Martijn Visser created FLINK-41248:
--------------------------------------
Summary:
UnalignedCheckpointRescaleITCase.shouldRescaleUnalignedCheckpoint fails with
OutOfMemoryError: Direct buffer memory in the Netty server
Key: FLINK-41248
URL: https://issues.apache.org/jira/browse/FLINK-41248
Project: Flink
Issue Type: Bug
Components: Runtime / Checkpointing
Reporter: Martijn Visser
UnalignedCheckpointRescaleITCase.shouldRescaleUnalignedCheckpoint failed on the
release-1.20 nightly of 2026-10-24 (Java 17, module tests):
https://github.com/apache/flink/actions/runs/40052503223/job/120756786222
{code}
[ERROR]
org.apache.flink.test.checkpointing.UnalignedCheckpointRescaleITCase.shouldRescaleUnalignedCheckpoint[upscale
multi_input from 1 to 2, sourceSleepMs = 0, buffersPerChannel = 0] -- Time
elapsed: 0.679 s <<< ERROR!
org.apache.flink.runtime.client.JobExecutionException: Job execution failed.
Caused by: org.apache.flink.runtime.JobException: Recovery is suppressed by
FixedDelayRestartBackoffTimeStrategy(maxNumberRestartAttempts=0,
backoffTimeMS=100)
Caused by:
org.apache.flink.runtime.io.network.netty.exception.RemoteTransportException:
Connection unexpectedly closed by remote task manager
'localhost/127.0.0.1:34639'. This might indicate that the remote task manager
was lost.
{code}
The fork log in the job's artifact shows the Netty server of that task manager
running out of direct memory just before, in the first run before the rescale:
{code}
03:16:14,819 [Flink Netty Server (0) Thread 0] ERROR
org.apache.flink.runtime.io.network.netty.PartitionRequestQueue [] -
Encountered error while consuming partitions (connection to /127.0.0.1:56798)
Caused by: java.lang.OutOfMemoryError: Cannot reserve 16384 bytes of direct
buffer memory (allocated: 1610596643, limit: 1610612736)
{code}
FLINK-30644 has the same Netty direct memory OOM in
ChangelogCompatibilityITCase on release-1.20.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
