndexerTest.java - Add TestNG group - Fix various IntelliJ inspection warnings - Reduce scope of helper methods since the TestNG group annotation on the class makes TestNG consider all public methods as test methods pom.xml - Allow enforce plugin to be run from command-line - Bump resources plugin version so that "[debug] execute contextualize" output is correctly suppressed by "mvn -q" - Bump exec plugin version so that skip property is renamed from "skip" to "exec.skip" web-console/pom.xml - Add property to allow disabling javascript-related work. This property is overridden in Travis CI to speed up the jobs.
2019-08-07 12:52:42 -04:00
env:
global:
- DOCKER_IP=127.0.0.1 # for integration tests
- MVN="mvn -B"
- > # Various options to make execution of maven goals faster (e.g., mvn install)
MAVEN_SKIP="-P skip-static-checks -Dweb.console.skip=true -Dmaven.javadoc.skip=true"
- MAVEN_SKIP_TESTS="-P skip-tests"
Speedup Travis CI jobs (#8240) Reorganize Travis CI jobs into smaller faster (and more) jobs. Add various maven options to skip unnecessary work and refactored Travis CI job definitions to follow DRY. Detailed changes: .travis.yml - Refactor build logic to get rid of copy-and-paste logic - Skip static checks and enable parallelism for maven install - Split static analysis into different jobs to ease triage - Use "name" attribute instead of NAME environment variable - Split "indexing" and "web console" out of "other modules test" - Split 2 integration test jobs into multiple smaller jobs build.sh - Enable parallelism - Disable more static checks travis_script_integration.sh travis_script_integration_part2.sh integration-tests/README.md - Use TestNG groups instead of shell scripts and move definition of jobs into Travis CI yaml integration-tests/pom.xml - Show elapsed time of individual tests to aid in future rebalancing of Travis CI integration test jobs run time TestNGGroup.java - Use TestNG groups to make it easy to have multiple Travis CI integration test jobs. TestNG groups also make it easier to have an "other" integration test group and make it less likely a test will accidentally not be included in a CI job. IT*Test.java AbstractITBatchIndexTest.java AbstractKafkaIndexerTest.java - Add TestNG group - Fix various IntelliJ inspection warnings - Reduce scope of helper methods since the TestNG group annotation on the class makes TestNG consider all public methods as test methods pom.xml - Allow enforce plugin to be run from command-line - Bump resources plugin version so that "[debug] execute contextualize" output is correctly suppressed by "mvn -q" - Bump exec plugin version so that skip property is renamed from "skip" to "exec.skip" web-console/pom.xml - Add property to allow disabling javascript-related work. This property is overridden in Travis CI to speed up the jobs.
2019-08-07 12:52:42 -04:00
addons:
apt:
packages:
- maven
- python3
# Add various options to make 'mvn install' fast and skip javascript compile (-Dweb.console.skip=true) since it is not
# needed. Depending on network speeds, "mvn -q install" may take longer than the default 10 minute timeout to print any
# output. To compensate, use travis_wait to extend the timeout.
install: ./check_test_suite.py && travis_terminate 0 || echo 'Running Maven install...' && MAVEN_OPTS='-Xmx3000m' travis_wait 15 ${MVN} clean install -q -ff -pl '!distribution,!:druid-it-image,!:druid-it-cases' ${MAVEN_SKIP} ${MAVEN_SKIP_TESTS} -T1C && ${MVN} install -q -ff -pl 'distribution' ${MAVEN_SKIP} ${MAVEN_SKIP_TESTS}
Speedup Travis CI jobs (#8240) Reorganize Travis CI jobs into smaller faster (and more) jobs. Add various maven options to skip unnecessary work and refactored Travis CI job definitions to follow DRY. Detailed changes: .travis.yml - Refactor build logic to get rid of copy-and-paste logic - Skip static checks and enable parallelism for maven install - Split static analysis into different jobs to ease triage - Use "name" attribute instead of NAME environment variable - Split "indexing" and "web console" out of "other modules test" - Split 2 integration test jobs into multiple smaller jobs build.sh - Enable parallelism - Disable more static checks travis_script_integration.sh travis_script_integration_part2.sh integration-tests/README.md - Use TestNG groups instead of shell scripts and move definition of jobs into Travis CI yaml integration-tests/pom.xml - Show elapsed time of individual tests to aid in future rebalancing of Travis CI integration test jobs run time TestNGGroup.java - Use TestNG groups to make it easy to have multiple Travis CI integration test jobs. TestNG groups also make it easier to have an "other" integration test group and make it less likely a test will accidentally not be included in a CI job. IT*Test.java AbstractITBatchIndexTest.java AbstractKafkaIndexerTest.java - Add TestNG group - Fix various IntelliJ inspection warnings - Reduce scope of helper methods since the TestNG group annotation on the class makes TestNG consider all public methods as test methods pom.xml - Allow enforce plugin to be run from command-line - Bump resources plugin version so that "[debug] execute contextualize" output is correctly suppressed by "mvn -q" - Bump exec plugin version so that skip property is renamed from "skip" to "exec.skip" web-console/pom.xml - Add property to allow disabling javascript-related work. This property is overridden in Travis CI to speed up the jobs.
2019-08-07 12:52:42 -04:00
2021-02-05 20:50:29 -05:00
# There are 3 stages of tests
# 1. Tests - phase 1 (completely migrated to github workflows)
2021-02-05 20:50:29 -05:00
# 2. Tests - phase 2
# 3. cron
#
# The cron type only runs jobs that are marked with stage cron. The cron stage also runs alongside
# phase 1 and phase 2 for commit builds on release branches.
2021-02-05 20:50:29 -05:00
# The test type is split into 2 stages. This is done so that more PRs can run their validation
# in parallel. The first phase is meant to include sanity test jobs. The jobs in this phase are
# meant to be fast. The second phase is meant to run all other tests. Cron stage does not run on pull requests.
2021-02-05 20:50:29 -05:00
# Jobs with known flaky tests should be put in the second phase since the second phase will not
# start if there are any failures in the second stage.
stages:
2021-02-05 20:50:29 -05:00
- name: Tests - phase 1
if: type != cron
2021-02-05 20:50:29 -05:00
- name: Tests - phase 2
if: type != cron
- name: cron
if: type = cron OR (type != pull_request AND branch != master)
jobs:
include:
- &test_processing_module
name: "(openjdk11) processing module test"
stage: Tests - phase 2
jdk: openjdk11
env:
Speedup Travis CI jobs (#8240) Reorganize Travis CI jobs into smaller faster (and more) jobs. Add various maven options to skip unnecessary work and refactored Travis CI job definitions to follow DRY. Detailed changes: .travis.yml - Refactor build logic to get rid of copy-and-paste logic - Skip static checks and enable parallelism for maven install - Split static analysis into different jobs to ease triage - Use "name" attribute instead of NAME environment variable - Split "indexing" and "web console" out of "other modules test" - Split 2 integration test jobs into multiple smaller jobs build.sh - Enable parallelism - Disable more static checks travis_script_integration.sh travis_script_integration_part2.sh integration-tests/README.md - Use TestNG groups instead of shell scripts and move definition of jobs into Travis CI yaml integration-tests/pom.xml - Show elapsed time of individual tests to aid in future rebalancing of Travis CI integration test jobs run time TestNGGroup.java - Use TestNG groups to make it easy to have multiple Travis CI integration test jobs. TestNG groups also make it easier to have an "other" integration test group and make it less likely a test will accidentally not be included in a CI job. IT*Test.java AbstractITBatchIndexTest.java AbstractKafkaIndexerTest.java - Add TestNG group - Fix various IntelliJ inspection warnings - Reduce scope of helper methods since the TestNG group annotation on the class makes TestNG consider all public methods as test methods pom.xml - Allow enforce plugin to be run from command-line - Bump resources plugin version so that "[debug] execute contextualize" output is correctly suppressed by "mvn -q" - Bump exec plugin version so that skip property is renamed from "skip" to "exec.skip" web-console/pom.xml - Add property to allow disabling javascript-related work. This property is overridden in Travis CI to speed up the jobs.
