[Impala-ASF-CR] IMPALA-8999: make union scheduling work with mt dop

2019-10-23 Thread Impala Public Jenkins (Code Review)
Impala Public Jenkins has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/14384 )

Change subject: IMPALA-8999: make union scheduling work with mt_dop
..


Patch Set 14: Verified+1


--
To view, visit http://gerrit.cloudera.org:8080/14384
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I0d2e9c86b530da3053e49d42b837dca0b1348ff2
Gerrit-Change-Number: 14384
Gerrit-PatchSet: 14
Gerrit-Owner: Tim Armstrong 
Gerrit-Reviewer: Andrew Sherman 
Gerrit-Reviewer: Bikramjeet Vig 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Reviewer: Tim Armstrong 
Gerrit-Comment-Date: Thu, 24 Oct 2019 04:41:57 +
Gerrit-HasComments: No


[Impala-ASF-CR] IMPALA-8999: make union scheduling work with mt dop

2019-10-23 Thread Impala Public Jenkins (Code Review)
Impala Public Jenkins has submitted this change and it was merged. ( 
http://gerrit.cloudera.org:8080/14384 )

Change subject: IMPALA-8999: make union scheduling work with mt_dop
..

IMPALA-8999: make union scheduling work with mt_dop

This change unifies mt_dop scheduling between the
union and scan cases.

Testing:
Manually checked that fragments with unions get parallelised
to the correct degree, both as a result of scans within the
fragment and input fragments.

Extend TestMtDopAdmissionSlots (renamed to TestMtDopScheduling)
to confirm that queries that were not parallelised before are
now parallelised. These tests verify the number of instances
of each operator using the ExecSummary embedded in the profile.

Change-Id: I0d2e9c86b530da3053e49d42b837dca0b1348ff2
Reviewed-on: http://gerrit.cloudera.org:8080/14384
Reviewed-by: Impala Public Jenkins 
Tested-by: Impala Public Jenkins 
---
M be/src/scheduling/query-schedule.cc
M be/src/scheduling/scheduler-test.cc
M be/src/scheduling/scheduler.cc
M be/src/scheduling/scheduler.h
M fe/src/main/java/org/apache/impala/planner/UnionNode.java
D 
testdata/workloads/functional-query/queries/QueryTest/mt-dop-parquet-admission-slots.test
A 
testdata/workloads/functional-query/queries/QueryTest/mt-dop-parquet-scheduling.test
M tests/query_test/test_mt_dop.py
8 files changed, 382 insertions(+), 219 deletions(-)

Approvals:
  Impala Public Jenkins: Looks good to me, approved; Verified

--
To view, visit http://gerrit.cloudera.org:8080/14384
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: merged
Gerrit-Change-Id: I0d2e9c86b530da3053e49d42b837dca0b1348ff2
Gerrit-Change-Number: 14384
Gerrit-PatchSet: 15
Gerrit-Owner: Tim Armstrong 
Gerrit-Reviewer: Andrew Sherman 
Gerrit-Reviewer: Bikramjeet Vig 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Reviewer: Tim Armstrong 


[Impala-ASF-CR] IMPALA-8999: make union scheduling work with mt dop

2019-10-23 Thread Impala Public Jenkins (Code Review)
Impala Public Jenkins has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/14384 )

Change subject: IMPALA-8999: make union scheduling work with mt_dop
..


Patch Set 14:

Build started: https://jenkins.impala.io/job/gerrit-verify-dryrun/5129/ 
DRY_RUN=false


--
To view, visit http://gerrit.cloudera.org:8080/14384
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I0d2e9c86b530da3053e49d42b837dca0b1348ff2
Gerrit-Change-Number: 14384
Gerrit-PatchSet: 14
Gerrit-Owner: Tim Armstrong 
Gerrit-Reviewer: Andrew Sherman 
Gerrit-Reviewer: Bikramjeet Vig 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Reviewer: Tim Armstrong 
Gerrit-Comment-Date: Thu, 24 Oct 2019 00:23:23 +
Gerrit-HasComments: No


[Impala-ASF-CR] IMPALA-8999: make union scheduling work with mt dop

2019-10-23 Thread Impala Public Jenkins (Code Review)
Impala Public Jenkins has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/14384 )

Change subject: IMPALA-8999: make union scheduling work with mt_dop
..


