[HDFS-13133] Ozone: OzoneFileSystem: Calling delete with non-existing path shouldn't be logged on ERROR level

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: HDFS-7240
• Fix Version/s: HDFS-7240
• Component/s: ozone
• Hadoop Flags: Reviewed
• Assignee: Marton Elek
• Reporter: Marton Elek
• Created: 11/Feb/18 16:36
• Updated: 26/Apr/18 21:56
• Resolved: 12/Feb/18 08:33

Description

During the test of OzoneFileSystem with spark I noticed ERROR messages multiple times:

Something like this:

2018-02-11 15:54:54 ERROR OzoneFileSystem:409 - Couldn't delete o3://bucket1.test/user/hadoop/.sparkStaging/application_1518349702045_0008 - does not exist

I checked the other implemetations, and they use DEBUG level. I think it's expected that the path sometimes points to a non-existing dir/file.

To be consistent with the other implemetation I propose to lower the log level to debug.

Examples from other file systems:

S3AFileSystem:

} catch (FileNotFoundException e) {
      LOG.debug("Couldn't delete {} - does not exist", f);
      instrumentation.errorIgnored();
      return false;
    } catch (AmazonClientException e) {
      throw translateException("delete", f, e);
    }

Alyun:

   try {
      return innerDelete(getFileStatus(path), recursive);
    } catch (FileNotFoundException e) {
      LOG.debug("Couldn't delete {} - does not exist", path);
      return false;
    }

SFTP:

   } catch (FileNotFoundException e) {
      // file not found, no need to delete, return true
      return false;
    }

SwiftNativeFileSystem:

    try {
      return store.delete(path, recursive);
    } catch (FileNotFoundException e) {
      //base path was not found.
      return false;
    }