2019-08-07 12:52:42 -04:00
- MAVEN_PROJECTS='processing'
before_script:
- export DRUID_USE_DEFAULT_VALUE_FOR_NULL=true
script:
Speedup Travis CI jobs (#8240) Reorganize Travis CI jobs into smaller faster (and more) jobs. Add various maven options to skip unnecessary work and refactored Travis CI job definitions to follow DRY. Detailed changes: .travis.yml - Refactor build logic to get rid of copy-and-paste logic - Skip static checks and enable parallelism for maven install - Split static analysis into different jobs to ease triage - Use "name" attribute instead of NAME environment variable - Split "indexing" and "web console" out of "other modules test" - Split 2 integration test jobs into multiple smaller jobs build.sh - Enable parallelism - Disable more static checks travis_script_integration.sh travis_script_integration_part2.sh integration-tests/README.md - Use TestNG groups instead of shell scripts and move definition of jobs into Travis CI yaml integration-tests/pom.xml - Show elapsed time of individual tests to aid in future rebalancing of Travis CI integration test jobs run time TestNGGroup.java - Use TestNG groups to make it easy to have multiple Travis CI integration test jobs. TestNG groups also make it easier to have an "other" integration test group and make it less likely a test will accidentally not be included in a CI job. IT*Test.java AbstractITBatchIndexTest.java AbstractKafkaIndexerTest.java - Add TestNG group - Fix various IntelliJ inspection warnings - Reduce scope of helper methods since the TestNG group annotation on the class makes TestNG consider all public methods as test methods pom.xml - Allow enforce plugin to be run from command-line - Bump resources plugin version so that "[debug] execute contextualize" output is correctly suppressed by "mvn -q" - Bump exec plugin version so that skip property is renamed from "skip" to "exec.skip" web-console/pom.xml - Add property to allow disabling javascript-related work. This property is overridden in Travis CI to speed up the jobs.
2019-08-07 12:52:42 -04:00
- unset _JAVA_OPTIONS
# Set MAVEN_OPTS for Surefire launcher. Skip remoteresources to avoid intermittent connection timeouts when
# resolving the SIGAR dependency.
- >
MAVEN_OPTS='-Xmx2048m' ${MVN} test -pl ${MAVEN_PROJECTS}
${MAVEN_SKIP} -Dremoteresources.skip=true -Ddruid.generic.useDefaultValueForNull=${DRUID_USE_DEFAULT_VALUE_FOR_NULL}
- sh -c "dmesg | egrep -i '(oom|out of memory|kill process|killed).*' -C 1 || exit 0"
- free -m
- ${MVN} -pl ${MAVEN_PROJECTS} jacoco:report
# Add merge target branch to determine diff (see https://github.com/travis-ci/travis-ci/issues/6069).
# This is not needed for build triggered by tags, since there will be no code diff.
- echo "TRAVIS_BRANCH=${TRAVIS_BRANCH}" # for debugging
- if [[ -z "${TRAVIS_TAG}" ]]; then git remote set-branches --add origin ${TRAVIS_BRANCH} && git fetch; fi
# Determine the modified files that match the maven projects being tested. We use maven project lists that
# either exclude (starts with "!") or include (does not start with "!"), so both cases need to be handled.
# If the build is triggered by a tag, an error will be printed, but `all_files` will be correctly set to empty
# so that the coverage check is skipped.
- all_files="$(git diff --name-only origin/${TRAVIS_BRANCH}...HEAD | grep "\.java$" || [[ $? == 1 ]])"
- for f in ${all_files}; do echo $f; done # for debugging
- >
if [[ "${MAVEN_PROJECTS}" = \!* ]]; then
regex="${MAVEN_PROJECTS:1}";
regex="^${regex//,\!/\\|^}";
project_files="$(echo "${all_files}" | grep -v "${regex}" || [[ $? == 1 ]])";
else
regex="^${MAVEN_PROJECTS//,/\\|^}";
project_files="$(echo "${all_files}" | grep "${regex}" || [[ $? == 1 ]])";
fi
- for f in ${project_files}; do echo $f; done # for debugging
# Check diff code coverage for the maven projects being tested (retry install in case of network error).
# Currently, the function coverage check is not reliable, so it is disabled.
- >
if [ -n "${project_files}" ]; then
travis_retry npm install @connectis/diff-test-coverage@1.5.3
&& git diff origin/${TRAVIS_BRANCH}...HEAD -- ${project_files}
| node_modules/.bin/diff-test-coverage
--coverage "**/target/site/jacoco/jacoco.xml"
--type jacoco
--line-coverage 50
--branch-coverage 50
--function-coverage 0
--log-template "coverage-lines-complete"
--log-template "coverage-files-complete"
--log-template "totals-complete"
--log-template "errors"
--
|| { printf "\n\n****FAILED****\nDiff code coverage check failed. To view coverage report, run 'mvn clean test jacoco:report' and open 'target/site/jacoco/index.html'\nFor more details on how to run code coverage locally, follow instructions here - https://github.com/apache/druid/blob/master/dev/code-review/code-coverage.md#running-code-coverage-locally\n\n" && false; }
fi
after_success:
# retry in case of network error
- travis_retry curl -o codecov.sh -s https://codecov.io/bash
- travis_retry bash codecov.sh -X gcov
2017-06-26 21:51:48 -04:00
- <<: *test_processing_module
name: "(openjdk17) processing module test"
stage: Tests - phase 2
jdk: openjdk17
- &test_processing_module_sqlcompat
<<: *test_processing_module
name: "(openjdk11) processing module test (SQL Compatibility)"
2021-02-05 20:50:29 -05:00
stage: Tests - phase 2
jdk: openjdk11
before_script: &setup_sqlcompat
- export DRUID_USE_DEFAULT_VALUE_FOR_NULL=false
- <<: *test_processing_module_sqlcompat
name: "(openjdk17) processing module test (SQL Compatibility)"
stage: Tests - phase 2
jdk: openjdk17
- &test_indexing_module
<<: *test_processing_module
name: "(openjdk11) indexing modules test"
2021-02-05 20:50:29 -05:00
stage: Tests - phase 2
jdk: openjdk11
env:
- MAVEN_PROJECTS='indexing-hadoop,indexing-service,extensions-core/kafka-indexing-service,extensions-core/kinesis-indexing-service'
- <<: *test_indexing_module
name: "(openjdk17) indexing modules test"
stage: Tests - phase 2
jdk: openjdk17
- &test_indexing_module_sqlcompat
<<: *test_indexing_module
name: "(openjdk11) indexing modules test (SQL Compatibility)"
2021-02-05 20:50:29 -05:00
stage: Tests - phase 2
jdk: openjdk11
before_script: *setup_sqlcompat
- <<: *test_indexing_module_sqlcompat
name: "(openjdk17) indexing modules test (SQL Compatibility)"
stage: Tests - phase 2
jdk: openjdk17
- &test_server_module
<<: *test_processing_module
name: "(openjdk11) server module test"
2021-02-05 20:50:29 -05:00
stage: Tests - phase 2
jdk: openjdk11
env:
- MAVEN_PROJECTS='server'
- <<: *test_server_module
name: "(openjdk17) server module test"
stage: Tests - phase 2
jdk: openjdk17
- &test_server_module_sqlcompat
<<: *test_server_module
name: "(openjdk11) server module test (SQL Compatibility)"
2021-02-05 20:50:29 -05:00
stage: Tests - phase 2
jdk: openjdk11
before_script: *setup_sqlcompat
- <<: *test_server_module_sqlcompat
name: "(openjdk17) server module test (SQL Compatibility)"
stage: Tests - phase 2
jdk: openjdk17
- &test_other_modules
<<: *test_processing_module
name: "(openjdk11) other modules test"
2021-02-05 20:50:29 -05:00
stage: Tests - phase 2
jdk: openjdk11
env:
- MAVEN_PROJECTS='!processing,!indexing-hadoop,!indexing-service,!extensions-core/kafka-indexing-service,!extensions-core/kinesis-indexing-service,!server,!web-console,!integration-tests,!:druid-it-tools,!:druid-it-image,!:druid-it-cases'
- <<: *test_other_modules
name: "(openjdk17) other modules test"
stage: Tests - phase 2
jdk: openjdk17
- &test_other_modules_sqlcompat
<<: *test_other_modules
name: "(openjdk11) other modules test (SQL Compatibility)"
2021-02-05 20:50:29 -05:00
stage: Tests - phase 2
jdk: openjdk11
before_script: *setup_sqlcompat
- <<: *test_other_modules_sqlcompat
name: "(openjdk17) other modules test (SQL Compatibility)"
stage: Tests - phase 2
jdk: openjdk17
- name: "Build and test on ARM64 CPU architecture (1)"
stage: Tests - phase 2
arch: arm64-graviton2
dist: focal
virt: vm
group: edge
jdk: openjdk11
env:
- MAVEN_PROJECTS='core,indexing-hadoop,indexing-service,processing'
script: ${MVN} test -B -pl ${MAVEN_PROJECTS} -Dweb.console.skip=true -DargLine=-Xmx3000m
- name: "Build and test on ARM64 CPU architecture (2)"
stage: Tests - phase 2
arch: arm64-graviton2
dist: focal
virt: vm
group: edge
jdk: openjdk11
env:
- MAVEN_PROJECTS='core,sql,server,services'
script: ${MVN} test -B -pl ${MAVEN_PROJECTS} -Dweb.console.skip=true -DargLine=-Xmx3000m
# Integration tests Java Compile version is set by the machine environment jdk (set by the jdk key)
# Integration tests Java Runtime version is set by the JVM_RUNTIME env property (set env key to -Djvm.runtime=<JVM_RUNTIME_VERSION>)
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
# Integration tests will either use MiddleManagers or Indexers
# (Currently integration tests only support running with jvm runtime 8 and 11)
# START - Integration tests for Compile with Java 8 and Run with Java 8
- &integration_batch_index
name: "(Compile=openjdk8, Run=openjdk8) batch index integration test"
2021-02-05 20:50:29 -05:00
stage: Tests - phase 2
jdk: openjdk8
Speedup Travis CI jobs (#8240) Reorganize Travis CI jobs into smaller faster (and more) jobs. Add various maven options to skip unnecessary work and refactored Travis CI job definitions to follow DRY. Detailed changes: .travis.yml - Refactor build logic to get rid of copy-and-paste logic - Skip static checks and enable parallelism for maven install - Split static analysis into different jobs to ease triage - Use "name" attribute instead of NAME environment variable - Split "indexing" and "web console" out of "other modules test" - Split 2 integration test jobs into multiple smaller jobs build.sh - Enable parallelism - Disable more static checks travis_script_integration.sh travis_script_integration_part2.sh integration-tests/README.md - Use TestNG groups instead of shell scripts and move definition of jobs into Travis CI yaml integration-tests/pom.xml - Show elapsed time of individual tests to aid in future rebalancing of Travis CI integration test jobs run time TestNGGroup.java - Use TestNG groups to make it easy to have multiple Travis CI integration test jobs. TestNG groups also make it easier to have an "other" integration test group and make it less likely a test will accidentally not be included in a CI job. IT*Test.java AbstractITBatchIndexTest.java AbstractKafkaIndexerTest.java - Add TestNG group - Fix various IntelliJ inspection warnings - Reduce scope of helper methods since the TestNG group annotation on the class makes TestNG consider all public methods as test methods pom.xml - Allow enforce plugin to be run from command-line - Bump resources plugin version so that "[debug] execute contextualize" output is correctly suppressed by "mvn -q" - Bump exec plugin version so that skip property is renamed from "skip" to "exec.skip" web-console/pom.xml - Add property to allow disabling javascript-related work. This property is overridden in Travis CI to speed up the jobs.
