Thomas Tauber-Marshall has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/14538 )

Change subject: IMPALA-4400: aggregate runtime filters locally
......................................................................


Patch Set 22:

(17 comments)

http://gerrit.cloudera.org:8080/#/c/14538/22/be/src/runtime/coordinator.cc
File be/src/runtime/coordinator.cc:

http://gerrit.cloudera.org:8080/#/c/14538/22/be/src/runtime/coordinator.cc@371
PS22, Line 371:           // TODO: having a shared RuntimeFilterBank between 
all fragments on a backend
Is there a JIRA for this?


http://gerrit.cloudera.org:8080/#/c/14538/22/be/src/runtime/coordinator.cc@1216
PS22, Line 1216: :
Formatting here was correct before, here and several places below (probably a 
merge conflict issue)


http://gerrit.cloudera.org:8080/#/c/14538/22/be/src/runtime/coordinator.cc@1217
PS22, Line 1217:       if (!IsExecuting() && rpc_params.has_bloom_filter()) {
I think you resolved a merge conflict incorrectly here? (a previous version of 
Fang-Yu's patch had it this way, and you probably rebased onto that verison 
first, but we changed it)

We'll still want to do the 'return' if !IsExecuting() even if its a min-max 
filter.


http://gerrit.cloudera.org:8080/#/c/14538/22/be/src/runtime/runtime-filter-bank.h
File be/src/runtime/runtime-filter-bank.h:

http://gerrit.cloudera.org:8080/#/c/14538/22/be/src/runtime/runtime-filter-bank.h@167
PS22, Line 167: //
nit: formatting


http://gerrit.cloudera.org:8080/#/c/14538/22/be/src/runtime/runtime-filter-bank.h@168
PS22, Line 168: producer
is_producer


http://gerrit.cloudera.org:8080/#/c/14538/22/be/src/runtime/runtime-filter-bank.h@184
PS22, Line 184: locked separately
typo


http://gerrit.cloudera.org:8080/#/c/14538/22/be/src/runtime/runtime-filter-bank.h@237
PS22, Line 237:  AllocateScratchBloomFilter
AllocateScratch*Filter


http://gerrit.cloudera.org:8080/#/c/14538/22/be/src/runtime/runtime-filter-bank.cc
File be/src/runtime/runtime-filter-bank.cc:

http://gerrit.cloudera.org:8080/#/c/14538/22/be/src/runtime/runtime-filter-bank.cc@69
PS22, Line 69: 
formatting


http://gerrit.cloudera.org:8080/#/c/14538/22/be/src/runtime/runtime-state.h
File be/src/runtime/runtime-state.h:

http://gerrit.cloudera.org:8080/#/c/14538/22/be/src/runtime/runtime-state.h@63
PS22, Line 63: /// Shared state for Impala's runtime query execution. Used in 
two contexts:
Thanks for doing this!


http://gerrit.cloudera.org:8080/#/c/14538/22/be/src/util/bloom-filter.cc
File be/src/util/bloom-filter.cc:

http://gerrit.cloudera.org:8080/#/c/14538/22/be/src/util/bloom-filter.cc@250
PS22, Line 250: DCHECK
DCHECK_NE?


http://gerrit.cloudera.org:8080/#/c/14538/22/common/thrift/Planner.thrift
File common/thrift/Planner.thrift:

http://gerrit.cloudera.org:8080/#/c/14538/22/common/thrift/Planner.thrift@64
PS22, Line 64: 12
The numbering here is kinda weird. Maybe worth just fully renumbering? Not a 
big deal if you prefer to leave as is.


http://gerrit.cloudera.org:8080/#/c/14538/22/fe/src/main/java/org/apache/impala/planner/PlanFragment.java
File fe/src/main/java/org/apache/impala/planner/PlanFragment.java:

http://gerrit.cloudera.org:8080/#/c/14538/22/fe/src/main/java/org/apache/impala/planner/PlanFragment.java@110
PS22, Line 110: an instance of
I find this wording confusing, since perBackend... is shared across instances.


http://gerrit.cloudera.org:8080/#/c/14538/22/fe/src/main/java/org/apache/impala/planner/PlanFragment.java@121
PS22, Line 121: globla
typo


http://gerrit.cloudera.org:8080/#/c/14538/22/fe/src/main/java/org/apache/impala/planner/PlanFragment.java@122
PS22, Line 122: is accounted
              :   // are accounted
typo


http://gerrit.cloudera.org:8080/#/c/14538/22/fe/src/main/java/org/apache/impala/planner/PlanFragment.java@292
PS22, Line 292: :
formatting


http://gerrit.cloudera.org:8080/#/c/14538/22/fe/src/main/java/org/apache/impala/planner/PlanFragment.java@295
PS22, Line 295: :
formatting


http://gerrit.cloudera.org:8080/#/c/14538/22/fe/src/main/java/org/apache/impala/planner/PlanFragment.java@496
PS22, Line 496: perHostProfile
Not a big deal, but I think you could reduce the code duplication here a lot by 
just having perHostProfile and totalRuntimeFilterReservation get set based on 
whether mt_dop == 0, and then all of the builder.appends could just be here 
once.



--
To view, visit http://gerrit.cloudera.org:8080/14538
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: Iabeeab5eec869ff2197250ad41c1eb5551704acc
Gerrit-Change-Number: 14538
Gerrit-PatchSet: 22
Gerrit-Owner: Tim Armstrong <[email protected]>
Gerrit-Reviewer: Bikramjeet Vig <[email protected]>
Gerrit-Reviewer: Fang-Yu Rao <[email protected]>
Gerrit-Reviewer: Impala Public Jenkins <[email protected]>
Gerrit-Reviewer: Michael Ho <[email protected]>
Gerrit-Reviewer: Thomas Tauber-Marshall <[email protected]>
Gerrit-Reviewer: Tim Armstrong <[email protected]>
Gerrit-Comment-Date: Mon, 27 Jan 2020 20:11:41 +0000
Gerrit-HasComments: Yes

Reply via email to