You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@druid.apache.org by GitBox <gi...@apache.org> on 2021/11/19 22:54:51 UTC
[GitHub] [druid] capistrant opened a new pull request #11960: clean up the balancing code around the batched vs deprecated way of sampling segments to balance
capistrant opened a new pull request #11960:
URL: https://github.com/apache/druid/pull/11960
<!-- Thanks for trying to help us make Apache Druid be the best it can be! Please fill out as much of the following information as is possible (where relevant, and remove it when irrelevant) to help make the intention and scope of this PR clear in order to ease review. -->
<!-- Please read the doc for contribution (https://github.com/apache/druid/blob/master/CONTRIBUTING.md) before making this PR. Also, once you open a PR, please _avoid using force pushes and rebasing_ since these make it difficult for reviewers to see what you've changed in response to their reviews. See [the 'If your pull request shows conflicts with master' section](https://github.com/apache/druid/blob/master/CONTRIBUTING.md#if-your-pull-request-shows-conflicts-with-master) for more details. -->
<!-- Replace XXXX with the id of the issue fixed in this PR. Remove this section if there is no corresponding issue. Don't reference the issue in the title of this pull-request. -->
<!-- If you are a committer, follow the PR action item checklist for committers:
https://github.com/apache/druid/blob/master/dev/committer-instructions.md#pr-and-issue-action-item-checklist-for-committers. -->
### Description
<!-- Describe the goal of this PR, what problem are you fixing. If there is a corresponding issue (referenced above), it's not necessary to repeat the description here, however, you may choose to keep one summary sentence. -->
<!-- Describe your patch: what did you change in code? How did you fix the problem? -->
<!-- If there are several relatively logically separate changes in this PR, create a mini-section for each of them. For example: -->
#### Fixed unintended code path execution in BalancerStrategy#pickSegmentsToMove
https://github.com/apache/druid/pull/11257 Introduced a new way of sampling segments for balancing that is much more performant than the legacy way that exists today. There is a case in the code where the deprecated code can be executed despite, the dynamic config for batched segment sampling being enabled. In [this](https://github.com/yuanlihan/druid/blob/b02a9e659846a88c7be375ffb805340c38d1dbf5/server/src/main/java/org/apache/druid/server/coordinator/duty/BalanceSegments.java#L198) line of code, if `useBatchedSegmentSampler() == true` and `maxSegmentsToMove == 1`, you would expect the batched segment sampler code to execute. However, you can see [here](https://github.com/yuanlihan/druid/blob/b02a9e659846a88c7be375ffb805340c38d1dbf5/server/src/main/java/org/apache/druid/server/coordinator/BalancerStrategy.java#L82) that the new batched sampler is not used in this case because reservoir size is 1. This PR fixes this behavior by splitting `BalancerStrategy#pickSegmentsToMove(List<S
erverHolder>, Set<String>, int, double)` into two methods using overloading - one for the new batched sampler, and one for the deprecated way of sampling one segment at a time. This change will also simplify cleanup of the deprecated code in what I assume will be the 0.24.X cycle.
#### Cleaned up the code for the competing segment pick strategies (deprecated vs batched segment sampler)
I did a few things to make the code more understandable and to make it abundantly clear that the old way of sampling segments to move with the help of the dynamic config `percentOfSegmentsToConsiderPerMove` is deprecated:
* Added `@Deprecated` annotations to the `CoordinatorDynamicConfig` code related to `percentOfSegmentsToConsiderPerMove`
* Updated the Coordinator Dynamic Config docs to explicitly state that `percentOfSegmentsToConsiderPerMove` is deprecated and will have no effect if `useBatchedSegmentSampler` dynamic config is `true`
* Split `BalancerStrategy#pickSegmentsToMove` into two methods with the same name and overloaded arguments. One implementation is deprecated and takes a double for `percentOfSegmentsToConsiderPerMove` while the other is the new batched sampler which takes an `int` for the reservoir size.
