replication_monitor/test: add diagnostic on assertion failure #20818
Conversation
Additionall reduces sleep durations to make the test faster.
|
Travis reported an issue here https://redpandadata.slack.com/archives/C05G315PGDR/p1719889095956289 where the test is failing to replicate even before the test does anything substantial, there is nothing in the log that indicates why (no drop of leadership etc). I looped through the test 1000s of times locally, couldn't reproduce, just adding additional error message when the test fails incase it reproduces again in CI 🤞 made it a bit quicker while I was there. |
| ASSERT_FALSE_CORO(all.available()); | ||
|
|
||
| for (auto i = 0; i < num_waiters(); i++) { | ||
| auto result = co_await raft->replicate( | ||
| make_batches({{"k", "v"}}), | ||
| replicate_options{raft::consistency_level::quorum_ack}); | ||
| ASSERT_TRUE_CORO(result.has_value()); | ||
| ASSERT_TRUE_CORO(result.has_value()) << result.error(); |
There was a problem hiding this comment.
I don't see how the evaluation of the RHS of << would be optionally evaluated. Is it? If not, this looks like undefined behavior to me, calling .error() on a result type that has a value.
There was a problem hiding this comment.
This is the output from preprocessor.
switch (0)
case 0:
default:
if (const ::testing::AssertionResult gtest_ar_ =
::testing::AssertionResult(result.has_value()))
;
else
co_return ::testing::internal::AssertHelper(
::testing::TestPartResult::kFatalFailure,
"/work/redpanda/src/v/raft/tests/replication_monitor_tests.cc", 69,
::testing::internal::GetBoolAssertionFailureMessage(
gtest_ar_, "result.has_value()", "false", "true")
.c_str()) = ::testing::Message() << result.error()
Additionally reduces sleep durations to make the test faster.
Backports Required
Release Notes