search

enso-org / enso

Enso Analytics is a self-service data prep and analysis platform designed for data teams.
https://ensoanalytics.com
Apache License 2.0
7.38k stars 323 forks source link

MacOS engine tests fail on CI #8408

Closed hubertp closed 11 months ago

hubertp commented 11 months ago 
 INFO ide_ci::program::command: sbtℹ️ [info] - should run visualization error preprocessor *** FAILED ***
 INFO ide_ci::program::command: sbtℹ️ [info]   List(Response(Some(b19aa5d2-a062-4368-8c99-a7a20618d55b), VisualizationAttached()), Response(Some(b19aa5d2-a062-4368-8c99-a7a20618d55b), VisualizationAttached())) did not contain all of (Response(Some(b19aa5d2-a062-4368-8c99-a7a20618d55b), VisualizationAttached()), Response(None, ExecutionComplete(ed5705c6-c90e-4052-a545-5a4c06fa1fa9))) (RuntimeVisualizationsTest.scala:2432)
 INFO ide_ci::program::command: sbt⚠️ 
 INFO ide_ci::program::command: sbt⚠️ Exception: java.lang.OutOfMemoryError thrown from the UncaughtExceptionHandler in thread "job-pool-1"
 INFO ide_ci::program::command: sbt⚠️ 
 INFO ide_ci::program::command: sbt⚠️ Exception: java.lang.OutOfMemoryError thrown from the UncaughtExceptionHandler in thread "pool-1-thread-1"

e.g. https://github.com/enso-org/enso/actions/runs/6999022093/job/19037773235?pr=8392

enso-bot[bot] commented 11 months ago

Hubert Plociniczak reports a new STANDUP for yesterday (2023-11-28):

Progress: Something is preventing IR and Metadata from being GC'ed which I believe to be especially painful on MacOS runners that OOM. Managed to reduce the explosion of threads being created during test runs. Needs more investigation. Created #8407 as a potential fix to build native macos builds. Some planning discussions. It should be finished by 2023-11-30.

Next Day: Next day I will be working on the #8408 task. Continue investigation.

enso-bot[bot] commented 11 months ago

Hubert Plociniczak reports a new STANDUP for yesterday (2023-11-29):

Progress: Found the leak in shutdown hooks. Removing it cuts heap space by 2-3 times during a regular test run. PR got updated. It should be finished by 2023-11-30.

Next Day: Next day I will be working on the #8408 task. Find any other place where we could be leaking in tests and not closing Context properly.

enso-bot[bot] commented 11 months ago

Hubert Plociniczak reports a new STANDUP for yesterday (2023-11-30):

Progress: Worked on a follow up PR that eliminated further potential leaks (#8432). Also trying to address problems in M1 build (#8407). It should be finished by 2023-11-30.

Next Day: Next day I will be working on the #8408 task. Pick up next item on the list