[GitHub] [kafka] showuon commented on pull request #12238: KIP-835: MetadataMaxIdleIntervalMs should be much bigger than brokerHeartbeatIntervalMs

showuon commented on PR #12238: URL: https://github.com/apache/kafka/pull/12238#issuecomment-1145703278 > should we allow the broker to be behind by A records (or time) to unfence the broker? I've similar solution came out yesterday, but then, I don't think this is a good solution

[GitHub] [kafka] showuon commented on pull request #12238: KIP-835: MetadataMaxIdleIntervalMs should be much bigger than brokerHeartbeatIntervalMs

showuon commented on PR #12238: URL: https://github.com/apache/kafka/pull/12238#issuecomment-1144876598 [KAFKA-13955](https://issues.apache.org/jira/browse/KAFKA-13955) is created for this failing tests, in case duplicated works by other people. (Because it failed many tests in PR build

[GitHub] [kafka] showuon commented on pull request #12238: KIP-835: MetadataMaxIdleIntervalMs should be much bigger than brokerHeartbeatIntervalMs

showuon commented on PR #12238: URL: https://github.com/apache/kafka/pull/12238#issuecomment-1144595607 Interesting bug. If we decided to increase the default `metadata.max.idle.interval.ms` value, we should note in the KIP discussion thread. However, I'm wondering if increasing the