Joe McDonnell has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/18616 )

Change subject: IMPALA-10545: Higher data_cache_write_concurrency for SSDs
......................................................................


Patch Set 4:

(2 comments)

This is looking good to me, only minor nits.

http://gerrit.cloudera.org:8080/#/c/18616/4/be/src/runtime/io/data-cache.cc
File be/src/runtime/io/data-cache.cc:

http://gerrit.cloudera.org:8080/#/c/18616/4/be/src/runtime/io/data-cache.cc@398
PS4, Line 398:     write_concurrency = 8;
Nit: Let's pull the 8 constant out as a global constant with its own variable. 
(Similar to THREADS_PER_SOLID_STATE_DISK in disk-io-mgr.cc.) Same with the 
above 1 constant for spinning disks.


http://gerrit.cloudera.org:8080/#/c/18616/4/be/src/util/disk-info.cc
File be/src/util/disk-info.cc:

http://gerrit.cloudera.org:8080/#/c/18616/4/be/src/util/disk-info.cc@96
PS4, Line 96:       FileSystemUtil::PathExists(sys + "/block/" + dev_name, 
&found_device);
Nit: Some C++ string manipulations create copies and can be wasteful. This is 
the reason you'll see a lot of use of Substitute() from gutil/strings and other 
forms of avoiding basic "+"'s (like stringstreams).

This isn't performance sensitive code, but our preference is to use those 
constructs. This is somewhat a style thing so that we aren't debating it 
frequently.



--
To view, visit http://gerrit.cloudera.org:8080/18616
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I60761faa2710f4795f1f3eaf66da866b5553f609
Gerrit-Change-Number: 18616
Gerrit-PatchSet: 4
Gerrit-Owner: Michael Smith <[email protected]>
Gerrit-Reviewer: Impala Public Jenkins <[email protected]>
Gerrit-Reviewer: Joe McDonnell <[email protected]>
Gerrit-Reviewer: Michael Smith <[email protected]>
Gerrit-Comment-Date: Fri, 24 Jun 2022 17:36:31 +0000
Gerrit-HasComments: Yes

Reply via email to