* This change also makes `BalanceSegments#balanceServers` more understandable to those with limited familiarity with this work to migrate to the batched segment sampler. Before the call to `BalancerStrategy#pickSegmentsToMove` was opaque because it included both the reservoir size and the percent of segments to consider per move despite only one of those arguments being used depending on the value of the reservoir size (this also relates to the unintended code execution I talk about above). The new way makes it clear that if batched segment sampling is enabled, the non-deprecated `pickSegmentsToMove` is called. And if it is not, the deprecated `pickSegmentsToMove` is called.
<!--
In each section, please describe design decisions made, including:
- Choice of algorithms
- Behavioral aspects. What configuration values are acceptable? How are corner cases and error conditions handled, such as when there are insufficient resources?
- Class organization and design (how the logic is split between classes, inheritance, composition, design patterns)
- Method organization and design (how the logic is split between methods, parameters and return types)
- Naming (class, method, API, configuration, HTTP endpoint, names of emitted metrics)
-->
<!-- It's good to describe an alternative design (or mention an alternative name) for every design (or naming) decision point and compare the alternatives with the designs that you've implemented (or the names you've chosen) to highlight the advantages of the chosen designs and names. -->
<!-- If there was a discussion of the design of the feature implemented in this PR elsewhere (e. g. a "Proposal" issue, any other issue, or a thread in the development mailing list), link to that discussion from this PR description and explain what have changed in your final design compared to your original proposal or the consensus version in the end of the discussion. If something hasn't changed since the original discussion, you can omit a detailed discussion of those aspects of the design here, perhaps apart from brief mentioning for the sake of readability of this PR description. -->
<!-- Some of the aspects mentioned above may be omitted for simple and small changes. -->
<hr>
##### Key changed/added classes in this PR
* `BalanceSegments`
* `CoordinatorDynamicConfig`
* `BalancerStrategy`
<hr>
<!-- Check the items by putting "x" in the brackets for the done things. Not all of these items apply to every PR. Remove the items which are not done or not relevant to the PR. None of the items from the checklist below are strictly necessary, but it would be very helpful if you at least self-review the PR. -->
This PR has:
- [ ] been self-reviewed.
- [X] added documentation for new or modified features or behaviors.
- [X] added Javadocs for most classes and all non-trivial methods. Linked related entities via Javadoc links.
- [X] added comments explaining the "why" and the intent of the code wherever would not be obvious for an unfamiliar reader.
- [X] added unit tests or modified existing tests to cover new code paths, ensuring the threshold for [code coverage](https://github.com/apache/druid/blob/master/dev/code-review/code-coverage.md) is met.
- [ ] been tested in a test Druid cluster.
--
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.
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For additional commands, e-mail: commits-help@druid.apache.org
[GitHub] [druid] yuanlihan commented on a change in pull request #11960: clean up the balancing code around the batched vs deprecated way of sampling segments to balance
Posted by GitBox <gi...@apache.org>.
yuanlihan commented on a change in pull request #11960:
URL: https://github.com/apache/druid/pull/11960#discussion_r754046654
##########
File path: server/src/main/java/org/apache/druid/server/coordinator/duty/BalanceSegments.java
##########
@@ -202,12 +202,20 @@ private void balanceTier(
final int maxToLoad = params.getCoordinatorDynamicConfig().getMaxSegmentsInNodeLoadingQueue();
int moved = 0, unmoved = 0;
- Iterator<BalancerSegmentHolder> segmentsToMove = strategy.pickSegmentsToMove(
- toMoveFrom,
- params.getBroadcastDatasources(),
- params.getCoordinatorDynamicConfig().useBatchedSegmentSampler() ? maxSegmentsToMove : DEFAULT_RESERVOIR_SIZE,
Review comment:
I think we can delete this unused constant variable
--
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.
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For additional commands, e-mail: commits-help@druid.apache.org
[GitHub] [druid] capistrant commented on pull request #11960: clean up the balancing code around the batched vs deprecated way of sampling segments to balance
Posted by GitBox <gi...@apache.org>.
capistrant commented on pull request #11960:
URL: https://github.com/apache/druid/pull/11960#issuecomment-974525318
@yuanlihan could you take a look at this since you wrote the original code for the batched sampler! I hope my code makes sense, but please comment on any mistakes I may have made!
