Philip Zeyliger has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/12437 )

Change subject: IMPALA-8185: Abstract out real/mock file system operations
......................................................................


Patch Set 4:

(2 comments)

Just kibitzing: would be good to move the comments to the interface so people 
can tell what's going on?

http://gerrit.cloudera.org:8080/#/c/12437/4/fe/src/main/java/org/apache/impala/analysis/FileSystemFacade.java
File fe/src/main/java/org/apache/impala/analysis/FileSystemFacade.java:

http://gerrit.cloudera.org:8080/#/c/12437/4/fe/src/main/java/org/apache/impala/analysis/FileSystemFacade.java@64
PS4, Line 64:   ScanAllocation allocateScans(FeFsTable table, TScanRangeSpec 
scanRangeSpecs,
Where do we document what this does?


http://gerrit.cloudera.org:8080/#/c/12437/4/fe/src/main/java/org/apache/impala/analysis/HdfsFileSystemFacade.java
File fe/src/main/java/org/apache/impala/analysis/HdfsFileSystemFacade.java:

http://gerrit.cloudera.org:8080/#/c/12437/4/fe/src/main/java/org/apache/impala/analysis/HdfsFileSystemFacade.java@102
PS4, Line 102:   public ScanAllocation allocateScans(FeFsTable table, 
TScanRangeSpec scanRangeSpecs,
Do you know what this is used for?

It's kind of different from the C++ version of the same 
(be/src/scheduling/scheduler.h). I don't think it matters, but it's weirdo.

I'd basically be surprised of computeNumNodes() isn't just:

max(1, min(numExecutors, numBlocks))

i.e., always at least 1 and at most numExecutors, but determined by the number 
of blocks.



--
To view, visit http://gerrit.cloudera.org:8080/12437
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I1a385923b64c9fb59cc6e700ee7ee14919398e6d
Gerrit-Change-Number: 12437
Gerrit-PatchSet: 4
Gerrit-Owner: Paul Rogers <[email protected]>
Gerrit-Reviewer: Bharath Vissapragada <[email protected]>
Gerrit-Reviewer: Impala Public Jenkins <[email protected]>
Gerrit-Reviewer: Paul Rogers <[email protected]>
Gerrit-Reviewer: Philip Zeyliger <[email protected]>
Gerrit-Comment-Date: Wed, 20 Feb 2019 22:43:04 +0000
Gerrit-HasComments: Yes

Reply via email to