Skip to content

Commit f494c33

Browse files
Run tests in Gitlab (#8656)
* first pass checks * add some debug * Revert "add some debug" This reverts commit 26a7adb. * Normalize node index and node total * fix normalization * actually use the normalized values * first pass: migrate tests to Gitlab * less scenarios for now * bump memory and workers * actually use task partitions * Add section start/end * Allow individual tests to control gradle memory * remove a little of the reportspam * use concrete docker image tags * configure test containers to use mirror * fix limit * fix missing container versions * Better gitlab naming with smarter split * update caches with test dep cache, save test results * Cleanup repositories {} blocks * update proxy url to use internal routing * no "s" * allow insecure everywhere * include apache with confluence * remove intentional fail * use gradlePluginCache, mavenCentral after pluginportal * proxy mule as well * change BUILD_CACHE_TYPE to CACHE_TYPE since its used for both * split cache like with circleci. Cache became too big * Some caches are too big to be saved * test fixes * more test fixes * disable some tests for now * more fixes * Use maven proxy with openliberty smoketests * missed one more mongodb case * enable some more of the matrix * Fix lettuce tests * Fix lettuce tests # Conflicts: # dd-java-agent/instrumentation/lettuce-5/src/test/groovy/Lettuce5AsyncClientTest.groovy # dd-java-agent/instrumentation/lettuce-5/src/test/groovy/Lettuce5ReactiveClientTest.groovy # dd-java-agent/instrumentation/lettuce-5/src/test/groovy/Lettuce5SyncClientTest.groovy * cleanup imports * create asyncCommands * codenarc * Cleanup aggregator tests to try to reduce flakiness * debug maven logs for now * try non-ssh url * Revert "debug maven logs for now" This reverts commit c5c158f. * add some debug logging * codenarc * try fixed value of parallelism * enable debug logs for smoketest building only * limit number of reserved threads * some debug for why the smoketest is not running * more debug for gradle launcher failure * try fixing graal test * try to force the logging * codenarc * set available processors for jetty * codenarc * bluntly bump max threads * some more debug * remove verbose logging * try memcached test with new location * opt out of split by parallel * Try with 3 gradle workers * try decreasing reporting interval * Refractor CrashTrackingSmokeTest to use OutputThreads * Simple forked memory test * better ergonomics * try increasing timeout * extend agent test runner * don't override forkedMinHeapSize and forkedMaxHeapSize * print out requests * remove logging statement * Better crash tracking * remove intentionally failing test * use a test UDP server to record oom events * Make crash tracking smoke test more resiliant * loop through messages until the correct event * some debug * configurable dogstatsd port * enable everything * use sleep instead of park nanos * save body data ahead of time * retry once for system failures * spotless * cleanup * spotless/merge issues * Skip ExceptionHistogramTest on J9 * add a print statement around the body for debugging * comment out all but 8, 17, and 21 for now * fix spring 6 forwarding headers * revert logging statements * needs to be set in 3.1 for latestDepTest * faster OOM * property was set in incorrect location * ssi_smoke, flaky, and esoteric jvms on master only * use normal runner when TestContainers isn't needed * skip forwarding header filter here too * tweaks * limit executor size and fix IPV4 assumptions * try more workers with less memory * set max memory for all crashtracking tests * double retry * scale up memory, less jobs * revert memory bump, increase forked tests * tweaks * add flag for nondefault jvms and flaky tests * correct description * Externalize runtime processors override * ExceptionHistogramTest exclusions already handled by master * increase spring native build memory * bump gradle build timeout * bump graalvm compile memory --------- Co-authored-by: Santiago Mola <santiago.mola@datadoghq.com>
1 parent c81bba3 commit f494c33

File tree

52 files changed

+567
-136
lines changed

Some content is hidden

Large Commits have some content hidden by default. Use the searchbox below for content that may be hidden.

52 files changed

+567
-136
lines changed

.circleci/collect_reports.sh

Lines changed: 10 additions & 10 deletions
Original file line numberDiff line numberDiff line change
@@ -35,12 +35,12 @@ done
3535

3636
mkdir -p $REPORTS_DIR >/dev/null 2>&1
3737

38-
cp /tmp/hs_err_pid*.log $REPORTS_DIR || true
39-
cp /tmp/java_pid*.hprof $REPORTS_DIR || true
40-
cp /tmp/javacore.* $REPORTS_DIR || true
41-
cp /tmp/*.trc $REPORTS_DIR || true
42-
cp /tmp/*.dmp $REPORTS_DIR || true
43-
cp /tmp/dd-profiler/*.jfr $REPORTS_DIR || true
38+
cp /tmp/hs_err_pid*.log $REPORTS_DIR 2>/dev/null || true
39+
cp /tmp/java_pid*.hprof $REPORTS_DIR 2>/dev/null || true
40+
cp /tmp/javacore.* $REPORTS_DIR 2>/dev/null || true
41+
cp /tmp/*.trc $REPORTS_DIR 2>/dev/null || true
42+
cp /tmp/*.dmp $REPORTS_DIR 2>/dev/null || true
43+
cp /tmp/dd-profiler/*.jfr $REPORTS_DIR 2>/dev/null || true
4444

4545
function process_reports () {
4646
project_to_save=$1
@@ -59,9 +59,9 @@ function process_reports () {
5959
else
6060
echo "copying reports for $project_to_save"
6161
mkdir -p $report_path
62-
cp -r workspace/$project_to_save/build/reports/* $report_path/ || true
63-
cp workspace/$project_to_save/build/hs_err_pid*.log $report_path/ || true
64-
cp workspace/$project_to_save/build/javacore*.txt $report_path/ || true
62+
cp -r workspace/$project_to_save/build/reports/* $report_path/ 2>/dev/null || true
63+
cp workspace/$project_to_save/build/hs_err_pid*.log $report_path/ 2>/dev/null || true
64+
cp workspace/$project_to_save/build/javacore*.txt $report_path/ 2>/dev/null || true
6565
fi
6666
}
6767

@@ -73,4 +73,4 @@ for report_path in workspace/**/build/reports; do
7373
process_reports $report_path
7474
done
7575

76-
tar -cvzf reports.tar $REPORTS_DIR
76+
tar -czf reports.tar $REPORTS_DIR

0 commit comments

Comments
 (0)