Patch Set 14: Code-Review+2


--
To view, visit http://gerrit.cloudera.org:8080/14384
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I0d2e9c86b530da3053e49d42b837dca0b1348ff2
Gerrit-Change-Number: 14384
Gerrit-PatchSet: 14
Gerrit-Owner: Tim Armstrong 
Gerrit-Reviewer: Andrew Sherman 
Gerrit-Reviewer: Bikramjeet Vig 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Reviewer: Tim Armstrong 
Gerrit-Comment-Date: Thu, 24 Oct 2019 00:23:22 +
Gerrit-HasComments: No


[Impala-ASF-CR] IMPALA-8999: make union scheduling work with mt dop

2019-10-23 Thread Bikramjeet Vig (Code Review)
Bikramjeet Vig has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/14384 )

Change subject: IMPALA-8999: make union scheduling work with mt_dop
..


Patch Set 13: Code-Review+2


--
To view, visit http://gerrit.cloudera.org:8080/14384
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I0d2e9c86b530da3053e49d42b837dca0b1348ff2
Gerrit-Change-Number: 14384
Gerrit-PatchSet: 13
Gerrit-Owner: Tim Armstrong 
Gerrit-Reviewer: Andrew Sherman 
Gerrit-Reviewer: Bikramjeet Vig 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Reviewer: Tim Armstrong 
Gerrit-Comment-Date: Wed, 23 Oct 2019 23:37:14 +
Gerrit-HasComments: No


[Impala-ASF-CR] IMPALA-8999: make union scheduling work with mt dop

2019-10-23 Thread Impala Public Jenkins (Code Review)
Impala Public Jenkins has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/14384 )

Change subject: IMPALA-8999: make union scheduling work with mt_dop
..


Patch Set 13:

Build Successful

https://jenkins.impala.io/job/gerrit-code-review-checks/4863/ : Initial code 
review checks passed. Use gerrit-verify-dryrun-external or gerrit-verify-dryrun 
to run full precommit tests.


--
To view, visit http://gerrit.cloudera.org:8080/14384
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I0d2e9c86b530da3053e49d42b837dca0b1348ff2
Gerrit-Change-Number: 14384
Gerrit-PatchSet: 13
Gerrit-Owner: Tim Armstrong 
Gerrit-Reviewer: Andrew Sherman 
Gerrit-Reviewer: Bikramjeet Vig 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Reviewer: Tim Armstrong 
Gerrit-Comment-Date: Wed, 23 Oct 2019 22:46:16 +
Gerrit-HasComments: No


[Impala-ASF-CR] IMPALA-8999: make union scheduling work with mt dop

2019-10-23 Thread Tim Armstrong (Code Review)
Tim Armstrong has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/14384 )

Change subject: IMPALA-8999: make union scheduling work with mt_dop
..


Patch Set 12:

(3 comments)

http://gerrit.cloudera.org:8080/#/c/14384/12/be/src/scheduling/scheduler.cc
File be/src/scheduling/scheduler.cc:

http://gerrit.cloudera.org:8080/#/c/14384/12/be/src/scheduling/scheduler.cc@338
PS12, Line 338:  for the scan range
> nit: remove?
Done


http://gerrit.cloudera.org:8080/#/c/14384/12/be/src/scheduling/scheduler.cc@344
PS12, Line 344: Note that this takes into account all of the input fragments,
  : // not just the leftmost because we expect unions to be 
symmetrical for purposes of
  : // planning, unlike joins.
> nit: how about: "Note that step 1 is modified to run on fragments with unio
Done


http://gerrit.cloudera.org:8080/#/c/14384/12/be/src/scheduling/scheduler.cc@392
PS12, Line 392:   vector scan_hosts;
> nit: maybe add a Dcheck(scan_node_ids.size() == 1 || has_union) so that we
Done



--
To view, visit http://gerrit.cloudera.org:8080/14384
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I0d2e9c86b530da3053e49d42b837dca0b1348ff2
Gerrit-Change-Number: 14384
Gerrit-PatchSet: 12
Gerrit-Owner: Tim Armstrong 
Gerrit-Reviewer: Andrew Sherman 
Gerrit-Reviewer: Bikramjeet Vig 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Reviewer: Tim Armstrong 
Gerrit-Comment-Date: Wed, 23 Oct 2019 22:01:44 +
Gerrit-HasComments: Yes


[Impala-ASF-CR] IMPALA-8999: make union scheduling work with mt dop

2019-10-23 Thread Tim Armstrong (Code Review)
Hello Andrew Sherman, Bikramjeet Vig, Impala Public Jenkins,

I'd like you to reexamine a change. Please visit

http://gerrit.cloudera.org:8080/14384

to look at the new patch set (#13).

