[
https://issues.apache.org/jira/browse/FLINK-33727?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17792534#comment-17792534
]
Jim Hughes commented on FLINK-33727:
------------------------------------
{quote}..., however tests are continuing passing. Is it expected?
{quote}
"Yes", but the reason is a little confusing.
The RestoreTest framework has two methods: `generateTestSetupFiles` and
`testRestore`.
Presently, the method `generateTestSetupFiles` is disabled and only run by test
developers before a PR is submitted. This method takes the SQL, gets and saves
a compiled plan, and runs through the beforeRestore data making some
comparisons, and finally stopping the job and taking a savepoint.
The second method uses the compiled plan and the savepoint.
Since you are only running the second method, changing the SQL is irrelevant
and not tested (unless you manually run `generateTestSetupFiles`).
{quote}what is wrong with current PR for this JIRA?
{quote}
CI failing is showing that the RestoreTestBase has some limitations/assumptions
around state which we need to address. The current PR fixes CI, but does not
address those, rather it works around them. I'd prefer that we fix the
limitations rather than work around them. That's why I'm suggesting to disable
the RestoreTests as a whole until Monday when Dawid and Timo can weigh in.
> JoinRestoreTest is failing on AZP
> ---------------------------------
>
> Key: FLINK-33727
> URL: https://issues.apache.org/jira/browse/FLINK-33727
> Project: Flink
> Issue Type: Bug
> Components: Table SQL / Planner
> Affects Versions: 1.19.0
> Reporter: Sergey Nuyanzin
> Priority: Critical
> Labels: pull-request-available, test-stability
>
> Since {{JoinRestoreTest}} was introduced in FLINK-33470 it seems to be a
> reason
> {noformat}
> Dec 02 04:42:26 04:42:26.408 [ERROR] Failures:
> Dec 02 04:42:26 04:42:26.408 [ERROR]
> JoinRestoreTest>RestoreTestBase.testRestore:283
> Dec 02 04:42:26 Expecting actual:
> Dec 02 04:42:26 ["+I[9, carol, apple, 9000]",
> Dec 02 04:42:26 "+I[8, bill, banana, 8000]",
> Dec 02 04:42:26 "+I[6, jerry, pen, 6000]"]
> Dec 02 04:42:26 to contain exactly in any order:
> Dec 02 04:42:26 ["+I[Adam, null]",
> Dec 02 04:42:26 "+I[Baker, Research]",
> Dec 02 04:42:26 "+I[Charlie, Human Resources]",
> Dec 02 04:42:26 "+I[Charlie, HR]",
> Dec 02 04:42:26 "+I[Don, Sales]",
> Dec 02 04:42:26 "+I[Victor, null]",
> Dec 02 04:42:26 "+I[Helena, Engineering]",
> Dec 02 04:42:26 "+I[Juliet, Engineering]",
> Dec 02 04:42:26 "+I[Ivana, Research]",
> Dec 02 04:42:26 "+I[Charlie, People Operations]"]
> {noformat}
> examples of failures
> https://dev.azure.com/apache-flink/apache-flink/_build/results?buildId=55120&view=logs&j=0c940707-2659-5648-cbe6-a1ad63045f0a&t=075c2716-8010-5565-fe08-3c4bb45824a4&l=12099
> https://dev.azure.com/apache-flink/apache-flink/_build/results?buildId=55129&view=logs&j=a9db68b9-a7e0-54b6-0f98-010e0aff39e2&t=cdd32e0b-6047-565b-c58f-14054472f1be&l=11786
> https://dev.azure.com/apache-flink/apache-flink/_build/results?buildId=55136&view=logs&j=0c940707-2659-5648-cbe6-a1ad63045f0a&t=075c2716-8010-5565-fe08-3c4bb45824a4&l=12099
> https://dev.azure.com/apache-flink/apache-flink/_build/results?buildId=55137&view=logs&j=0c940707-2659-5648-cbe6-a1ad63045f0a&t=075c2716-8010-5565-fe08-3c4bb45824a4&l=11779
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
