orangeCatDeveloper commented on issue #3197: URL: https://github.com/apache/hertzbeat/issues/3197#issuecomment-5861052855
Root cause, still present on master: the IPMI client opens a blocking `DatagramChannel` and calls `channel.receive()` without any timeout. If a UDP response is lost (UDP gives no delivery guarantee), the collection thread blocks until the task-level timeout, which matches the hang described here. https://github.com/apache/hertzbeat/blob/8104d7057c72f9c18dea31d32e076750d2c34020/hertzbeat-collector/hertzbeat-collector-basic/src/main/java/org/apache/hertzbeat/collector/collect/ipmi2/client/UdpConnection.java#L64-L66 A fix would be to receive with a timeout (e.g. a `Selector` or a non-blocking channel with a deadline) and retry or fail the request. -- 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]