Change subject: IMPALA-8999: make union scheduling work with mt_dop
..

IMPALA-8999: make union scheduling work with mt_dop

This change unifies mt_dop scheduling between the
union and scan cases.

Testing:
Manually checked that fragments with unions get parallelised
to the correct degree, both as a result of scans within the
fragment and input fragments.

Extend TestMtDopAdmissionSlots (renamed to TestMtDopScheduling)
to confirm that queries that were not parallelised before are
now parallelised. These tests verify the number of instances
of each operator using the ExecSummary embedded in the profile.

Change-Id: I0d2e9c86b530da3053e49d42b837dca0b1348ff2
---
M be/src/scheduling/query-schedule.cc
M be/src/scheduling/scheduler-test.cc
M be/src/scheduling/scheduler.cc
M be/src/scheduling/scheduler.h
M fe/src/main/java/org/apache/impala/planner/UnionNode.java
D 
testdata/workloads/functional-query/queries/QueryTest/mt-dop-parquet-admission-slots.test
A 
testdata/workloads/functional-query/queries/QueryTest/mt-dop-parquet-scheduling.test
M tests/query_test/test_mt_dop.py
8 files changed, 382 insertions(+), 219 deletions(-)


  git pull ssh://gerrit.cloudera.org:29418/Impala-ASF refs/changes/84/14384/13
--
To view, visit http://gerrit.cloudera.org:8080/14384
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: newpatchset
Gerrit-Change-Id: I0d2e9c86b530da3053e49d42b837dca0b1348ff2
Gerrit-Change-Number: 14384
Gerrit-PatchSet: 13
Gerrit-Owner: Tim Armstrong 
Gerrit-Reviewer: Andrew Sherman 
Gerrit-Reviewer: Bikramjeet Vig 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Reviewer: Tim Armstrong 


[Impala-ASF-CR] IMPALA-8999: make union scheduling work with mt dop

2019-10-21 Thread Bikramjeet Vig (Code Review)
Bikramjeet Vig has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/14384 )

Change subject: IMPALA-8999: make union scheduling work with mt_dop
..


Patch Set 12:

(4 comments)

http://gerrit.cloudera.org:8080/#/c/14384/12/be/src/scheduling/scheduler.cc
File be/src/scheduling/scheduler.cc:

http://gerrit.cloudera.org:8080/#/c/14384/12/be/src/scheduling/scheduler.cc@338
PS12, Line 338:  for the scan range
nit: remove?


http://gerrit.cloudera.org:8080/#/c/14384/12/be/src/scheduling/scheduler.cc@344
PS12, Line 344: Note that this takes into account all of the input fragments,
  : // not just the leftmost because we expect unions to be 
symmetrical for purposes of
  : // planning, unlike joins.
nit: how about: "Note that step 1 is modified to run on fragments with union 
nodes, by considering all input fragments and not just the leftmost because we 
expect unions to be symmetrical for purposes of planning, unlike joins."


http://gerrit.cloudera.org:8080/#/c/14384/12/be/src/scheduling/scheduler.cc@392
PS12, Line 392:   vector scan_hosts;
nit: maybe add a Dcheck(scan_node_ids.size() == 1 || has_union) so that we make 
sure this methods gets another look whenever this changes.


http://gerrit.cloudera.org:8080/#/c/14384/11/testdata/workloads/functional-query/queries/QueryTest/mt-dop-parquet-scheduling.test
File 
testdata/workloads/functional-query/queries/QueryTest/mt-dop-parquet-scheduling.test:

http://gerrit.cloudera.org:8080/#/c/14384/11/testdata/workloads/functional-query/queries/QueryTest/mt-dop-parquet-scheduling.test@133
PS11, Line 133: 
> I think this is covered by the tests on l32 and l52 - alltypes has 24 files
Got it, thanks for the explanation.



