https://linen.dev logo
Join Slack
Powered by
This content can't be displayed.
# oss-master-build-failure
g
This content can't be displayed.
e
So... what should I do since I broke the build? Looks like it has been broken for a while today.
Copy code
The RUN_DATABASE_MIGRATION_ON_STARTUP variable is not set. Defaulting to a blank string.
The DEPLOYMENT_MODE variable is not set. Defaulting to a blank string.
The SECRET_PERSISTENCE variable is not set. Defaulting to a blank string.
The WORKER_ENVIRONMENT variable is not set. Defaulting to a blank string.
The JOB_ERROR_REPORTING_SENTRY_DSN variable is not set. Defaulting to a blank string.
The NEW_SCHEDULER variable is not set. Defaulting to a blank string.
Removing network airbyte_default
Network airbyte_default not found.
Removing volume airbyte_workspace
Volume airbyte_workspace not found.
Removing volume airbyte_data
Volume airbyte_data not found.
Removing volume airbyte_db
Volume airbyte_db not found.
Error: Process completed with exit code 1.
https://github.com/airbytehq/airbyte/runs/7544137928?check_suite_focus=true
t
the test that’s failing is on line 876
Copy code
BasicAcceptanceTests > testDiscoverSourceSchema() FAILED
    org.opentest4j.AssertionFailedError: expected: <class AirbyteCatalog {
        streams: [class AirbyteStreamAndConfiguration {
            stream: class AirbyteStream {
                name: id_and_name
                jsonSchema: {"type":"object","properties":{"id":{"type":"number"},"name":{"type":"string"}}}
                supportedSyncModes: [full_refresh, incremental]
                sourceDefinedCursor: null
                defaultCursorField: []
                sourceDefinedPrimaryKey: []
                namespace: public
            }
            config: class AirbyteStreamConfiguration {
                syncMode: full_refresh
                cursorField: []
                destinationSyncMode: append
                primaryKey: []
                aliasName: id_and_name
                selected: true
            }
        }]
    }> but was: <class AirbyteCatalog {
        streams: [class AirbyteStreamAndConfiguration {
            stream: class AirbyteStream {
                name: id_and_name
                jsonSchema: {"type":"object","properties":{"name":{"type":"string"},"id":{"type":"number","airbyte_type":"integer"}}}
                supportedSyncModes: [full_refresh, incremental]
                sourceDefinedCursor: null
                defaultCursorField: []
                sourceDefinedPrimaryKey: []
                namespace: public
            }
            config: class AirbyteStreamConfiguration {
                syncMode: full_refresh
                cursorField: []
                destinationSyncMode: append
                primaryKey: []
                aliasName: id_and_name
                selected: true
            }
        }]
    }>
at app//io.airbyte.test.acceptance.BasicAcceptanceTests.testDiscoverSourceSchema(BasicAcceptanceTests.java:312)
looks like something changed and the acceptance tests were not updated. might be related to this PR?
@Edward Gao (Airbyte) fyi. edit: the only reason i think it might be related to your PR is because i searched for 
airbyte_type
which was the unexpected thing in the failed acceptance test
e
seems plausible, looks like my merge somehow didn’t trigger the normal master build actions
looking now
@Terence Cho (Airbyte) I tagged you on https://github.com/airbytehq/airbyte/pull/15078, hopefully that’s a reasonable approach I’m a bit confused because (a) I’m pretty sure my PR is the culprit, but (b) this test didn’t seem to fail on my PR? I.e. the acceptance tests failures I saw were all transient, and plenty of commits had green builds
going to wait for pr build, but does master run a different set of tests than branches?
t
the change seems reasonable to me. tests shouldn’t be different between master and branches unless the code is different between them
can you show me examples of where the acceptance tests passed with the code changes already there?
e
yeah, e.g. https://github.com/airbytehq/airbyte/runs/7526877454?check_suite_focus=true#step:14:877 has 
BasicAcceptanceTests > testDiscoverSourceSchema() PASSED
the diff after that commit is just regenerating normalization sample outputs + bumping versions/changelogs https://github.com/airbytehq/airbyte/pull/14362/files/0f0c93116093a16b39483a844a94566a9cc0af60..cf7dcef695a43b6e31171627705210f844200b36
which is spooky, and makes me wonder if my PR doesn’t actually work 100% of the time? but that seems kind of far-fetched
e
Thanks to both of you for jumping on this issue! I'm glad it wasn't my fault, and that a fix is forthcoming!
also, 👻 - spooky
e
also, am I supposed to be able to BasicAcceptanceTest locally? It fails on my laptop with 
io.airbyte.api.client.invoker.generated.ApiException: java.net.ConnectException
probably more relevant: 
Suppressed: java.lang.NullPointerException: Cannot invoke "io.airbyte.test.utils.AirbyteAcceptanceTestHarness.stopDbAndContainers()" because "io.airbyte.test.acceptance.BasicAcceptanceTests.testHarness" is null
t
yeah you are supposed to be able to run it locally. are you running it through the script? 
./tools/bin/acceptance_test.sh
you also might need to pull in more recent changes and rebuild the platform code with 
SUB_BUILD=PLATFORM ./gradlew build -x test
e
ah, I was just running from intellij. will try with the script
going to try just reverting my fix PR and rerunning, in the hopes that I can repro the failure. If that succeeds then I’m going to be very concerned
t
yeah, this is very odd. no idea why the one you linked passed and all the others failed. i’ll probably step away from this issue for now, but feel free to ping a platform OC for more help or an extra pair of eyes!
5 Views
Open in Slack
PreviousNext
Powered by