2019-08-07 12:52:42 -04:00
services: &integration_test_services
- docker
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
env: TESTNG_GROUPS='-Dgroups=batch-index' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='middleManager'
Speedup Travis CI jobs (#8240) Reorganize Travis CI jobs into smaller faster (and more) jobs. Add various maven options to skip unnecessary work and refactored Travis CI job definitions to follow DRY. Detailed changes: .travis.yml - Refactor build logic to get rid of copy-and-paste logic - Skip static checks and enable parallelism for maven install - Split static analysis into different jobs to ease triage - Use "name" attribute instead of NAME environment variable - Split "indexing" and "web console" out of "other modules test" - Split 2 integration test jobs into multiple smaller jobs build.sh - Enable parallelism - Disable more static checks travis_script_integration.sh travis_script_integration_part2.sh integration-tests/README.md - Use TestNG groups instead of shell scripts and move definition of jobs into Travis CI yaml integration-tests/pom.xml - Show elapsed time of individual tests to aid in future rebalancing of Travis CI integration test jobs run time TestNGGroup.java - Use TestNG groups to make it easy to have multiple Travis CI integration test jobs. TestNG groups also make it easier to have an "other" integration test group and make it less likely a test will accidentally not be included in a CI job. IT*Test.java AbstractITBatchIndexTest.java AbstractKafkaIndexerTest.java - Add TestNG group - Fix various IntelliJ inspection warnings - Reduce scope of helper methods since the TestNG group annotation on the class makes TestNG consider all public methods as test methods pom.xml - Allow enforce plugin to be run from command-line - Bump resources plugin version so that "[debug] execute contextualize" output is correctly suppressed by "mvn -q" - Bump exec plugin version so that skip property is renamed from "skip" to "exec.skip" web-console/pom.xml - Add property to allow disabling javascript-related work. This property is overridden in Travis CI to speed up the jobs.
2019-08-07 12:52:42 -04:00
script: &run_integration_test
- ${MVN} verify -pl integration-tests -P integration-tests ${TESTNG_GROUPS} ${JVM_RUNTIME} -Dit.indexer=${USE_INDEXER} ${MAVEN_SKIP} -Doverride.config.path=${OVERRIDE_CONFIG_PATH}
Speedup Travis CI jobs (#8240) Reorganize Travis CI jobs into smaller faster (and more) jobs. Add various maven options to skip unnecessary work and refactored Travis CI job definitions to follow DRY. Detailed changes: .travis.yml - Refactor build logic to get rid of copy-and-paste logic - Skip static checks and enable parallelism for maven install - Split static analysis into different jobs to ease triage - Use "name" attribute instead of NAME environment variable - Split "indexing" and "web console" out of "other modules test" - Split 2 integration test jobs into multiple smaller jobs build.sh - Enable parallelism - Disable more static checks travis_script_integration.sh travis_script_integration_part2.sh integration-tests/README.md - Use TestNG groups instead of shell scripts and move definition of jobs into Travis CI yaml integration-tests/pom.xml - Show elapsed time of individual tests to aid in future rebalancing of Travis CI integration test jobs run time TestNGGroup.java - Use TestNG groups to make it easy to have multiple Travis CI integration test jobs. TestNG groups also make it easier to have an "other" integration test group and make it less likely a test will accidentally not be included in a CI job. IT*Test.java AbstractITBatchIndexTest.java AbstractKafkaIndexerTest.java - Add TestNG group - Fix various IntelliJ inspection warnings - Reduce scope of helper methods since the TestNG group annotation on the class makes TestNG consider all public methods as test methods pom.xml - Allow enforce plugin to be run from command-line - Bump resources plugin version so that "[debug] execute contextualize" output is correctly suppressed by "mvn -q" - Bump exec plugin version so that skip property is renamed from "skip" to "exec.skip" web-console/pom.xml - Add property to allow disabling javascript-related work. This property is overridden in Travis CI to speed up the jobs.
2019-08-07 12:52:42 -04:00
after_failure: &integration_test_diags
- for v in ~/shared/logs/*.log ; do
echo $v logtail ======================== ; tail -100 $v ;
done
- for v in broker middlemanager overlord router coordinator historical ; do
echo $v dmesg ======================== ;
docker exec -it druid-$v sh -c 'dmesg | tail -3' ;
done
- &integration_input_format
name: "(Compile=openjdk8, Run=openjdk8) input format integration test"
2021-02-05 20:50:29 -05:00
stage: Tests - phase 2
jdk: openjdk8
services: *integration_test_services
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
env: TESTNG_GROUPS='-Dgroups=input-format' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='middleManager'
script: *run_integration_test
after_failure: *integration_test_diags
- &integration_input_source
name: "(Compile=openjdk8, Run=openjdk8) input source integration test"
2021-02-05 20:50:29 -05:00
stage: Tests - phase 2
jdk: openjdk8
services: *integration_test_services
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
env: TESTNG_GROUPS='-Dgroups=input-source' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='middleManager'
script: *run_integration_test
after_failure: *integration_test_diags
Parallel indexing single dim partitions (#8925) * Parallel indexing single dim partitions Implements single dimension range partitioning for native parallel batch indexing as described in #8769. This initial version requires the druid-datasketches extension to be loaded. The algorithm has 5 phases that are orchestrated by the supervisor in `ParallelIndexSupervisorTask#runRangePartitionMultiPhaseParallel()`. These phases and the main classes involved are described below: 1) In parallel, determine the distribution of dimension values for each input source split. `PartialDimensionDistributionTask` uses `StringSketch` to generate the approximate distribution of dimension values for each input source split. If the rows are ungrouped, `PartialDimensionDistributionTask.UngroupedRowDimensionValueFilter` uses a Bloom filter to skip rows that would be grouped. The final distribution is sent back to the supervisor via `DimensionDistributionReport`. 2) The range partitions are determined. In `ParallelIndexSupervisorTask#determineAllRangePartitions()`, the supervisor uses `StringSketchMerger` to merge the individual `StringSketch`es created in the preceding phase. The merged sketch is then used to create the range partitions. 3) In parallel, generate partial range-partitioned segments. `PartialRangeSegmentGenerateTask` uses the range partitions determined in the preceding phase and `RangePartitionCachingLocalSegmentAllocator` to generate `SingleDimensionShardSpec`s. The partition information is sent back to the supervisor via `GeneratedGenericPartitionsReport`. 4) The partial range segments are grouped. In `ParallelIndexSupervisorTask#groupGenericPartitionLocationsPerPartition()`, the supervisor creates the `PartialGenericSegmentMergeIOConfig`s necessary for the next phase. 5) In parallel, merge partial range-partitioned segments. `PartialGenericSegmentMergeTask` uses `GenericPartitionLocation` to retrieve the partial range-partitioned segments generated earlier and then merges and publishes them. * Fix dependencies & forbidden apis * Fixes for integration test * Address review comments * Fix docs, strict compile, sketch check, rollup check * Fix first shard spec, partition serde, single subtask * Fix first partition check in test * Misc rewording/refactoring to address code review * Fix doc link * Split batch index integration test * Do not run parallel-batch-index twice * Adjust last partition * Split ITParallelIndexTest to reduce runtime * Rename test class * Allow null values in range partitions * Indicate which phase failed * Improve asserts in tests
2019-12-10 02:05:49 -05:00
- &integration_perfect_rollup_parallel_batch_index
name: "(Compile=openjdk8, Run=openjdk8) perfect rollup parallel batch index integration test"
jdk: openjdk8
2021-02-05 20:50:29 -05:00
stage: Tests - phase 2
