anmolanmol1234 commented on code in PR #8830:
URL: https://github.com/apache/hadoop/pull/8830#discussion_r4523868287
##########
hadoop-tools/hadoop-azure/src/test/java/org/apache/hadoop/fs/azurebfs/services/ITestAbfsInputStreamSmallFileReads.java:
##########
@@ -97,11 +99,18 @@ private void validateNumBackendCalls(final boolean
readSmallFilesCompletely,
long requestsMadeAfterTest = metricMap
.get(CONNECTIONS_MADE.getStatName());
+ long expected;
if (readSmallFilesCompletely) {
- assertEquals(1, requestsMadeAfterTest - requestsMadeBeforeTest);
+ expected = 1;
} else {
- assertEquals(3, requestsMadeAfterTest - requestsMadeBeforeTest);
+ expected = 3;
}
+ Assertions.assertThat(requestsMadeAfterTest - requestsMadeBeforeTest)
+ .describedAs(
+ "Number of calls made will be equal to expected if "
+ + "data layout is already present else greater than 1 if
call to "
+ + "fetch data layout is made")
+ .isIn(Arrays.asList(expected, expected + 1));
Review Comment:
This now accepts either `expected` or `expected + 1` backend calls, so an
unrelated extra request would pass unnoticed. Can we keep the original exact
count for the non-layout calls and assert the GetBlobLayout call count
separately, or disable data locality in this test? That way a real
extra-request regression is still caught.
--
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
