[GitHub] [flink] buptljy commented on issue #11104: [FLINK-16051] Subtask ID in Overview-Subtasks should start from 1

2020-02-17 Thread GitBox
buptljy commented on issue #11104: [FLINK-16051] Subtask ID in 
Overview-Subtasks should start from 1
URL: https://github.com/apache/flink/pull/11104#issuecomment-587245879
 
 
   > I remember a similar discussion in the past (can't find it right now), 
with the conclusion being that the only proper solution would be to be 
consistent across the entire project.
   > However, at the time we did not see a good way to implement this, and I 
still don't see it now.
   > 
   > As a result we sticked with the existing inconsistency. Without a clear 
goal of what the final solution should be there is no point in taking 
intermediate steps, as in the worst case they go in the wrong direction.
   > 
   > Sure, starting the index with 1 is "nicer", but implies that in every log 
statement we have to remember to increment the subtask index.
   > At the same time, changing the subtask index for metrics/logs may create 
headaches for or even break existing setups.
   
   Makes sense to me. We can keep the JIRA issue for further discussion. 
Closing this now...


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:
us...@infra.apache.org


With regards,
Apache Git Services


[GitHub] [flink] buptljy commented on issue #11104: [FLINK-16051] Subtask ID in Overview-Subtasks should start from 1

2020-02-17 Thread GitBox
buptljy commented on issue #11104: [FLINK-16051] Subtask ID in 
Overview-Subtasks should start from 1
URL: https://github.com/apache/flink/pull/11104#issuecomment-586890133
 
 
   > Why did you opt for incrementing this subtask id? Now both the checkpoint 
and subtask id are inconsistent with the entire Flink runtime, including 
logging and metrics.
   
   @zentol I've struggled for a little bit here because:
   1. ID in subtasks starts from 0, while ID in checkpoint / watermark / 
backpressure starts from 1.
   2. Logging with #ExecutionVertex#subtaskIndex starts from 0, while Logging 
with #ExecutionVertex#taskNameWithSubtask starts from 1.
   
   The reason I choose incrementing ID in subtasks is that it may look more 
reasonable and intuitive from user's perspective(maybe not from developer's 
perspective). I didn't think of the metrics before but it should be included.


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:
us...@infra.apache.org


With regards,
Apache Git Services