Parallel indexing single dim partitions (#8925) * Parallel indexing single dim partitions Implements single dimension range partitioning for native parallel batch indexing as described in #8769. This initial version requires the druid-datasketches extension to be loaded. The algorithm has 5 phases that are orchestrated by the supervisor in `ParallelIndexSupervisorTask#runRangePartitionMultiPhaseParallel()`. These phases and the main classes involved are described below: 1) In parallel, determine the distribution of dimension values for each input source split. `PartialDimensionDistributionTask` uses `StringSketch` to generate the approximate distribution of dimension values for each input source split. If the rows are ungrouped, `PartialDimensionDistributionTask.UngroupedRowDimensionValueFilter` uses a Bloom filter to skip rows that would be grouped. The final distribution is sent back to the supervisor via `DimensionDistributionReport`. 2) The range partitions are determined. In `ParallelIndexSupervisorTask#determineAllRangePartitions()`, the supervisor uses `StringSketchMerger` to merge the individual `StringSketch`es created in the preceding phase. The merged sketch is then used to create the range partitions. 3) In parallel, generate partial range-partitioned segments. `PartialRangeSegmentGenerateTask` uses the range partitions determined in the preceding phase and `RangePartitionCachingLocalSegmentAllocator` to generate `SingleDimensionShardSpec`s. The partition information is sent back to the supervisor via `GeneratedGenericPartitionsReport`. 4) The partial range segments are grouped. In `ParallelIndexSupervisorTask#groupGenericPartitionLocationsPerPartition()`, the supervisor creates the `PartialGenericSegmentMergeIOConfig`s necessary for the next phase. 5) In parallel, merge partial range-partitioned segments. `PartialGenericSegmentMergeTask` uses `GenericPartitionLocation` to retrieve the partial range-partitioned segments generated earlier and then merges and publishes them. * Fix dependencies & forbidden apis * Fixes for integration test * Address review comments * Fix docs, strict compile, sketch check, rollup check * Fix first shard spec, partition serde, single subtask * Fix first partition check in test * Misc rewording/refactoring to address code review * Fix doc link * Split batch index integration test * Do not run parallel-batch-index twice * Adjust last partition * Split ITParallelIndexTest to reduce runtime * Rename test class * Allow null values in range partitions * Indicate which phase failed * Improve asserts in tests
2019-12-10 02:05:49 -05:00
services: *integration_test_services
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
env: TESTNG_GROUPS='-Dgroups=perfect-rollup-parallel-batch-index' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='middleManager'
Parallel indexing single dim partitions (#8925) * Parallel indexing single dim partitions Implements single dimension range partitioning for native parallel batch indexing as described in #8769. This initial version requires the druid-datasketches extension to be loaded. The algorithm has 5 phases that are orchestrated by the supervisor in `ParallelIndexSupervisorTask#runRangePartitionMultiPhaseParallel()`. These phases and the main classes involved are described below: 1) In parallel, determine the distribution of dimension values for each input source split. `PartialDimensionDistributionTask` uses `StringSketch` to generate the approximate distribution of dimension values for each input source split. If the rows are ungrouped, `PartialDimensionDistributionTask.UngroupedRowDimensionValueFilter` uses a Bloom filter to skip rows that would be grouped. The final distribution is sent back to the supervisor via `DimensionDistributionReport`. 2) The range partitions are determined. In `ParallelIndexSupervisorTask#determineAllRangePartitions()`, the supervisor uses `StringSketchMerger` to merge the individual `StringSketch`es created in the preceding phase. The merged sketch is then used to create the range partitions. 3) In parallel, generate partial range-partitioned segments. `PartialRangeSegmentGenerateTask` uses the range partitions determined in the preceding phase and `RangePartitionCachingLocalSegmentAllocator` to generate `SingleDimensionShardSpec`s. The partition information is sent back to the supervisor via `GeneratedGenericPartitionsReport`. 4) The partial range segments are grouped. In `ParallelIndexSupervisorTask#groupGenericPartitionLocationsPerPartition()`, the supervisor creates the `PartialGenericSegmentMergeIOConfig`s necessary for the next phase. 5) In parallel, merge partial range-partitioned segments. `PartialGenericSegmentMergeTask` uses `GenericPartitionLocation` to retrieve the partial range-partitioned segments generated earlier and then merges and publishes them. * Fix dependencies & forbidden apis * Fixes for integration test * Address review comments * Fix docs, strict compile, sketch check, rollup check * Fix first shard spec, partition serde, single subtask * Fix first partition check in test * Misc rewording/refactoring to address code review * Fix doc link * Split batch index integration test * Do not run parallel-batch-index twice * Adjust last partition * Split ITParallelIndexTest to reduce runtime * Rename test class * Allow null values in range partitions * Indicate which phase failed * Improve asserts in tests
2019-12-10 02:05:49 -05:00
script: *run_integration_test
after_failure: *integration_test_diags
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
- <<: *integration_perfect_rollup_parallel_batch_index
name: "(Compile=openjdk8, Run=openjdk8) perfect rollup parallel batch index integration test with Indexer"
env: TESTNG_GROUPS='-Dgroups=perfect-rollup-parallel-batch-index' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='indexer'
- <<: *integration_perfect_rollup_parallel_batch_index
name: "(Compile=openjdk8, Run=openjdk8) perfect rollup parallel batch index integration test with deep storage as intermediate store"
env: TESTNG_GROUPS='-Dgroups=shuffle-deep-store' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='middleManager' OVERRIDE_CONFIG_PATH='./environment-configs/test-groups/shuffle-deep-store'
- <<: *integration_perfect_rollup_parallel_batch_index
name: "(Compile=openjdk8, Run=openjdk8) perfect rollup parallel batch index integration test with deep storage as intermediate store with indexer"
env: TESTNG_GROUPS='-Dgroups=shuffle-deep-store' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='indexer' OVERRIDE_CONFIG_PATH='./environment-configs/test-groups/shuffle-deep-store'
- &integration_kafka_index
name: "(Compile=openjdk8, Run=openjdk8) kafka index integration test"
2021-02-05 20:50:29 -05:00
stage: Tests - phase 2
jdk: openjdk8
Speedup Travis CI jobs (#8240) Reorganize Travis CI jobs into smaller faster (and more) jobs. Add various maven options to skip unnecessary work and refactored Travis CI job definitions to follow DRY. Detailed changes: .travis.yml - Refactor build logic to get rid of copy-and-paste logic - Skip static checks and enable parallelism for maven install - Split static analysis into different jobs to ease triage - Use "name" attribute instead of NAME environment variable - Split "indexing" and "web console" out of "other modules test" - Split 2 integration test jobs into multiple smaller jobs build.sh - Enable parallelism - Disable more static checks travis_script_integration.sh travis_script_integration_part2.sh integration-tests/README.md - Use TestNG groups instead of shell scripts and move definition of jobs into Travis CI yaml integration-tests/pom.xml - Show elapsed time of individual tests to aid in future rebalancing of Travis CI integration test jobs run time TestNGGroup.java - Use TestNG groups to make it easy to have multiple Travis CI integration test jobs. TestNG groups also make it easier to have an "other" integration test group and make it less likely a test will accidentally not be included in a CI job. IT*Test.java AbstractITBatchIndexTest.java AbstractKafkaIndexerTest.java - Add TestNG group - Fix various IntelliJ inspection warnings - Reduce scope of helper methods since the TestNG group annotation on the class makes TestNG consider all public methods as test methods pom.xml - Allow enforce plugin to be run from command-line - Bump resources plugin version so that "[debug] execute contextualize" output is correctly suppressed by "mvn -q" - Bump exec plugin version so that skip property is renamed from "skip" to "exec.skip" web-console/pom.xml - Add property to allow disabling javascript-related work. This property is overridden in Travis CI to speed up the jobs.