--
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.
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For additional commands, e-mail: commits-help@druid.apache.org
[GitHub] [druid] a2l007 commented on pull request #11960: clean up the balancing code around the batched vs deprecated way of sampling segments to balance
Posted by GitBox <gi...@apache.org>.
a2l007 commented on pull request #11960:
URL: https://github.com/apache/druid/pull/11960#issuecomment-976850081
Tagging this under Release Notes so that users are aware of the deprecation of `percentOfSegmentsToConsiderPerMove`
--
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.
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For additional commands, e-mail: commits-help@druid.apache.org
[GitHub] [druid] capistrant commented on pull request #11960: clean up the balancing code around the batched vs deprecated way of sampling segments to balance
Posted by GitBox <gi...@apache.org>.
capistrant commented on pull request #11960:
URL: https://github.com/apache/druid/pull/11960#issuecomment-988037607
@a2l007 @yuanlihan thank you both for reviewing! I synced up with master and plan on merging once CI is green.
--
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.
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For additional commands, e-mail: commits-help@druid.apache.org
[GitHub] [druid] jihoonson commented on pull request #11960: clean up the balancing code around the batched vs deprecated way of sampling segments to balance
Posted by GitBox <gi...@apache.org>.
jihoonson commented on pull request #11960:
URL: https://github.com/apache/druid/pull/11960#issuecomment-988316724
The "docs' CI passed. Merging this PR.
--
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.
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For additional commands, e-mail: commits-help@druid.apache.org
[GitHub] [druid] jihoonson merged pull request #11960: clean up the balancing code around the batched vs deprecated way of sampling segments to balance
Posted by GitBox <gi...@apache.org>.
jihoonson merged pull request #11960:
URL: https://github.com/apache/druid/pull/11960
--
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.
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For additional commands, e-mail: commits-help@druid.apache.org
[GitHub] [druid] jihoonson commented on a change in pull request #11960: clean up the balancing code around the batched vs deprecated way of sampling segments to balance
Posted by GitBox <gi...@apache.org>.
jihoonson commented on a change in pull request #11960:
URL: https://github.com/apache/druid/pull/11960#discussion_r764238416
##########
File path: web-console/src/druid-models/coordinator-dynamic-config.tsx
##########
@@ -213,14 +213,16 @@ export const COORDINATOR_DYNAMIC_CONFIG_FIELDS: Field<CoordinatorDynamicConfig>[
defaultValue: 100,
info: (
<>
- The percentage of the total number of segments in the cluster that are considered every time
- a segment needs to be selected for a move. Druid orders servers by available capacity
- ascending (the least available capacity first) and then iterates over the servers. For each
- server, Druid iterates over the segments on the server, considering them for moving. The
- default config of 100% means that every segment on every server is a candidate to be moved.
- This should make sense for most small to medium-sized clusters. However, an admin may find
- it preferable to drop this value lower if they don't think that it is worthwhile to
- consider every single segment in the cluster each time it is looking for a segment to move.
+ Deprecated. This will be phased out by the batched segment sampler. Only used if
Review comment:
Sorry for ambiguity. Yes, encouraging using `useBatchedSegmentSampler` instead will be enough. The one in your comment looks good to me.
--
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.
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For additional commands, e-mail: commits-help@druid.apache.org
[GitHub] [druid] capistrant commented on a change in pull request #11960: clean up the balancing code around the batched vs deprecated way of sampling segments to balance
Posted by GitBox <gi...@apache.org>.
capistrant commented on a change in pull request #11960:
URL: https://github.com/apache/druid/pull/11960#discussion_r764235666
##########
File path: web-console/src/druid-models/coordinator-dynamic-config.tsx
##########
@@ -213,14 +213,16 @@ export const COORDINATOR_DYNAMIC_CONFIG_FIELDS: Field<CoordinatorDynamicConfig>[
defaultValue: 100,
info: (
<>
- The percentage of the total number of segments in the cluster that are considered every time
- a segment needs to be selected for a move. Druid orders servers by available capacity
- ascending (the least available capacity first) and then iterates over the servers. For each
- server, Druid iterates over the segments on the server, considering them for moving. The
- default config of 100% means that every segment on every server is a candidate to be moved.
