You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@iceberg.apache.org by GitBox <gi...@apache.org> on 2020/10/30 09:29:31 UTC

[GitHub] [iceberg] wg1026688210 commented on pull request #1687: make load hive table more robust

wg1026688210 commented on pull request #1687:
URL: https://github.com/apache/iceberg/pull/1687#issuecomment-719444845


   hi~ @RussellSpitzer  thanks for your concerns . 
   I do some improvments on trying previous_location use quite a long time on a new commit . 
   The parsing metadata retry will stop when  it throws NotFoundException.  
   
   about the use case for this PR 
   I found the metadata_location file was deleted when commit fail as I mentioned in #1688 ,
   but the metadata_location property of the hive table  still became  the metadata_location which is deleted.
   Flink job fail to restart from checkpoint all the time due to the deleted file      
   
   https://github.com/apache/iceberg/blob/4ba48be69963fded08bcd9d0a28e14202191585d/hive-metastore/src/main/java/org/apache/iceberg/hive/HiveTableOperations.java#L201-L204
   
   I want to  make a improvment that  trying  the previous_metadata_location  only  if  geting the metadata_location throws NotFoundException, and if the metadata_location is no found , I will set the previous_metadata_location value to it in case of both metadata_location are  not found. 


----------------------------------------------------------------
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:
users@infra.apache.org



---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For additional commands, e-mail: issues-help@iceberg.apache.org