dongjoon-hyun commented on PR #43743:
URL: https://github.com/apache/spark/pull/43743#issuecomment-1809845273

   To @mridulm . The above was the exact answer for your following question. 
What I meant is that Apache Spark already provide more patterns including `the 
prefix customazable`.
   > Why not continue to use the existing scheme (iirc prefix + time: I am afk, 
so can't confirm), but make the prefix customizable instead ?
   
   For the following question, that is the exact design goal of this PR where a 
upper service system uses `Apache Spark Standalone cluster as an execution 
subsystem`. The upper service layer already provides many features like the 
unique appName and security layers. So, there is no such things like `hard to 
debug` or `minimizing the issue`.
   > why we are allowing explicitly setting an application id, and introducing 
hard to debug modes; instead of relying on minimizing the issue.
   
   In addition, I already answered with the examples. Let me reiterate that. 
You can generate a SHS link like `https://shs/history/app-202311140014-0204/` 
with `appId` while you cannot with `appName`. `appId` is used as a foreign key 
across Apache Spark systems (Storage like S3 event log and Spark driver logs, 
SHS link, Spark Master link). It opens a wide range of opportunity to us.
   
   - https://github.com/apache/spark/pull/43743#issuecomment-1806033968
   
   Lastly, this configuration is introduced as `Experimental` as you see 
because we are already aware of the risk. We don't recommend this configuration 
for novice users. This is more like automated subsystems in the industrial 
production usages.
   
   
https://github.com/apache/spark/blob/2cac768fabaa7cad40390b2205dd9c5000011e4c/core/src/main/scala/org/apache/spark/internal/config/package.scala#L1852


-- 
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.

To unsubscribe, e-mail: [email protected]

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


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

Reply via email to