The GitHub Actions job "Flink CI (beta)" on flink.git/master has failed.
Run started by GitHub user MartijnVisser (triggered by MartijnVisser).

Head commit for run:
e91a0ce09e5b944853bca713be2d445971112cb2 / Martijn Visser 
<[email protected]>
[FLINK-21995][yarn-tests] Whitelist netty-wrapped Pekko connection-refused WARN

The checkForProhibitedLogContents check in YARNSessionFIFOSecuredITCase
(and the other YARN session ITCases) flags any "Exception" substring in
the TaskManager/JobManager logs. A benign WARN is emitted by Pekko during
association teardown when a remote system is gated and the reconnect is
refused. That line was already whitelisted, but the whitelist regex
assumed the cause was a bare "java.net.ConnectException: Connection
refused: <host>". Current Pekko/Netty wraps the cause in
"org.apache.flink.shaded.netty4.io.netty.channel.AbstractChannel$AnnotatedConnectException",
so the line no longer matches and the transient WARN trips the check,
producing flaky failures such as testDetachedMode (see FLINK-26514).

Extend the "Caused by:" portion of the existing pattern to accept both
the old direct ConnectException form and the new Netty-wrapped
AnnotatedConnectException form. The cause class is pinned to those two
known types rather than matching any class that mentions "Connection
refused", so an unrelated exception cannot slip through the whitelist.
The rest of the pattern stays specific to the gated re-association WARN.

This removes one source of false positives in the prohibited-log-contents
check; it does not address every flaky cause tracked under FLINK-21995.

YarnTestBaseTest gains a positive probe for the wrapped form and a
negative probe asserting that a gated re-association WARN with a
non-connection cause (still containing "Connection refused" text) is not
whitelisted, guarding the discriminating property of the pattern.

Generated-by: Claude Code (Opus 4.8)

Report URL: https://github.com/apache/flink/actions/runs/28426126517

With regards,
GitHub Actions via GitBox

Reply via email to