hubcio opened a new pull request, #4407: URL: https://github.com/apache/iggy/pull/4407
The cluster tests inject scripted faults at chosen points, for example one SIGKILL of a backup or of the primary. No test kept producers, consumer groups and segment expiry running while nodes died one after another. The new test SIGKILLs one random node of five every interval and restarts it with its data. Message ids carry the producer and the sequence, so the test can account for every acknowledged message without trusting payloads. At the end, it compares the nodes over HTTP and, after a clean stop, their segment files byte for byte. A second topic without expiry keeps the full history for these compares, because expiry leaves only active segments. The module doc lists the server gaps that a run can hit or that limit its checks. It is opt-in (#[ignore]) because it runs for minutes. A nextest override runs it alone, so other tests do not skew the kill timing, and allows up to one hour for long runs. Its knobs use the IGGY_TEST_ prefix of the other harness variables. The harness forwards every IGGY_* variable to the servers, and a debug server refuses to boot on an unknown one, so the boot check now accepts that prefix instead of three listed names. -- 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]
