velctor commented on pull request #3276: URL: https://github.com/apache/iceberg/pull/3276#issuecomment-948183868
> @velctor, why do you think that the stream was closed by the finalizer and not by a normal call to `close`? In order to determine that the file stream closed by the `Finalizer thread` is the same as the stream in the error log, I add informations to the log and recompile the relevant jar package. The log is as follows: ``` 2021-*-*T14:47:43.211+0800 DEBUG 20210927_064723_00001_ajsi3.1.0-13-103 org.apache.hadoop.fs.FileSystem Bypassing cache to create filesystem ***00000-1-b5ccab76-0db2-4f2e-887f-f70b9833cb91-00227.parquet 2021-*-*T14:47:43.381+0800 DEBUG Finalizer org.apache.hadoop.hdfs.DFSClient close DFSInputStream ***00000-1-b5ccab76-0db2-4f2e-887f-f70b9833cb91-00227.parquet thread name:Finalizer client:DFSClient_NONMAPREDUCE_373867554_103 2021-*-*T14:47:43.382+0800 DEBUG 20210927_064723_00001_ajsi3.1.0-13-103 org.apache.hadoop.hdfs.DFSClient DFSInputStream has been closed already ***00000-1-b5ccab76-0db2-4f2e-887f-f70b9833cb91-00227.parquet thread name:20210927_064723_00001_ajsi3.1.0-13-103 client:DFSClient_NONMAPREDUCE_373867554_103 org.apache.iceberg.exceptions.RuntimeIOException: Failed to open Parquet file: ***00000-1-b5ccab76-0db2-4f2e-887f-f70b9833cb91-00227.parquet, thread name: 20210927_064723_00001_ajsi3.1.0-13-103 ``` In addition, when I applied these changes to iceberg, no related errors were reported after testing. -- 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]
