You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@impala.apache.org by "Zoltan Borok-Nagy (Code Review)" <ge...@cloudera.org> on 2019/01/21 14:32:39 UTC

[Impala-ASF-CR] IMPALA-8090: race when reusing ScanRange in test

Zoltan Borok-Nagy has posted comments on this change. ( http://gerrit.cloudera.org:8080/12238 )

Change subject: IMPALA-8090: race when reusing ScanRange in test
......................................................................


Patch Set 2:

(2 comments)

When I have some time I'll come back to scan ranges in the context of IMPALA-7556, because I always find some peculiar and shady stuff going on here :)

http://gerrit.cloudera.org:8080/#/c/12238/2/be/src/runtime/io/scan-range.cc
File be/src/runtime/io/scan-range.cc:

http://gerrit.cloudera.org:8080/#/c/12238/2/be/src/runtime/io/scan-range.cc@211
PS2, Line 211: Open
side note: In HdfsFileReader::Open() there is an unnecessary hdfsSeek() that always seeks to the beginning of the ScanRange. Then we also perform an hdfsSeek()/hdfsPread() in ReadFromPos().

I'm not sure about the overhead of it, but if some parts of the system do some speculative pre-fetching based on hdfsSeek(), it can hurt.

Though it only happens when file handle caching is used.


http://gerrit.cloudera.org:8080/#/c/12238/2/be/src/runtime/io/scan-range.cc@240
PS2, Line 240:     file_reader_->Close();
This bug existed earlier as well, but FileReader::Close() is not idempotent, and it is also called in CancelInternal().



-- 
To view, visit http://gerrit.cloudera.org:8080/12238
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I3122e5b2efea60ffe82d780930301d5be108876b
Gerrit-Change-Number: 12238
Gerrit-PatchSet: 2
Gerrit-Owner: Tim Armstrong <ta...@cloudera.com>
Gerrit-Reviewer: Zoltan Borok-Nagy <bo...@cloudera.com>
Gerrit-Comment-Date: Mon, 21 Jan 2019 14:32:39 +0000
Gerrit-HasComments: Yes