kbendick commented on issue #2991: URL: https://github.com/apache/iceberg/issues/2991#issuecomment-901575352
Also, YARN and Spark both make use of `java.io.tmpdir` by default. You might consider setting `spark.local.dir` or even configuring YARN to use a different temporary directory: https://stackoverflow.com/questions/49317521/how-to-change-java-io-tmpdir-for-spark-job-running-on-yarn There is also a possibility that the directory is full based on your EBS mounting strategy or instance type, could you check if that is the case? Also, can you verify the permissions of the directory? You may want to reach out to your TAMs as well (if you have support) to see if they can help you troubleshoot the EMR issue, as this has only been brought up one other time to my knowledge - in that case, it was a disk full issue AFAIK. Yours might be different though! -- 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]
