You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@doris.apache.org by GitBox <gi...@apache.org> on 2020/12/17 08:08:50 UTC
[GitHub] [incubator-doris] xinghuayu007 opened a new pull request #5104: [Enhancement]Make Cholocate table join more load balance
xinghuayu007 opened a new pull request #5104:
URL: https://github.com/apache/incubator-doris/pull/5104
## Proposed changes
When two cholocate tables make `join` operation, to make `join` operation locally, the tablet belongs to the same bucket sequence will be distributed to the same host. When chosing which host for a bucket sequence, it takes random strategy. Random strategy can not make query task load balance logically for one query. Therefore, this patch takes round-robin strategy, make buckets distributed evenly. For example, if there are 6 bucket sequences and 3 hosts, it is better to distributed 2 buket sequence for every host.
## Types of changes
What types of changes does your code introduce to Doris?
_Put an `x` in the boxes that apply_
- [] Bugfix (non-breaking change which fixes an issue)
- [] New feature (non-breaking change which adds functionality)
- [x] Breaking change (fix or feature that would cause existing functionality to not work as expected)
- [] Documentation Update (if none of the other choices apply)
- [] Code refactor (Modify the code structure, format the code, etc...)
## Checklist
_Put an `x` in the boxes that apply. You can also fill these out after creating the PR. If you're unsure about any of them, don't hesitate to ask. We're here to help! This is simply a reminder of what we are going to look for before merging your code._
- [] I have create an issue on (Fix #ISSUE), and have described the bug/feature there in detail
- [x] Compiling and unit tests pass locally with my changes
- [x] I have added tests that prove my fix is effective or that my feature works
- [x] If this change need a document change, I have updated the document
- [x] Any dependent changes have been merged
## Further comments
If this is a relatively large or complex change, kick off the discussion at dev@doris.apache.org by explaining why you chose the solution you did and what alternatives you considered, etc...
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
[GitHub] [incubator-doris] xinghuayu007 commented on a change in pull request #5104: [Enhancement]Make Cholocate table join more load balance
Posted by GitBox <gi...@apache.org>.
xinghuayu007 commented on a change in pull request #5104:
URL: https://github.com/apache/incubator-doris/pull/5104#discussion_r549651544
##########
File path: fe/fe-core/src/main/java/org/apache/doris/qe/SimpleScheduler.java
##########
@@ -92,6 +92,38 @@ public static TNetworkAddress getHost(long backendId,
backends, locations.size()));
}
+ public static TScanRangeLocation getLocation(TScanRangeLocation minLocation,
Review comment:
Because function `computeScanRangeAssignmentByScheduler` need `minLocation` to set volume_id. Therefore it must return param `minLocation`
// Volume is optional, so we need to set the value and the is-set bit
scanRangeParams.setVolumeId(minLocation.volume_id);
##########
File path: fe/fe-core/src/main/java/org/apache/doris/qe/Coordinator.java
##########
@@ -1275,15 +1275,40 @@ private void computeScanRangeAssignmentByColocate(
}
}
- // randomly choose a backend from the TScanRangeLocations for a certain bucket sequence.
- private void getExecHostPortForFragmentIDAndBucketSeq(TScanRangeLocations seqLocation, PlanFragmentId fragmentId, Integer bucketSeq) throws Exception {
- int randomLocation = new Random().nextInt(seqLocation.locations.size());
+ //ensure bucket sequence distribued to every host evenly
+ private void getExecHostPortForFragmentIDAndBucketSeq(TScanRangeLocations seqLocation, PlanFragmentId fragmentId, Integer bucketSeq,
+ HashMap<TNetworkAddress, Long> assignedBytesPerHost) throws Exception {
Reference<Long> backendIdRef = new Reference<Long>();
- TNetworkAddress execHostPort = SimpleScheduler.getHost(seqLocation.locations.get(randomLocation).backend_id, seqLocation.locations, this.idToBackend, backendIdRef);
+ distributeHost(seqLocation, assignedBytesPerHost, backendIdRef);
+ Backend backend = this.idToBackend.get(backendIdRef.getRef());
+ TNetworkAddress execHostPort = new TNetworkAddress(backend.getHost(), backend.getBePort());
this.addressToBackendID.put(execHostPort, backendIdRef.getRef());
this.fragmentIdToSeqToAddressMap.get(fragmentId).put(bucketSeq, execHostPort);
}
+ public TScanRangeLocation distributeHost(TScanRangeLocations seqLocation,
Review comment:
ok. I will modify it.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
[GitHub] [incubator-doris] kangkaisen commented on a change in pull request #5104: [Enhancement]Make Cholocate table join more load balance
Posted by GitBox <gi...@apache.org>.
kangkaisen commented on a change in pull request #5104:
URL: https://github.com/apache/incubator-doris/pull/5104#discussion_r549550083
##########
File path: fe/fe-core/src/main/java/org/apache/doris/qe/Coordinator.java
##########
@@ -1276,10 +1276,22 @@ private void computeScanRangeAssignmentByColocate(
}
// randomly choose a backend from the TScanRangeLocations for a certain bucket sequence.