- This should make sense for most small to medium-sized clusters. However, an admin may find
- it preferable to drop this value lower if they don't think that it is worthwhile to
- consider every single segment in the cluster each time it is looking for a segment to move.
+ Deprecated. This will be phased out by the batched segment sampler. Only used if
Review comment:
could you clarify what you mean by this? I take it as you asking me to state that the preferred alternative is to set `useBatchedSegmentSampler == true`. so perhaps?
> Deprecated. Will eventually be phased out by the batched segment sampler. You can enable the batched segment sampler now by setting the dynamic Coordinator config, `useBatchedSegmentSampler`, to `true`. Note that if you choose enable the batched segment sampler, this config has no effect on balancing. If `useBatchedSegmentSampler == false`, this config defines the percentage of the total number of segments in the cluster that are considered every time a segment needs to be selected for a move. Druid orders servers by available capacity ascending (the least available capacity first) and then iterates over the servers. For each server, Druid iterates over the segments on the server, considering them for moving. The default config of 100% means that every segment on every server is a candidate to be moved. This should make sense for most small to medium-sized clusters. However, an admin may find it preferable to drop this value lower if they don't think that it is worthwhile to con
sider every single segment in the cluster each time it is looking for a segment to move.|100|
##########
File path: web-console/src/druid-models/coordinator-dynamic-config.tsx
##########
@@ -213,14 +213,16 @@ export const COORDINATOR_DYNAMIC_CONFIG_FIELDS: Field<CoordinatorDynamicConfig>[
defaultValue: 100,
info: (
<>
- The percentage of the total number of segments in the cluster that are considered every time
- a segment needs to be selected for a move. Druid orders servers by available capacity
- ascending (the least available capacity first) and then iterates over the servers. For each
- server, Druid iterates over the segments on the server, considering them for moving. The
- default config of 100% means that every segment on every server is a candidate to be moved.
- This should make sense for most small to medium-sized clusters. However, an admin may find
- it preferable to drop this value lower if they don't think that it is worthwhile to
- consider every single segment in the cluster each time it is looking for a segment to move.
+ Deprecated. This will be phased out by the batched segment sampler. Only used if
Review comment:
could you clarify what you mean by this? I take it as you asking me to state that the preferred alternative is to set `useBatchedSegmentSampler == true`. so perhaps:
> Deprecated. Will eventually be phased out by the batched segment sampler. You can enable the batched segment sampler now by setting the dynamic Coordinator config, `useBatchedSegmentSampler`, to `true`. Note that if you choose enable the batched segment sampler, this config has no effect on balancing. If `useBatchedSegmentSampler == false`, this config defines the percentage of the total number of segments in the cluster that are considered every time a segment needs to be selected for a move. Druid orders servers by available capacity ascending (the least available capacity first) and then iterates over the servers. For each server, Druid iterates over the segments on the server, considering them for moving. The default config of 100% means that every segment on every server is a candidate to be moved. This should make sense for most small to medium-sized clusters. However, an admin may find it preferable to drop this value lower if they don't think that it is worthwhile to con
sider every single segment in the cluster each time it is looking for a segment to move.|100|
--
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.
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For additional commands, e-mail: commits-help@druid.apache.org
[GitHub] [druid] jihoonson commented on pull request #11960: clean up the balancing code around the batched vs deprecated way of sampling segments to balance
Posted by GitBox <gi...@apache.org>.
jihoonson commented on pull request #11960:
URL: https://github.com/apache/druid/pull/11960#issuecomment-988206282
@capistrant thank you for updating docs. Feel free to merge once the "docs" CI passes.
--
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.