2019-08-07 12:52:42 -04:00
services: *integration_test_services
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
env: TESTNG_GROUPS='-Dgroups=kafka-index' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='middleManager'
Speedup Travis CI jobs (#8240) Reorganize Travis CI jobs into smaller faster (and more) jobs. Add various maven options to skip unnecessary work and refactored Travis CI job definitions to follow DRY. Detailed changes: .travis.yml - Refactor build logic to get rid of copy-and-paste logic - Skip static checks and enable parallelism for maven install - Split static analysis into different jobs to ease triage - Use "name" attribute instead of NAME environment variable - Split "indexing" and "web console" out of "other modules test" - Split 2 integration test jobs into multiple smaller jobs build.sh - Enable parallelism - Disable more static checks travis_script_integration.sh travis_script_integration_part2.sh integration-tests/README.md - Use TestNG groups instead of shell scripts and move definition of jobs into Travis CI yaml integration-tests/pom.xml - Show elapsed time of individual tests to aid in future rebalancing of Travis CI integration test jobs run time TestNGGroup.java - Use TestNG groups to make it easy to have multiple Travis CI integration test jobs. TestNG groups also make it easier to have an "other" integration test group and make it less likely a test will accidentally not be included in a CI job. IT*Test.java AbstractITBatchIndexTest.java AbstractKafkaIndexerTest.java - Add TestNG group - Fix various IntelliJ inspection warnings - Reduce scope of helper methods since the TestNG group annotation on the class makes TestNG consider all public methods as test methods pom.xml - Allow enforce plugin to be run from command-line - Bump resources plugin version so that "[debug] execute contextualize" output is correctly suppressed by "mvn -q" - Bump exec plugin version so that skip property is renamed from "skip" to "exec.skip" web-console/pom.xml - Add property to allow disabling javascript-related work. This property is overridden in Travis CI to speed up the jobs.
2019-08-07 12:52:42 -04:00
script: *run_integration_test
after_failure: *integration_test_diags
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
- <<: *integration_kafka_index
name: "(Compile=openjdk8, Run=openjdk8) kafka index, transactional kafka index integration test with Indexer"
env: TESTNG_GROUPS='-Dgroups=kafka-index,kafka-transactional-index' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='indexer'
- <<: *integration_kafka_index
name: "(Compile=openjdk8, Run=openjdk8) custom coordinator duties integration test"
env: TESTNG_GROUPS='-Dgroups=custom-coordinator-duties' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='middleManager' OVERRIDE_CONFIG_PATH='./environment-configs/test-groups/custom-coordinator-duties'
- &integration_kafka_index_slow
name: "(Compile=openjdk8, Run=openjdk8) kafka index integration test slow"
2021-02-05 20:50:29 -05:00
stage: Tests - phase 2
jdk: openjdk8
services: *integration_test_services
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
env: TESTNG_GROUPS='-Dgroups=kafka-index-slow' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='middleManager'
script: *run_integration_test
after_failure: *integration_test_diags
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
- <<: *integration_kafka_index_slow
name: "(Compile=openjdk8, Run=openjdk8) kafka index integration test slow with Indexer"
env: TESTNG_GROUPS='-Dgroups=kafka-index-slow' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='indexer'
- &integration_kafka_transactional_index
name: "(Compile=openjdk8, Run=openjdk8) transactional kafka index integration test"
2021-02-05 20:50:29 -05:00
stage: Tests - phase 2
jdk: openjdk8
services: *integration_test_services
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
env: TESTNG_GROUPS='-Dgroups=kafka-transactional-index' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='middleManager'
script: *run_integration_test
after_failure: *integration_test_diags
- &integration_kafka_transactional_index_slow
name: "(Compile=openjdk8, Run=openjdk8) transactional kafka index integration test slow"
2021-02-05 20:50:29 -05:00
stage: Tests - phase 2
jdk: openjdk8
services: *integration_test_services
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
env: TESTNG_GROUPS='-Dgroups=kafka-transactional-index-slow' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='middleManager'
script: *run_integration_test
after_failure: *integration_test_diags
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
- <<: *integration_kafka_transactional_index_slow
name: "(Compile=openjdk8, Run=openjdk8) transactional kafka index integration test slow with Indexer"
env: TESTNG_GROUPS='-Dgroups=kafka-transactional-index-slow' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='indexer'
- &integration_kafka_format_tests
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
name: "(Compile=openjdk8, Run=openjdk8) Kafka index integration test with various formats"
2021-02-05 20:50:29 -05:00
stage: Tests - phase 2
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
jdk: openjdk8
services: *integration_test_services
env: TESTNG_GROUPS='-Dgroups=kafka-data-format' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='middleManager'
script: *run_integration_test
after_failure: *integration_test_diags
- <<: *integration_kafka_format_tests
name: "(Compile=openjdk8, Run=openjdk8) Kafka index integration test with various formats with Indexer"
env: TESTNG_GROUPS='-Dgroups=kafka-data-format' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='indexer'
- &integration_query
name: "(Compile=openjdk8, Run=openjdk8) query integration test"
2021-02-05 20:50:29 -05:00
stage: Tests - phase 2
jdk: openjdk8
Speedup Travis CI jobs (#8240) Reorganize Travis CI jobs into smaller faster (and more) jobs. Add various maven options to skip unnecessary work and refactored Travis CI job definitions to follow DRY. Detailed changes: .travis.yml - Refactor build logic to get rid of copy-and-paste logic - Skip static checks and enable parallelism for maven install - Split static analysis into different jobs to ease triage - Use "name" attribute instead of NAME environment variable - Split "indexing" and "web console" out of "other modules test" - Split 2 integration test jobs into multiple smaller jobs build.sh - Enable parallelism - Disable more static checks travis_script_integration.sh travis_script_integration_part2.sh integration-tests/README.md - Use TestNG groups instead of shell scripts and move definition of jobs into Travis CI yaml integration-tests/pom.xml - Show elapsed time of individual tests to aid in future rebalancing of Travis CI integration test jobs run time TestNGGroup.java - Use TestNG groups to make it easy to have multiple Travis CI integration test jobs. TestNG groups also make it easier to have an "other" integration test group and make it less likely a test will accidentally not be included in a CI job. IT*Test.java AbstractITBatchIndexTest.java AbstractKafkaIndexerTest.java - Add TestNG group - Fix various IntelliJ inspection warnings - Reduce scope of helper methods since the TestNG group annotation on the class makes TestNG consider all public methods as test methods pom.xml - Allow enforce plugin to be run from command-line - Bump resources plugin version so that "[debug] execute contextualize" output is correctly suppressed by "mvn -q" - Bump exec plugin version so that skip property is renamed from "skip" to "exec.skip" web-console/pom.xml - Add property to allow disabling javascript-related work. This property is overridden in Travis CI to speed up the jobs.
2019-08-07 12:52:42 -04:00
services: *integration_test_services
env: TESTNG_GROUPS='-Dgroups=query' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='middleManager' OVERRIDE_CONFIG_PATH='./environment-configs/test-groups/prepopulated-data'
Speedup Travis CI jobs (#8240) Reorganize Travis CI jobs into smaller faster (and more) jobs. Add various maven options to skip unnecessary work and refactored Travis CI job definitions to follow DRY. Detailed changes: .travis.yml - Refactor build logic to get rid of copy-and-paste logic - Skip static checks and enable parallelism for maven install - Split static analysis into different jobs to ease triage - Use "name" attribute instead of NAME environment variable - Split "indexing" and "web console" out of "other modules test" - Split 2 integration test jobs into multiple smaller jobs build.sh - Enable parallelism - Disable more static checks travis_script_integration.sh travis_script_integration_part2.sh integration-tests/README.md - Use TestNG groups instead of shell scripts and move definition of jobs into Travis CI yaml integration-tests/pom.xml - Show elapsed time of individual tests to aid in future rebalancing of Travis CI integration test jobs run time TestNGGroup.java - Use TestNG groups to make it easy to have multiple Travis CI integration test jobs. TestNG groups also make it easier to have an "other" integration test group and make it less likely a test will accidentally not be included in a CI job. IT*Test.java AbstractITBatchIndexTest.java AbstractKafkaIndexerTest.java - Add TestNG group - Fix various IntelliJ inspection warnings - Reduce scope of helper methods since the TestNG group annotation on the class makes TestNG consider all public methods as test methods pom.xml - Allow enforce plugin to be run from command-line - Bump resources plugin version so that "[debug] execute contextualize" output is correctly suppressed by "mvn -q" - Bump exec plugin version so that skip property is renamed from "skip" to "exec.skip" web-console/pom.xml - Add property to allow disabling javascript-related work. This property is overridden in Travis CI to speed up the jobs.