--
To view, visit http://gerrit.cloudera.org:8080/14384
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I0d2e9c86b530da3053e49d42b837dca0b1348ff2
Gerrit-Change-Number: 14384
Gerrit-PatchSet: 12
Gerrit-Owner: Tim Armstrong 
Gerrit-Reviewer: Andrew Sherman 
Gerrit-Reviewer: Bikramjeet Vig 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Reviewer: Tim Armstrong 
Gerrit-Comment-Date: Mon, 21 Oct 2019 20:42:40 +
Gerrit-HasComments: Yes


[Impala-ASF-CR] IMPALA-8999: make union scheduling work with mt dop

2019-10-19 Thread Impala Public Jenkins (Code Review)
Impala Public Jenkins has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/14384 )

Change subject: IMPALA-8999: make union scheduling work with mt_dop
..


Patch Set 12:

Build Successful

https://jenkins.impala.io/job/gerrit-code-review-checks/4836/ : Initial code 
review checks passed. Use gerrit-verify-dryrun-external or gerrit-verify-dryrun 
to run full precommit tests.


--
To view, visit http://gerrit.cloudera.org:8080/14384
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I0d2e9c86b530da3053e49d42b837dca0b1348ff2
Gerrit-Change-Number: 14384
Gerrit-PatchSet: 12
Gerrit-Owner: Tim Armstrong 
Gerrit-Reviewer: Andrew Sherman 
Gerrit-Reviewer: Bikramjeet Vig 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Reviewer: Tim Armstrong 
Gerrit-Comment-Date: Sat, 19 Oct 2019 20:55:55 +
Gerrit-HasComments: No


[Impala-ASF-CR] IMPALA-8999: make union scheduling work with mt dop

2019-10-19 Thread Tim Armstrong (Code Review)
Tim Armstrong has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/14384 )

Change subject: IMPALA-8999: make union scheduling work with mt_dop
..


Patch Set 11:

(9 comments)

Thanks for the feedback, this should make it a bit less cryptic.

http://gerrit.cloudera.org:8080/#/c/14384/11/be/src/scheduling/scheduler.h
File be/src/scheduling/scheduler.h:

http://gerrit.cloudera.org:8080/#/c/14384/11/be/src/scheduling/scheduler.h@381
PS11, Line 381: For HDFS, this attempts to load balance among instances by 
computing the average
  :   /// number of bytes per instances and then in a single pass 
assigning scan ranges to
  :   /// each instance to roughly meet that average.
> nit: looks like we forgot to update this in the commit for  IMPALA-9015
nice catch


http://gerrit.cloudera.org:8080/#/c/14384/11/be/src/scheduling/scheduler.cc
File be/src/scheduling/scheduler.cc:

http://gerrit.cloudera.org:8080/#/c/14384/11/be/src/scheduling/scheduler.cc@291
PS11, Line 291: one or more scan nodes
> for the cases with non-union fragments, we are now considering the parallel
We don't actually have plans like this at the moment. The operators that have 
multiple inputs are joins, which always have an exchange on the RHS, unions, 
and subplans, which never have a scan on the RHS.

That could potentially change, but that's the current state of things.

This makes me realise that case 2 is incorrectly described, there's no "or 
more".


http://gerrit.cloudera.org:8080/#/c/14384/11/be/src/scheduling/scheduler.cc@332
PS11, Line 332: Instance selection for an interior fragment
> what does instance selection and interior fragment mean here?
elaborated a bit


http://gerrit.cloudera.org:8080/#/c/14384/11/be/src/scheduling/scheduler.cc@360
PS11, Line 360: hosts
> nit: instances_per_host
Done


http://gerrit.cloudera.org:8080/#/c/14384/11/be/src/scheduling/scheduler.cc@363
PS11, Line 363: only factor in
> nit: got confused with this, read it as "only factor... in" vs the intended
Done


http://gerrit.cloudera.org:8080/#/c/14384/11/be/src/scheduling/scheduler.cc@368
PS11, Line 368: input_fragment_hosts
> nit: input_fragment_instances_per_host
Done


http://gerrit.cloudera.org:8080/#/c/14384/11/fe/src/main/java/org/apache/impala/planner/UnionNode.java
File fe/src/main/java/org/apache/impala/planner/UnionNode.java:

http://gerrit.cloudera.org:8080/#/c/14384/11/fe/src/main/java/org/apache/impala/planner/UnionNode.java@122
PS11, Line 122: Union fragments are scheduled on the union of hosts that the 
child fragments run
  :   // on.
> nit: Union fragments are scheduled on the union of hosts of all scans in th
Done


http://gerrit.cloudera.org:8080/#/c/14384/11/testdata/workloads/functional-query/queries/QueryTest/mt-dop-parquet-scheduling.test
File 
testdata/workloads/functional-query/queries/QueryTest/mt-dop-parquet-scheduling.test:

http://gerrit.cloudera.org:8080/#/c/14384/11/testdata/workloads/functional-query/queries/QueryTest/mt-dop-parquet-scheduling.test@52
PS11, Line 52: it
> nit: if
Done


http://gerrit.cloudera.org:8080/#/c/14384/11/testdata/workloads/functional-query/queries/QueryTest/mt-dop-parquet-scheduling.test@133
PS11, Line 133:
> add a test for a union fragment where the scan ranges for scan node are mor
I think this is covered by the tests on l32 and l52 - alltypes has 24 files, so 
at least 8 ranges per backend. I updated the comments to be specific about that.



--
To view, visit http://gerrit.cloudera.org:8080/14384
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I0d2e9c86b530da3053e49d42b837dca0b1348ff2
Gerrit-Change-Number: 14384
Gerrit-PatchSet: 11
Gerrit-Owner: Tim Armstrong 
Gerrit-Reviewer: Andrew Sherman 
Gerrit-Reviewer: Bikramjeet Vig 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Reviewer: Tim Armstrong 
Gerrit-Comment-Date: Sat, 19 Oct 2019 20:11:40 +
Gerrit-HasComments: Yes


[Impala-ASF-CR] IMPALA-8999: make union scheduling work with mt dop

2019-10-19 Thread Tim Armstrong (Code Review)
Hello Andrew Sherman, Bikramjeet Vig, Impala Public Jenkins,

I'd like you to reexamine a change. Please visit

http://gerrit.cloudera.org:8080/14384

to look at the new patch set (#12).

Change subject: IMPALA-8999: make union scheduling work with mt_dop
..

IMPALA-8999: make union scheduling work with mt_dop

This change unifies mt_dop scheduling between the
union and scan cases.

Testing:
Manually checked that fragments with unions get parallelised
to the correct degree, both as a result of scans within the
fragment and input fragments.

Extend TestMtDopAdmissionSlots (renamed to TestMtDopScheduling)
to confirm that queries that were not parallelised before are
now parallelised. These tests verify the number of instances
of each operator using the ExecSummary embedded in the profile.

Change-Id: I0d2e9c86b530da3053e49d42b837dca0b1348ff2
---
M be/src/scheduling/query-schedule.cc
M be/src/scheduling/scheduler-test.cc
M be/src/scheduling/scheduler.cc
M be/src/scheduling/scheduler.h
M fe/src/main/java/org/apache/impala/planner/UnionNode.java
D 
testdata/workloads/functional-query/queries/QueryTest/mt-dop-parquet-admission-slots.test
A 
testdata/workloads/functional-query/queries/QueryTest/mt-dop-parquet-scheduling.test
M tests/query_test/test_mt_dop.py
8 files changed, 380 insertions(+), 219 deletions(-)


  git pull ssh://gerrit.cloudera.org:29418/Impala-ASF refs/changes/84/14384/12
--
To view, visit http://gerrit.cloudera.org:8080/14384
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: newpatchset
Gerrit-Change-Id: I0d2e9c86b530da3053e49d42b837dca0b1348ff2
Gerrit-Change-Number: 14384
Gerrit-PatchSet: 12
Gerrit-Owner: Tim Armstrong 
Gerrit-Reviewer: Andrew Sherman 
Gerrit-Reviewer: Bikramjeet Vig 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Reviewer: Tim Armstrong 


[Impala-ASF-CR] IMPALA-8999: make union scheduling work with mt dop

2019-10-18 Thread Bikramjeet Vig (Code Review)
Bikramjeet Vig has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/14384 )

Change subject: IMPALA-8999: make union scheduling work with mt_dop
..


