steveloughran commented on PR #7164: URL: https://github.com/apache/hadoop/pull/7164#issuecomment-2486111016
regarding ITestUploadRecovery we try to regenerate the failure condition "whole block uploaded then server returns 500" that is: it is testing that the SDK will recover from such a failure by resubmitting the block again. If this doesn't work then will CSE code be able to recover from a failure in a part upload? the implication is "no" In which case, although this is a rare occurrence, it does surface in the wild sporadically and it seems like this going to result in a failure with this message. Ideally, I would like this recovery to be handled. Is this possible? what do other clients do here? do you have a special "unreliable test endpoint" to generate these failures for more accurate testing? If it can't be recovered from and the error message can be expected * test should actually expect upload failure and the error if encryption is enabled. This will highlight if/when there's a change. * troubleshooting docs need something on the error and the meaning. -- 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]