2019-08-07 12:52:42 -04:00
script: *run_integration_test
after_failure: *integration_test_diags
- &integration_query_retry
name: "(Compile=openjdk8, Run=openjdk8) query retry integration test for missing segments"
2021-02-05 20:50:29 -05:00
stage: Tests - phase 2
jdk: openjdk8
services: *integration_test_services
env: TESTNG_GROUPS='-Dgroups=query-retry' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='middleManager' OVERRIDE_CONFIG_PATH='./environment-configs/test-groups/prepopulated-data'
script: *run_integration_test
after_failure: *integration_test_diags
- &integration_query_error
name: "(Compile=openjdk8, Run=openjdk8) query error integration test"
stage: Tests - phase 2
jdk: openjdk8
services: *integration_test_services
env: TESTNG_GROUPS='-Dgroups=query-error' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='middleManager' OVERRIDE_CONFIG_PATH='./environment-configs/test-groups/prepopulated-data'
script: *run_integration_test
after_failure: *integration_test_diags
- &integration_security
name: "(Compile=openjdk8, Run=openjdk8) security integration test"
2021-02-05 20:50:29 -05:00
stage: Tests - phase 2
jdk: openjdk8
services: *integration_test_services
env: TESTNG_GROUPS='-Dgroups=security' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='middleManager' OVERRIDE_CONFIG_PATH='./environment-configs/test-groups/prepopulated-data'
script: *run_integration_test
after_failure: *integration_test_diags
- &integration_ldap_security
name: "(Compile=openjdk8, Run=openjdk8) ldap security integration test"
stage: Tests - phase 2
jdk: openjdk8
services: *integration_test_services
env: TESTNG_GROUPS='-Dgroups=ldap-security' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='middleManager'
script: *run_integration_test
after_failure: *integration_test_diags
- &integration_realtime_index
name: "(Compile=openjdk8, Run=openjdk8) realtime index integration test"
2021-02-05 20:50:29 -05:00
stage: Tests - phase 2
jdk: openjdk8
Speedup Travis CI jobs (#8240) Reorganize Travis CI jobs into smaller faster (and more) jobs. Add various maven options to skip unnecessary work and refactored Travis CI job definitions to follow DRY. Detailed changes: .travis.yml - Refactor build logic to get rid of copy-and-paste logic - Skip static checks and enable parallelism for maven install - Split static analysis into different jobs to ease triage - Use "name" attribute instead of NAME environment variable - Split "indexing" and "web console" out of "other modules test" - Split 2 integration test jobs into multiple smaller jobs build.sh - Enable parallelism - Disable more static checks travis_script_integration.sh travis_script_integration_part2.sh integration-tests/README.md - Use TestNG groups instead of shell scripts and move definition of jobs into Travis CI yaml integration-tests/pom.xml - Show elapsed time of individual tests to aid in future rebalancing of Travis CI integration test jobs run time TestNGGroup.java - Use TestNG groups to make it easy to have multiple Travis CI integration test jobs. TestNG groups also make it easier to have an "other" integration test group and make it less likely a test will accidentally not be included in a CI job. IT*Test.java AbstractITBatchIndexTest.java AbstractKafkaIndexerTest.java - Add TestNG group - Fix various IntelliJ inspection warnings - Reduce scope of helper methods since the TestNG group annotation on the class makes TestNG consider all public methods as test methods pom.xml - Allow enforce plugin to be run from command-line - Bump resources plugin version so that "[debug] execute contextualize" output is correctly suppressed by "mvn -q" - Bump exec plugin version so that skip property is renamed from "skip" to "exec.skip" web-console/pom.xml - Add property to allow disabling javascript-related work. This property is overridden in Travis CI to speed up the jobs.
2019-08-07 12:52:42 -04:00
services: *integration_test_services
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
env: TESTNG_GROUPS='-Dgroups=realtime-index' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='middleManager'
Speedup Travis CI jobs (#8240) Reorganize Travis CI jobs into smaller faster (and more) jobs. Add various maven options to skip unnecessary work and refactored Travis CI job definitions to follow DRY. Detailed changes: .travis.yml - Refactor build logic to get rid of copy-and-paste logic - Skip static checks and enable parallelism for maven install - Split static analysis into different jobs to ease triage - Use "name" attribute instead of NAME environment variable - Split "indexing" and "web console" out of "other modules test" - Split 2 integration test jobs into multiple smaller jobs build.sh - Enable parallelism - Disable more static checks travis_script_integration.sh travis_script_integration_part2.sh integration-tests/README.md - Use TestNG groups instead of shell scripts and move definition of jobs into Travis CI yaml integration-tests/pom.xml - Show elapsed time of individual tests to aid in future rebalancing of Travis CI integration test jobs run time TestNGGroup.java - Use TestNG groups to make it easy to have multiple Travis CI integration test jobs. TestNG groups also make it easier to have an "other" integration test group and make it less likely a test will accidentally not be included in a CI job. IT*Test.java AbstractITBatchIndexTest.java AbstractKafkaIndexerTest.java - Add TestNG group - Fix various IntelliJ inspection warnings - Reduce scope of helper methods since the TestNG group annotation on the class makes TestNG consider all public methods as test methods pom.xml - Allow enforce plugin to be run from command-line - Bump resources plugin version so that "[debug] execute contextualize" output is correctly suppressed by "mvn -q" - Bump exec plugin version so that skip property is renamed from "skip" to "exec.skip" web-console/pom.xml - Add property to allow disabling javascript-related work. This property is overridden in Travis CI to speed up the jobs.
2019-08-07 12:52:42 -04:00
script: *run_integration_test
after_failure: *integration_test_diags
- &integration_append_ingestion
name: "(Compile=openjdk8, Run=openjdk8) append ingestion integration test"
2021-02-05 20:50:29 -05:00
stage: Tests - phase 2
jdk: openjdk8
services: *integration_test_services
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
env: TESTNG_GROUPS='-Dgroups=append-ingestion' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='middleManager'
script: *run_integration_test
after_failure: *integration_test_diags
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
- <<: *integration_append_ingestion
name: "(Compile=openjdk8, Run=openjdk8) append ingestion integration test with Indexer"
env: TESTNG_GROUPS='-Dgroups=append-ingestion' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='indexer'
- &integration_compaction_tests
name: "(Compile=openjdk8, Run=openjdk8) compaction integration test"
2021-02-05 20:50:29 -05:00
stage: Tests - phase 2
jdk: openjdk8
services: *integration_test_services
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
env: TESTNG_GROUPS='-Dgroups=compaction' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='middleManager'
script: *run_integration_test
after_failure: *integration_test_diags
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
- <<: *integration_compaction_tests
name: "(Compile=openjdk8, Run=openjdk8) compaction integration test with Indexer"
env: TESTNG_GROUPS='-Dgroups=compaction' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='indexer'
- &integration_tests
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
name: "(Compile=openjdk8, Run=openjdk8) other integration tests"
2021-02-05 20:50:29 -05:00
stage: Tests - phase 2
jdk: openjdk8
Speedup Travis CI jobs (#8240) Reorganize Travis CI jobs into smaller faster (and more) jobs. Add various maven options to skip unnecessary work and refactored Travis CI job definitions to follow DRY. Detailed changes: .travis.yml - Refactor build logic to get rid of copy-and-paste logic - Skip static checks and enable parallelism for maven install - Split static analysis into different jobs to ease triage - Use "name" attribute instead of NAME environment variable - Split "indexing" and "web console" out of "other modules test" - Split 2 integration test jobs into multiple smaller jobs build.sh - Enable parallelism - Disable more static checks travis_script_integration.sh travis_script_integration_part2.sh integration-tests/README.md - Use TestNG groups instead of shell scripts and move definition of jobs into Travis CI yaml integration-tests/pom.xml - Show elapsed time of individual tests to aid in future rebalancing of Travis CI integration test jobs run time TestNGGroup.java - Use TestNG groups to make it easy to have multiple Travis CI integration test jobs. TestNG groups also make it easier to have an "other" integration test group and make it less likely a test will accidentally not be included in a CI job. IT*Test.java AbstractITBatchIndexTest.java AbstractKafkaIndexerTest.java - Add TestNG group - Fix various IntelliJ inspection warnings - Reduce scope of helper methods since the TestNG group annotation on the class makes TestNG consider all public methods as test methods pom.xml - Allow enforce plugin to be run from command-line - Bump resources plugin version so that "[debug] execute contextualize" output is correctly suppressed by "mvn -q" - Bump exec plugin version so that skip property is renamed from "skip" to "exec.skip" web-console/pom.xml - Add property to allow disabling javascript-related work. This property is overridden in Travis CI to speed up the jobs.
