Integration test script failures

[n-jay]

Describe the bug
run_integration_test.sh script fails causing CI pipeline to fail.

To Reproduce
Run the run_integration_test.sh script or initiate the packaging-pipeline workflow in Github Actions.

Expected behavior
Integration test script runs successfully causing packaging-pipeline workflow to succeed.

Screenshots

Version Number
0.25

Additional context
Failure is triggered by the following line in the integration test script:

OpenCue/ci/run_integration_test.sh

Line 89 in 9a033a9

container=$(docker compose ps --all --format json | jq ".[] | select(.Service==\"$1\")")

Cause is docker compose ps --format json no longer returning a JSON array after v2.21,

[n-jay]

@DiegoTavares there seems to be another issue popping up with the script's database migration portion after the jq command issue was resolved.

Any thoughts on what the reason might be for mismatch in values? Will create a separate PR to fix that.

OpenCue/ci/run_integration_test.sh

Lines 116 to 123 in 9a033a9

	verify_migration_versions() {
	migrations_in_db=$(docker compose exec -e PGUSER=cuebot db psql -Aqtc "SELECT COUNT(*) FROM flyway_schema_history")
	migrations_in_code=$(ls cuebot/src/main/resources/conf/ddl/postgres/migrations/ | wc -l | tr -d ' ')
	if [[ ${migrations_in_db} = ${migrations_in_code} ]]; then
	log INFO "Database and code both contain ${migrations_in_db} migrations (PASS)"
	else
	log ERROR "Database contains ${migrations_in_db} migrations, code contains ${migrations_in_code} (FAIL)"
	exit 1

[bcipriano]

Any thoughts on what the reason might be for mismatch in values?

You'll want to dump the two migration lists to see where the discrepancy is -- something like SELECT * FROM flyway_schema_history to dump the full list from the db, and the list of files in migrations/.

I seem to remember running into this in the past but I don't remember the ultimate cause.

[n-jay]

You'll want to dump the two migration lists to see where the discrepancy is

Noted @bcipriano. Will try that out. Thanks!

EDIT:
Seems like V18_Add_New_Indexes is the odd one out from the list of files:

Should this be removed from the migrations directory or added to the flyway_schema_history?

[DiegoTavares]

Part os the issue is fixed. There seems to no longer exist a migration issue. but the pipeline is still failing on the cuebot build step. The error I see when running manually is:

Invocation of init method failed; nested exception is java.lang.IllegalStateException: Missing embedded postgres binaries

Looking online, the exception I'm getting locally might be caused by running on the apple silicon.

[n-jay]

@DiegoTavares I was just attempting to run the script to see if everything was working. 😅

but the pipeline is still failing on the cuebot build step. The error I see when running manually is:

I feel like its better we keep this issue open until we get the whole thing running successfully.
I'll try and see what this exception is about on my end as well.

EDIT:
By Cuebot build step do you mean the Dockerfile build?

OpenCue/ci/run_integration_test.sh

Line 227 in b3e433b

docker build -t opencue/cuebot -f cuebot/Dockerfile . &>"${TEST_LOGS}/docker-build-cuebot.log"

[DiegoTavares]

@DiegoTavares I was just attempting to run the script to see if everything was working. 😅

but the pipeline is still failing on the cuebot build step. The error I see when running manually is:

I feel like its better we keep this issue open until we get the whole thing running successfully. I'll try and see what this exception is about on my end as well.

EDIT: By Cuebot build step do you mean the Dockerfile build?

OpenCue/ci/run_integration_test.sh

Line 227 in b3e433b

docker build -t opencue/cuebot -f cuebot/Dockerfile . &>"${TEST_LOGS}/docker-build-cuebot.log"

Yes

[DiegoTavares]

Now looking at the error on the repo pipeline, it happens after the cuebot docker build step, so was looking at the wrong thing. The error is on the test_pycue() on run_integration_test. Fix coming soon

[n-jay]

@DiegoTavares did you run this locally? Seems like that step is passing in the CI pipeline with the latest changes.
https://github.com/AcademySoftwareFoundation/OpenCue/actions/runs/9602218280/job/26482575373?pr=1309
According to this the error seem to be with the jq command. Seems to be within the test_pycue() function, but have to verify.

EDIT:

Looking online, the exception I'm getting locally might be caused by running on the apple silicon.

Saw this now. Seems like that might be the case.

[DiegoTavares]

Integration tests got fixed by #1392
Now working on the Sonar pipeline failure

[DiegoTavares]

Based on the sonar exception, it looks like the version of sonarqube we are using is incompatible with java 8:

org/sonar/batch/bootstrapper/EnvironmentInformation has been compiled by a more recent version of the Java Runtime (class file version 61.0), this version of the Java Runtime only recognizes class file versions up to 55.0

@bcipriano Do you know if there has been any changes on the sonar pipeline recently that might have caused this incompatibility?

[bcipriano]

Commented over on your Sonarcloud PR -- no changes that I know of but it looks like the image we're using is pretty old so if it's easy to just upgrade the image let's do it.