[GitHub] [hadoop-ozone] elek commented on pull request #1028: HDDS-3735. Improve SCM performance with 3.7% by remove unnecessary lock and unlock

[GitBox <gi...@apache.org>]

elek commented on pull request #1028:
URL: https://github.com/apache/hadoop-ozone/pull/1028#issuecomment-647466786


   Actually I am not sure about the other part.
   
   Using read / write locks in the `ContainerStateMap` provides a consistent view. Read calls can see any containers only if all steps from the write method are done (all the indexes are updated).
   
   The patch seems to introduce a consistency problem. If you don't have a read lock the write may not be finished when you read. The  `containerMap` itself is ok, but if you do any operation with any other indexes from `ContainerStateMap` you might see a different view.
   
   Did I miss something? @runzhiwang 


----------------------------------------------------------------
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: ozone-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: ozone-issues-help@hadoop.apache.org