bgaborg commented on issue #1079: HADOOP-16380: test to show that it is the 
root directory where the "tombstone problem" can be replicated
URL: https://github.com/apache/hadoop/pull/1079#issuecomment-511409184
 
 
   > OK. Is there any other issue which this problem creates.
   > I'm worried about whether if there's a file `/test/dir1/file.avro` under, 
say, `/test`, and there's a tombstone for `/test`, will the file get found in a 
listing, _and is that behaviour different in the root path from elsewhere?_. I 
know a listFiles() is likely to find it, I'd expect a treewalk with 
listStatus() will miss it just because the topmost directory listing won't find 
the /test dir to scan
   
   If there's a tombstone for `/test` it won't be visible, but we should expect 
that behaviour.
   This is why we have tombstone expiry. So eventually we don't have this 
problem - if /test/ has bee created by an OOB operation (I guess that's why we 
have a tombstone and still have the directory structure behind it) then the 
metadata and tombstone expiry will be eventually solved.
   
   The problem starts to surface if we have a `/test1/` tombstoned, and we also 
have `/test9`, or even `/test2/dir1/file.avro`. I will write a test for that.

----------------------------------------------------------------
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.
 
For queries about this service, please contact Infrastructure at:
[email protected]


With regards,
Apache Git Services

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to