- private void getExecHostPortForFragmentIDAndBucketSeq(TScanRangeLocations seqLocation, PlanFragmentId fragmentId, Integer bucketSeq) throws Exception {
- int randomLocation = new Random().nextInt(seqLocation.locations.size());
+ private void getExecHostPortForFragmentIDAndBucketSeq(TScanRangeLocations seqLocation, PlanFragmentId fragmentId, Integer bucketSeq,
+ HashMap<TNetworkAddress, Long> assignedBytesPerHost) throws Exception {
+ Long minAssignedBytes = Long.MAX_VALUE;
Review comment:
These codes are also used in `computeScanRangeAssignmentByScheduler`, would better abstract these codes to a method
```
Long minAssignedBytes = Long.MAX_VALUE;
TScanRangeLocation minLocation = null;
for (final TScanRangeLocation location : scanRangeLocations.getLocations()) {
Long assignedBytes = findOrInsert(assignedBytesPerHost, location.server, 0L);
if (assignedBytes < minAssignedBytes) {
minAssignedBytes = assignedBytes;
minLocation = location;
}
}
```
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
[GitHub] [incubator-doris] kangkaisen commented on a change in pull request #5104: [Enhancement]Make Cholocate table join more load balance
Posted by GitBox <gi...@apache.org>.
kangkaisen commented on a change in pull request #5104:
URL: https://github.com/apache/incubator-doris/pull/5104#discussion_r549645737
##########
File path: fe/fe-core/src/main/java/org/apache/doris/qe/SimpleScheduler.java
##########
@@ -92,6 +92,38 @@ public static TNetworkAddress getHost(long backendId,
backends, locations.size()));
}
+ public static TScanRangeLocation getLocation(TScanRangeLocation minLocation,
Review comment:
Why add this method? The diff between `getLocation` and `getHost` is first arg, but we could get `long backendId` from `TScanRangeLocation minLocation`?
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
[GitHub] [incubator-doris] morningman merged pull request #5104: [Enhancement]Make Cholocate table join more load balance
Posted by GitBox <gi...@apache.org>.
morningman merged pull request #5104:
URL: https://github.com/apache/incubator-doris/pull/5104
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
[GitHub] [incubator-doris] xinghuayu007 commented on a change in pull request #5104: [Enhancement]Make Cholocate table join more load balance
Posted by GitBox <gi...@apache.org>.
xinghuayu007 commented on a change in pull request #5104:
URL: https://github.com/apache/incubator-doris/pull/5104#discussion_r549629355
##########
File path: fe/fe-core/src/main/java/org/apache/doris/qe/Coordinator.java
##########
@@ -1276,10 +1276,22 @@ private void computeScanRangeAssignmentByColocate(
}
// randomly choose a backend from the TScanRangeLocations for a certain bucket sequence.
- private void getExecHostPortForFragmentIDAndBucketSeq(TScanRangeLocations seqLocation, PlanFragmentId fragmentId, Integer bucketSeq) throws Exception {
- int randomLocation = new Random().nextInt(seqLocation.locations.size());
+ private void getExecHostPortForFragmentIDAndBucketSeq(TScanRangeLocations seqLocation, PlanFragmentId fragmentId, Integer bucketSeq,
+ HashMap<TNetworkAddress, Long> assignedBytesPerHost) throws Exception {
+ Long minAssignedBytes = Long.MAX_VALUE;
Review comment:
Thanks for your code review. The code has been improved.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
[GitHub] [incubator-doris] kangkaisen commented on a change in pull request #5104: [Enhancement]Make Cholocate table join more load balance
Posted by GitBox <gi...@apache.org>.
kangkaisen commented on a change in pull request #5104:
URL: https://github.com/apache/incubator-doris/pull/5104#discussion_r549646655
##########
File path: fe/fe-core/src/main/java/org/apache/doris/qe/Coordinator.java
##########
@@ -1275,15 +1275,40 @@ private void computeScanRangeAssignmentByColocate(
}
}
- // randomly choose a backend from the TScanRangeLocations for a certain bucket sequence.
- private void getExecHostPortForFragmentIDAndBucketSeq(TScanRangeLocations seqLocation, PlanFragmentId fragmentId, Integer bucketSeq) throws Exception {
- int randomLocation = new Random().nextInt(seqLocation.locations.size());
+ //ensure bucket sequence distribued to every host evenly
+ private void getExecHostPortForFragmentIDAndBucketSeq(TScanRangeLocations seqLocation, PlanFragmentId fragmentId, Integer bucketSeq,
+ HashMap<TNetworkAddress, Long> assignedBytesPerHost) throws Exception {
Reference<Long> backendIdRef = new Reference<Long>();
- TNetworkAddress execHostPort = SimpleScheduler.getHost(seqLocation.locations.get(randomLocation).backend_id, seqLocation.locations, this.idToBackend, backendIdRef);
+ distributeHost(seqLocation, assignedBytesPerHost, backendIdRef);
+ Backend backend = this.idToBackend.get(backendIdRef.getRef());
+ TNetworkAddress execHostPort = new TNetworkAddress(backend.getHost(), backend.getBePort());
this.addressToBackendID.put(execHostPort, backendIdRef.getRef());
this.fragmentIdToSeqToAddressMap.get(fragmentId).put(bucketSeq, execHostPort);
}
+ public TScanRangeLocation distributeHost(TScanRangeLocations seqLocation,
Review comment:
I think we need a more concrete name. Such as `selectBackendsByRoundRobin`
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
[GitHub] [incubator-doris] kangkaisen commented on a change in pull request #5104: [Enhancement]Make Cholocate table join more load balance
Posted by GitBox <gi...@apache.org>.
kangkaisen commented on a change in pull request #5104:
URL: https://github.com/apache/incubator-doris/pull/5104#discussion_r549549710
##########
File path: fe/fe-core/src/main/java/org/apache/doris/qe/Coordinator.java
##########
@@ -1276,10 +1276,22 @@ private void computeScanRangeAssignmentByColocate(
}
// randomly choose a backend from the TScanRangeLocations for a certain bucket sequence.
Review comment:
Please update the old comment
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org