Andrew Wong has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/11605 )

Change subject: [compaction] Small improvements to compaction policy tests
......................................................................


Patch Set 2:

(4 comments)

http://gerrit.cloudera.org:8080/#/c/11605/2/src/kudu/tablet/compaction_policy-test.cc
File src/kudu/tablet/compaction_policy-test.cc:

http://gerrit.cloudera.org:8080/#/c/11605/2/src/kudu/tablet/compaction_policy-test.cc@117
PS2, Line 117: // Similar to the above, but with some small overlap between 
adjacent
             : // rowsets.
In compaction-policy parlance, this means is testing compaction on a "short" 
(keyspace height is small) tablet, right? Whereas TestSignificantOverlap is 
"tall"?


http://gerrit.cloudera.org:8080/#/c/11605/2/src/kudu/tablet/compaction_policy-test.cc@131
PS2, Line 131: %09d", i * 10000),
             :         StringPrintf("%09d
Was this change important?


http://gerrit.cloudera.org:8080/#/c/11605/2/src/kudu/tablet/compaction_policy-test.cc@178
PS2, Line 178: the
nit: remove


http://gerrit.cloudera.org:8080/#/c/11605/2/src/kudu/tablet/compaction_policy-test.cc@185
PS2, Line 185: // compacting {A, B, C} results in the same quality score as {A, 
B}, 0.67, but
             : // uses more budget. By penalizing the wider solution {A, B, C}, 
we ensure we
             : // don't waste I/O.
Not really related to this change, but thought I'd ask. Conceptually, I get 
that we're reducing the summed width an equal amount in both cases, but in one 
case we're using 2 "units" of IO and in the other we're using 3, but I haven't 
found a doc or comment explaining what 'quality' is in this case. Do they exist 
somewhere? Is there a formula for 'quality'?



--
To view, visit http://gerrit.cloudera.org:8080/11605
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I8f7a0e2c69f301d7fc7ca2fac657569ea240d4e3
Gerrit-Change-Number: 11605
Gerrit-PatchSet: 2
Gerrit-Owner: Will Berkeley <[email protected]>
Gerrit-Reviewer: Adar Dembo <[email protected]>
Gerrit-Reviewer: Andrew Wong <[email protected]>
Gerrit-Reviewer: Attila Bukor <[email protected]>
Gerrit-Reviewer: Kudu Jenkins
Gerrit-Reviewer: Mike Percy <[email protected]>
Gerrit-Reviewer: Todd Lipcon <[email protected]>
Gerrit-Comment-Date: Tue, 09 Oct 2018 07:29:49 +0000
Gerrit-HasComments: Yes

Reply via email to