2019-08-07 12:52:42 -04:00
services: *integration_test_services
env: TESTNG_GROUPS='-DexcludedGroups=batch-index,input-format,input-source,perfect-rollup-parallel-batch-index,kafka-index,query,query-retry,query-error,realtime-index,security,ldap-security,s3-deep-storage,gcs-deep-storage,azure-deep-storage,hdfs-deep-storage,s3-ingestion,kinesis-index,kinesis-data-format,kafka-transactional-index,kafka-index-slow,kafka-transactional-index-slow,kafka-data-format,hadoop-s3-to-s3-deep-storage,hadoop-s3-to-hdfs-deep-storage,hadoop-azure-to-azure-deep-storage,hadoop-azure-to-hdfs-deep-storage,hadoop-gcs-to-gcs-deep-storage,hadoop-gcs-to-hdfs-deep-storage,aliyun-oss-deep-storage,append-ingestion,compaction,high-availability,upgrade,shuffle-deep-store,custom-coordinator-duties' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='middleManager'
Speedup Travis CI jobs (#8240) Reorganize Travis CI jobs into smaller faster (and more) jobs. Add various maven options to skip unnecessary work and refactored Travis CI job definitions to follow DRY. Detailed changes: .travis.yml - Refactor build logic to get rid of copy-and-paste logic - Skip static checks and enable parallelism for maven install - Split static analysis into different jobs to ease triage - Use "name" attribute instead of NAME environment variable - Split "indexing" and "web console" out of "other modules test" - Split 2 integration test jobs into multiple smaller jobs build.sh - Enable parallelism - Disable more static checks travis_script_integration.sh travis_script_integration_part2.sh integration-tests/README.md - Use TestNG groups instead of shell scripts and move definition of jobs into Travis CI yaml integration-tests/pom.xml - Show elapsed time of individual tests to aid in future rebalancing of Travis CI integration test jobs run time TestNGGroup.java - Use TestNG groups to make it easy to have multiple Travis CI integration test jobs. TestNG groups also make it easier to have an "other" integration test group and make it less likely a test will accidentally not be included in a CI job. IT*Test.java AbstractITBatchIndexTest.java AbstractKafkaIndexerTest.java - Add TestNG group - Fix various IntelliJ inspection warnings - Reduce scope of helper methods since the TestNG group annotation on the class makes TestNG consider all public methods as test methods pom.xml - Allow enforce plugin to be run from command-line - Bump resources plugin version so that "[debug] execute contextualize" output is correctly suppressed by "mvn -q" - Bump exec plugin version so that skip property is renamed from "skip" to "exec.skip" web-console/pom.xml - Add property to allow disabling javascript-related work. This property is overridden in Travis CI to speed up the jobs.
2019-08-07 12:52:42 -04:00
script: *run_integration_test
after_failure: *integration_test_diags
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
- <<: *integration_tests
name: "(Compile=openjdk8, Run=openjdk8) other integration tests with Indexer"
env: TESTNG_GROUPS='-DexcludedGroups=batch-index,input-format,input-source,perfect-rollup-parallel-batch-index,kafka-index,query,query-retry,query-error,realtime-index,security,ldap-security,s3-deep-storage,gcs-deep-storage,azure-deep-storage,hdfs-deep-storage,s3-ingestion,kinesis-index,kinesis-data-format,kafka-transactional-index,kafka-index-slow,kafka-transactional-index-slow,kafka-data-format,hadoop-s3-to-s3-deep-storage,hadoop-s3-to-hdfs-deep-storage,hadoop-azure-to-azure-deep-storage,hadoop-azure-to-hdfs-deep-storage,hadoop-gcs-to-gcs-deep-storage,hadoop-gcs-to-hdfs-deep-storage,aliyun-oss-deep-storage,append-ingestion,compaction,high-availability,upgrade,shuffle-deep-store,custom-coordinator-duties' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='indexer'
#- <<: *integration_tests
# name: "(Compile=openjdk8, Run=openjdk8) leadership and high availability integration tests"
# jdk: openjdk8
# env: TESTNG_GROUPS='-Dgroups=high-availability' JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='middleManager' OVERRIDE_CONFIG_PATH='./environment-configs/test-groups/prepopulated-data'
- <<: *integration_query
name: "(Compile=openjdk8, Run=openjdk8) query integration test (mariaDB)"
jdk: openjdk8
env: TESTNG_GROUPS='-Dgroups=query' USE_INDEXER='middleManager' MYSQL_DRIVER_CLASSNAME='org.mariadb.jdbc.Driver' OVERRIDE_CONFIG_PATH='./environment-configs/test-groups/prepopulated-data'
# Revised ITs.
- &integration_tests_ex
name: "(Compile=openjdk8, Run=openjdk8) leadership and high availability integration tests (new)"
stage: Tests - phase 2
jdk: openjdk8
services: *integration_test_services
env: JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='middleManager'
# Uses the installation defined above. Then, builds the test tools and docker image,
# and runs one IT. If tests fail, echos log lines of any of
# the Druid services that did not exit normally.
script: ./it.sh travis HighAvailability
- &integration_tests_ex
name: "(Compile=openjdk8, Run=openjdk8) multi stage query tests"
stage: Tests - phase 2
jdk: openjdk8
services: *integration_test_services
env: JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='middleManager'
script: ./it.sh travis MultiStageQuery
2023-01-10 21:08:29 -05:00
- &integration_tests_ex
name: "(Compile=openjdk8, Run=openjdk8) multi stage query tests with MM"
stage: Tests - phase 2
jdk: openjdk8
services: *integration_test_services
env: JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='middleManager'
script: ./it.sh travis MultiStageQueryWithMM
- &integration_tests_ex
name: "(Compile=openjdk8, Run=openjdk8) catalog integration tests"
stage: Tests - phase 2
jdk: openjdk8
services: *integration_test_services
env: JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='middleManager'
script: ./it.sh travis Catalog
- &integration_tests_ex
name: "(Compile=openjdk8, Run=openjdk8) batch index integration test with Indexer (new)"
stage: Tests - phase 2
jdk: openjdk8
services: *integration_test_services
env: JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='indexer'
script: ./it.sh travis BatchIndex
- &integration_tests_ex
name: "(Compile=openjdk8, Run=openjdk8) input source integration test with Indexer (new)"
stage: Tests - phase 2
jdk: openjdk8
services: *integration_test_services
env: JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='indexer'
script: ./it.sh travis InputSource
- &integration_tests_ex
name: "(Compile=openjdk8, Run=openjdk8) input format integration test with Indexer (new)"
stage: Tests - phase 2
jdk: openjdk8
services: *integration_test_services
env: JVM_RUNTIME='-Djvm.runtime=8' USE_INDEXER='indexer'
script: ./it.sh travis InputFormat
# END - Integration tests for Compile with Java 8 and Run with Java 8
# START - Integration tests for Compile with Java 8 and Run with Java 11
- <<: *integration_batch_index
name: "(Compile=openjdk8, Run=openjdk11) batch index integration test"
jdk: openjdk8
stage: cron
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
env: TESTNG_GROUPS='-Dgroups=batch-index' JVM_RUNTIME='-Djvm.runtime=11' USE_INDEXER='middleManager'
- <<: *integration_input_format
name: "(Compile=openjdk8, Run=openjdk11) input format integration test"
jdk: openjdk8
stage: cron
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
env: TESTNG_GROUPS='-Dgroups=input-format' JVM_RUNTIME='-Djvm.runtime=11' USE_INDEXER='middleManager'
- <<: *integration_input_source
name: "(Compile=openjdk8, Run=openjdk11) input source integration test"
jdk: openjdk8
stage: cron
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
env: TESTNG_GROUPS='-Dgroups=input-source' JVM_RUNTIME='-Djvm.runtime=11' USE_INDEXER='middleManager'
- <<: *integration_perfect_rollup_parallel_batch_index
name: "(Compile=openjdk8, Run=openjdk11) perfect rollup parallel batch index integration test"
jdk: openjdk8
stage: cron
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
env: TESTNG_GROUPS='-Dgroups=perfect-rollup-parallel-batch-index' JVM_RUNTIME='-Djvm.runtime=11' USE_INDEXER='middleManager'
- <<: *integration_query
name: "(Compile=openjdk8, Run=openjdk11) query integration test"
jdk: openjdk8
env: TESTNG_GROUPS='-Dgroups=query' JVM_RUNTIME='-Djvm.runtime=11' USE_INDEXER='middleManager' OVERRIDE_CONFIG_PATH='./environment-configs/test-groups/prepopulated-data'
- <<: *integration_query_retry
