davidzollo commented on code in PR #10925:
URL: https://github.com/apache/seatunnel/pull/10925#discussion_r3683944420
##########
seatunnel-e2e/seatunnel-connector-v2-e2e/connector-file-ftp-e2e/src/test/java/org/apache/seatunnel/e2e/connector/file/ftp/FtpFileIT.java:
##########
@@ -765,6 +884,38 @@ private void ensureReadJsonInputFile() throws IOException,
InterruptedException
Assertions.assertEquals(0, chmodResult.getExitCode(),
chmodResult.getStderr());
}
+ /**
+ * The multiple-table FTP JSON test reads two logical tables from two
table configs. Give each
+ * table its own physical source file so one reader never depends on a
file already consumed by
+ * the other table path.
+ */
+ private void ensureMultipleTableJsonInputFiles() throws IOException,
InterruptedException {
Review Comment:
Rechecked and documented in the helper Javadoc:
`ensureMultipleTableJsonInputFiles` owns `tmp/seatunnel/read/json` for the
suite and rebuilds that tree before the multiple-table run. The single-table
helper also resets the same directory before its own run, so the shared
directory ownership is intentional and now explicit.
##########
seatunnel-e2e/seatunnel-connector-v2-e2e/connector-rocketmq-e2e/src/test/java/org/apache/seatunnel/e2e/connector/rocketmq/RocketMqIT.java:
##########
@@ -806,6 +837,15 @@ private long getTopicMaxOffset(String topicName) {
}
}
+ /**
+ * Returns a collision-resistant suffix for broker resources shared by
template invocations.
+ *
+ * @return UUID text without separators, suitable for topic and
consumer-group names
+ */
+ private String uniqueTestSuffix() {
Review Comment:
Confirmed. `uniqueTestSuffix()` uses UUID-backed names and RocketMQ accepts
independent topic creation requests; these resources are per test invocation,
so synchronization around UUID generation is not needed. The helper Javadoc now
documents that assumption.
##########
seatunnel-e2e/seatunnel-connector-v2-e2e/connector-cdc-mysql-e2e/src/test/java/org/apache/seatunnel/connectors/seatunnel/cdc/mysql/MysqlCDCWithFlinkSchemaChangeIT.java:
##########
@@ -68,7 +68,7 @@ public class MysqlCDCWithFlinkSchemaChangeIT extends
TestSuiteBase implements Te
* Flink schema evolution can restart after XA recover/rollback on loaded
CI runners, so these
* assertions need enough time for the job to recover and replay the
schema-change event.
*/
- private static final long SCHEMA_EVOLUTION_ASSERT_TIMEOUT_MILLIS =
300_000L;
+ private static final long SCHEMA_EVOLUTION_ASSERT_TIMEOUT_MILLIS =
600_000L;
Review Comment:
Rechecked this timeout. I kept the 600s bound as a failure deadline rather
than expected recovery duration, because this E2E covers schema-change replay
after Flink XA recover/rollback on loaded CI runners. The assertions still poll
real schema/data equality and fail if the propagation stalls. I did tighten the
timestamp drift guard separately from 60s to 20s in `c21a0d39c1` so
minute-level lag is no longer hidden by the timestamp assertion.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