Patch Set 11:

(9 comments)

http://gerrit.cloudera.org:8080/#/c/14384/11/be/src/scheduling/scheduler.h
File be/src/scheduling/scheduler.h:

http://gerrit.cloudera.org:8080/#/c/14384/11/be/src/scheduling/scheduler.h@381
PS11, Line 381: For HDFS, this attempts to load balance among instances by 
computing the average
  :   /// number of bytes per instances and then in a single pass 
assigning scan ranges to
  :   /// each instance to roughly meet that average.
nit: looks like we forgot to update this in the commit for  IMPALA-9015


http://gerrit.cloudera.org:8080/#/c/14384/11/be/src/scheduling/scheduler.cc
File be/src/scheduling/scheduler.cc:

http://gerrit.cloudera.org:8080/#/c/14384/11/be/src/scheduling/scheduler.cc@291
PS11, Line 291: one or more scan nodes
for the cases with non-union fragments, we are now considering the parallelism 
of all scan nodes in the fragment  as opposed to the left most scan node, what 
are the cases that this can happen in a plan? Maybe add that as a test case too


http://gerrit.cloudera.org:8080/#/c/14384/11/be/src/scheduling/scheduler.cc@332
PS11, Line 332: Instance selection for an interior fragment
what does instance selection and interior fragment mean here?


http://gerrit.cloudera.org:8080/#/c/14384/11/be/src/scheduling/scheduler.cc@360
PS11, Line 360: hosts
nit: instances_per_host


http://gerrit.cloudera.org:8080/#/c/14384/11/be/src/scheduling/scheduler.cc@363
PS11, Line 363: only factor in
nit: got confused with this, read it as "only factor... in" vs the intended 
meaning of "only... factor in". maybe say "only consider parallelism"


http://gerrit.cloudera.org:8080/#/c/14384/11/be/src/scheduling/scheduler.cc@368
PS11, Line 368: input_fragment_hosts
nit: input_fragment_instances_per_host


http://gerrit.cloudera.org:8080/#/c/14384/11/fe/src/main/java/org/apache/impala/planner/UnionNode.java
File fe/src/main/java/org/apache/impala/planner/UnionNode.java:

http://gerrit.cloudera.org:8080/#/c/14384/11/fe/src/main/java/org/apache/impala/planner/UnionNode.java@122
PS11, Line 122: Union fragments are scheduled on the union of hosts that the 
child fragments run
  :   // on.
nit: Union fragments are scheduled on the union of hosts of all scans in the 
fragment as well as the hosts of all its input fragments


http://gerrit.cloudera.org:8080/#/c/14384/11/testdata/workloads/functional-query/queries/QueryTest/mt-dop-parquet-scheduling.test
File 
testdata/workloads/functional-query/queries/QueryTest/mt-dop-parquet-scheduling.test:

http://gerrit.cloudera.org:8080/#/c/14384/11/testdata/workloads/functional-query/queries/QueryTest/mt-dop-parquet-scheduling.test@52
PS11, Line 52: it
nit: if


http://gerrit.cloudera.org:8080/#/c/14384/11/testdata/workloads/functional-query/queries/QueryTest/mt-dop-parquet-scheduling.test@133
PS11, Line 133:
add a test for a union fragment where the scan ranges for scan node are more 
than 4 but the max instances gets bounded by mt_dop



--
To view, visit http://gerrit.cloudera.org:8080/14384
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I0d2e9c86b530da3053e49d42b837dca0b1348ff2
Gerrit-Change-Number: 14384
Gerrit-PatchSet: 11
Gerrit-Owner: Tim Armstrong 
Gerrit-Reviewer: Andrew Sherman 
Gerrit-Reviewer: Bikramjeet Vig 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Reviewer: Tim Armstrong 
Gerrit-Comment-Date: Sat, 19 Oct 2019 00:54:34 +
Gerrit-HasComments: Yes


[Impala-ASF-CR] IMPALA-8999: make union scheduling work with mt dop

2019-10-17 Thread Tim Armstrong (Code Review)
Tim Armstrong has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/14384 )

Change subject: IMPALA-8999: make union scheduling work with mt_dop
..