name: "(Compile=openjdk8, Run=openjdk11) query retry integration test for missing segments"
jdk: openjdk8
stage: cron
env: TESTNG_GROUPS='-Dgroups=query-retry' JVM_RUNTIME='-Djvm.runtime=11' USE_INDEXER='middleManager' OVERRIDE_CONFIG_PATH='./environment-configs/test-groups/prepopulated-data'
- <<: *integration_query_error
name: "(Compile=openjdk8, Run=openjdk11) query error integration test for missing segments"
jdk: openjdk8
stage: cron
env: TESTNG_GROUPS='-Dgroups=query-error' JVM_RUNTIME='-Djvm.runtime=11' USE_INDEXER='middleManager' OVERRIDE_CONFIG_PATH='./environment-configs/test-groups/prepopulated-data'
- <<: *integration_security
name: "(Compile=openjdk8, Run=openjdk11) security integration test"
jdk: openjdk8
env: TESTNG_GROUPS='-Dgroups=security' JVM_RUNTIME='-Djvm.runtime=11' USE_INDEXER='middleManager' OVERRIDE_CONFIG_PATH='./environment-configs/test-groups/prepopulated-data'
- <<: *integration_ldap_security
name: "(Compile=openjdk8, Run=openjdk11) ldap security integration test"
jdk: openjdk8
env: TESTNG_GROUPS='-Dgroups=ldap-security' JVM_RUNTIME='-Djvm.runtime=11' USE_INDEXER='middleManager'
- <<: *integration_realtime_index
name: "(Compile=openjdk8, Run=openjdk11) realtime index integration test"
jdk: openjdk8
stage: cron
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
env: TESTNG_GROUPS='-Dgroups=realtime-index' JVM_RUNTIME='-Djvm.runtime=11' USE_INDEXER='middleManager'
- <<: *integration_append_ingestion
name: "(Compile=openjdk8, Run=openjdk11) append ingestion integration test"
jdk: openjdk8
stage: cron
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
env: TESTNG_GROUPS='-Dgroups=append-ingestion' JVM_RUNTIME='-Djvm.runtime=11' USE_INDEXER='middleManager'
- <<: *integration_compaction_tests
name: "(Compile=openjdk8, Run=openjdk11) compaction integration test"
jdk: openjdk8
stage: cron
Fixes and tests related to the Indexer process. (#10631) * Fixes and tests related to the Indexer process. Three bugs fixed: 1) Indexers would not announce themselves as segment servers if they did not have storage locations defined. This used to work, but was broken in #9971. Fixed this by adding an "isSegmentServer" method to ServerType and updating SegmentLoadDropHandler to always announce if this method returns true. 2) Certain batch task types were written in a way that assumed "isReady" would be called before "run", which is not guaranteed. In particular, they relied on it in order to initialize "taskLockHelper". Fixed this by updating AbstractBatchIndexTask to ensure "isReady" is called before "run" for these tasks. 3) UnifiedIndexerAppenderatorsManager did not properly handle complex datasources. Introduced DataSourceAnalysis in order to fix this. Test changes: 1) Add a new "docker-compose.cli-indexer.yml" config that spins up an Indexer instead of a MiddleManager. 2) Introduce a "USE_INDEXER" environment variable that determines if docker-compose will start up an Indexer or a MiddleManager. 3) Duplicate all the jdk8 tests and run them in both MiddleManager and Indexer mode. 4) Various adjustments to encourage fail-fast errors in the Docker build scripts. 5) Various adjustments to speed up integration tests and reduce memory usage. 6) Add another Mac-specific approach to determining a machine's own IP. This was useful on my development machine. 7) Update segment-count check in ITCompactionTaskTest to eliminate a race condition (it was looking for 6 segments, which only exist together briefly, until the older 4 are marked unused). Javadoc updates: 1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX that make it clear when taskLockHelper will be initialized as a side effect. (Related to the second bug above.) 2) Task: Clarified that "isReady" is not guaranteed to be called before "run". It was already implied, but now it's explicit. 3) ZkCoordinator: Clarified deprecation message. 4) DataSegmentServerAnnouncer: Clarified deprecation message. * Fix stop_cluster script. * Fix sanity check in script. * Fix hashbang lines. * Test and doc adjustments. * Additional tests, and adjustments for tests. * Split ITs back out. * Revert change to druid_coordinator_period_indexingPeriod. * Set Indexer capacity to match MM. * Bump up Historical memory. * Bump down coordinator, overlord memory. * Bump up Broker memory.
2020-12-08 19:02:26 -05:00
env: TESTNG_GROUPS='-Dgroups=compaction' JVM_RUNTIME='-Djvm.runtime=11' USE_INDEXER='middleManager'
- <<: *integration_tests
name: "(Compile=openjdk8, Run=openjdk11) other integration test"
jdk: openjdk8
stage: cron
env: TESTNG_GROUPS='-DexcludedGroups=batch-index,input-format,input-source,perfect-rollup-parallel-batch-index,kafka-index,query,query-retry,query-error,realtime-index,security,ldap-security,s3-deep-storage,gcs-deep-storage,azure-deep-storage,hdfs-deep-storage,s3-ingestion,kinesis-index,kinesis-data-format,kafka-transactional-index,kafka-index-slow,kafka-transactional-index-slow,kafka-data-format,hadoop-s3-to-s3-deep-storage,hadoop-s3-to-hdfs-deep-storage,hadoop-azure-to-azure-deep-storage,hadoop-azure-to-hdfs-deep-storage,hadoop-gcs-to-gcs-deep-storage,hadoop-gcs-to-hdfs-deep-storage,aliyun-oss-deep-storage,append-ingestion,compaction,high-availability,upgrade,shuffle-deep-store,custom-coordinator-duties' JVM_RUNTIME='-Djvm.runtime=11' USE_INDEXER='middleManager'
- <<: *integration_tests
name: "(Compile=openjdk8, Run=openjdk11) leadership and high availability integration tests"
jdk: openjdk8
env: TESTNG_GROUPS='-Dgroups=high-availability' JVM_RUNTIME='-Djvm.runtime=11' USE_INDEXER='middleManager' OVERRIDE_CONFIG_PATH='./environment-configs/test-groups/prepopulated-data'
- <<: *integration_query
name: "(Compile=openjdk8, Run=openjdk11) query integration test (mariaDB)"
jdk: openjdk8
env: TESTNG_GROUPS='-Dgroups=query' JVM_RUNTIME='-Djvm.runtime=11' USE_INDEXER='middleManager' MYSQL_DRIVER_CLASSNAME='org.mariadb.jdbc.Driver' OVERRIDE_CONFIG_PATH='./environment-configs/test-groups/prepopulated-data'
# END - Integration tests for Compile with Java 8 and Run with Java 11
- &integration_batch_index_k8s
name: "(Compile=openjdk8, Run=openjdk8, Cluster Build On K8s) ITNestedQueryPushDownTest integration test"
2021-02-05 20:50:29 -05:00
stage: Tests - phase 2
jdk: openjdk8
services: &integration_test_services_k8s
- docker
env: CONFIG_FILE='k8s_run_config_file.json' IT_TEST='-Dit.test=ITNestedQueryPushDownTest' POD_NAME=int-test POD_NAMESPACE=default BUILD_DRUID_CLSUTER=true
script: &run_integration_test_k8s
- ${MVN} verify -pl integration-tests -P int-tests-config-file ${IT_TEST} ${MAVEN_SKIP} -Dpod.name=${POD_NAME} -Dpod.namespace=${POD_NAMESPACE} -Dbuild.druid.cluster=${BUILD_DRUID_CLSUTER}
after_failure: &integration_test_diags_k8s
- for v in broker middlemanager router coordinator historical ; do
echo "------------------------druid-tiny-cluster-"$v"s-0-------------------------";
sudo /usr/local/bin/kubectl logs --tail 1000 druid-tiny-cluster-"$v"s-0;
done
- &security_vulnerabilities
name: "security vulnerabilities"
stage: cron
install: skip
script: |-
${MVN} dependency-check:purge dependency-check:check ${HADOOP_PROFILE} || { echo "
The OWASP dependency check has found security vulnerabilities. Please use a newer version
of the dependency that does not have vulnerabilities. To see a report run
`mvn dependency-check:check`
If the analysis has false positives,
they can be suppressed by adding entries to owasp-dependency-check-suppressions.xml (for more
information, see https://jeremylong.github.io/DependencyCheck/general/suppression.html).
" && false; }
- <<: *security_vulnerabilities
name: "security vulnerabilities with Hadoop3"
env:
- HADOOP_PROFILE='-Phadoop3'
# Travis CI only supports per build (and not per-job notifications): https://github.com/travis-ci/travis-ci/issues/9888
notifications:
email:
if: type = cron
recipients:
# This is the string "dev@druid.apache.org" encrypted against the apache/druid repo so that forks are unable to
# use this notification:
# https://github.com/travis-ci/travis-ci/issues/1094#issuecomment-215019909
# https://github.com/travis-ci/travis-ci/issues/2711
- secure: "MupjX/0jLwh3XzHPl74BTk2/Kp5r+8TrEewfRhpQdWKFMBXLKNqu0k2VXf5C/NIg3uvPianq3REk+qeTHI8dL2ShjiWS/eIRkJOHLfObdNNBuos5fo4TxAuBQcXyT4VjAq5jnAkH84Pxf2Nl0rkisWoIhvwSX7+kNrjW1qdu7K0="
on_success: change
on_failure: always