Patch Set 11:

Exhaustive tests passed


--
To view, visit http://gerrit.cloudera.org:8080/14384
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I0d2e9c86b530da3053e49d42b837dca0b1348ff2
Gerrit-Change-Number: 14384
Gerrit-PatchSet: 11
Gerrit-Owner: Tim Armstrong 
Gerrit-Reviewer: Andrew Sherman 
Gerrit-Reviewer: Bikramjeet Vig 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Reviewer: Tim Armstrong 
Gerrit-Comment-Date: Thu, 17 Oct 2019 15:45:58 +
Gerrit-HasComments: No


[Impala-ASF-CR] IMPALA-8999: make union scheduling work with mt dop

2019-10-16 Thread Impala Public Jenkins (Code Review)
Impala Public Jenkins has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/14384 )

Change subject: IMPALA-8999: make union scheduling work with mt_dop
..


Patch Set 11:

Build Successful

https://jenkins.impala.io/job/gerrit-code-review-checks/4810/ : Initial code 
review checks passed. Use gerrit-verify-dryrun-external or gerrit-verify-dryrun 
to run full precommit tests.


--
To view, visit http://gerrit.cloudera.org:8080/14384
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I0d2e9c86b530da3053e49d42b837dca0b1348ff2
Gerrit-Change-Number: 14384
Gerrit-PatchSet: 11
Gerrit-Owner: Tim Armstrong 
Gerrit-Reviewer: Andrew Sherman 
Gerrit-Reviewer: Bikramjeet Vig 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Reviewer: Tim Armstrong 
Gerrit-Comment-Date: Thu, 17 Oct 2019 00:23:20 +
Gerrit-HasComments: No


[Impala-ASF-CR] IMPALA-8999: make union scheduling work with mt dop

2019-10-16 Thread Tim Armstrong (Code Review)
Tim Armstrong has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/14384 )

Change subject: IMPALA-8999: make union scheduling work with mt_dop
..


Patch Set 11:

I'm running exhaustive tests still, but this should be ready for review.


--
To view, visit http://gerrit.cloudera.org:8080/14384
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I0d2e9c86b530da3053e49d42b837dca0b1348ff2
Gerrit-Change-Number: 14384
Gerrit-PatchSet: 11
Gerrit-Owner: Tim Armstrong 
Gerrit-Reviewer: Bikramjeet Vig 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Reviewer: Tim Armstrong 
Gerrit-Comment-Date: Wed, 16 Oct 2019 23:40:47 +
Gerrit-HasComments: No


[Impala-ASF-CR] IMPALA-8999: make union scheduling work with mt dop

2019-10-16 Thread Tim Armstrong (Code Review)
Hello Bikramjeet Vig, Impala Public Jenkins,

I'd like you to reexamine a change. Please visit

http://gerrit.cloudera.org:8080/14384

to look at the new patch set (#11).

Change subject: IMPALA-8999: make union scheduling work with mt_dop
..

IMPALA-8999: make union scheduling work with mt_dop

This change unifies mt_dop scheduling between the
union and scan cases.

Testing:
Manually checked that fragments with unions get parallelised
to the correct degree, both as a result of scans within the
fragment and input fragments.

Extend TestMtDopAdmissionSlots (renamed to TestMtDopScheduling)
to confirm that queries that were not parallelised before are
now parallelised. These tests verify the number of instances
of each operator using the ExecSummary embedded in the profile.

Change-Id: I0d2e9c86b530da3053e49d42b837dca0b1348ff2
---
M be/src/scheduling/query-schedule.cc
M be/src/scheduling/scheduler-test.cc
M be/src/scheduling/scheduler.cc
M be/src/scheduling/scheduler.h
M fe/src/main/java/org/apache/impala/planner/UnionNode.java
D 
testdata/workloads/functional-query/queries/QueryTest/mt-dop-parquet-admission-slots.test
A 
testdata/workloads/functional-query/queries/QueryTest/mt-dop-parquet-scheduling.test
M tests/query_test/test_mt_dop.py
8 files changed, 372 insertions(+), 213 deletions(-)


  git pull ssh://gerrit.cloudera.org:29418/Impala-ASF refs/changes/84/14384/11
--
To view, visit http://gerrit.cloudera.org:8080/14384
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: newpatchset
Gerrit-Change-Id: I0d2e9c86b530da3053e49d42b837dca0b1348ff2
Gerrit-Change-Number: 14384
Gerrit-PatchSet: 11
Gerrit-Owner: Tim Armstrong 
Gerrit-Reviewer: Bikramjeet Vig 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Reviewer: Tim Armstrong