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For additional commands, e-mail: commits-help@druid.apache.org
[GitHub] [druid] a2l007 commented on a change in pull request #11960: clean up the balancing code around the batched vs deprecated way of sampling segments to balance
Posted by GitBox <gi...@apache.org>.
a2l007 commented on a change in pull request #11960:
URL: https://github.com/apache/druid/pull/11960#discussion_r755332953
##########
File path: docs/configuration/index.md
##########
@@ -886,7 +886,7 @@ Issuing a GET request at the same URL will return the spec that is currently in
|`mergeSegmentsLimit`|The maximum number of segments that can be in a single [append task](../ingestion/tasks.md).|100|
|`maxSegmentsToMove`|The maximum number of segments that can be moved at any given time.|5|
|`useBatchedSegmentSampler`|Boolean flag for whether or not we should use the Reservoir Sampling with a reservoir of size k instead of fixed size 1 to pick segments to move. This option can be enabled to speed up segment balancing process, especially if there are huge number of segments in the cluster or if there are too many segments to move.|false|
-|`percentOfSegmentsToConsiderPerMove`|The percentage of the total number of segments in the cluster that are considered every time a segment needs to be selected for a move. Druid orders servers by available capacity ascending (the least available capacity first) and then iterates over the servers. For each server, Druid iterates over the segments on the server, considering them for moving. The default config of 100% means that every segment on every server is a candidate to be moved. This should make sense for most small to medium-sized clusters. However, an admin may find it preferable to drop this value lower if they don't think that it is worthwhile to consider every single segment in the cluster each time it is looking for a segment to move.|100|
+|`percentOfSegmentsToConsiderPerMove`|Deprecated. Will eventually be phased out by the batched segment sampler. Only used if `useBatchedSegmentSampler == false`. The percentage of the total number of segments in the cluster that are considered every time a segment needs to be selected for a move. Druid orders servers by available capacity ascending (the least available capacity first) and then iterates over the servers. For each server, Druid iterates over the segments on the server, considering them for moving. The default config of 100% means that every segment on every server is a candidate to be moved. This should make sense for most small to medium-sized clusters. However, an admin may find it preferable to drop this value lower if they don't think that it is worthwhile to consider every single segment in the cluster each time it is looking for a segment to move.|100|
Review comment:
Should we add a similar deprecation message in the web console description for this property?
--
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.
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For additional commands, e-mail: commits-help@druid.apache.org
[GitHub] [druid] jihoonson commented on a change in pull request #11960: clean up the balancing code around the batched vs deprecated way of sampling segments to balance
Posted by GitBox <gi...@apache.org>.
jihoonson commented on a change in pull request #11960:
URL: https://github.com/apache/druid/pull/11960#discussion_r764219324
##########
File path: web-console/src/druid-models/coordinator-dynamic-config.tsx
##########
@@ -213,14 +213,16 @@ export const COORDINATOR_DYNAMIC_CONFIG_FIELDS: Field<CoordinatorDynamicConfig>[
defaultValue: 100,
info: (
<>
- The percentage of the total number of segments in the cluster that are considered every time
- a segment needs to be selected for a move. Druid orders servers by available capacity
- ascending (the least available capacity first) and then iterates over the servers. For each
- server, Druid iterates over the segments on the server, considering them for moving. The
- default config of 100% means that every segment on every server is a candidate to be moved.
- This should make sense for most small to medium-sized clusters. However, an admin may find
- it preferable to drop this value lower if they don't think that it is worthwhile to
- consider every single segment in the cluster each time it is looking for a segment to move.
+ Deprecated. This will be phased out by the batched segment sampler. Only used if
Review comment:
Can we document what the preferred alternative is here as well?
##########
File path: docs/configuration/index.md
##########
@@ -886,7 +886,7 @@ Issuing a GET request at the same URL will return the spec that is currently in
|`mergeSegmentsLimit`|The maximum number of segments that can be in a single [append task](../ingestion/tasks.md).|100|
|`maxSegmentsToMove`|The maximum number of segments that can be moved at any given time.|5|
|`useBatchedSegmentSampler`|Boolean flag for whether or not we should use the Reservoir Sampling with a reservoir of size k instead of fixed size 1 to pick segments to move. This option can be enabled to speed up segment balancing process, especially if there are huge number of segments in the cluster or if there are too many segments to move.|false|
-|`percentOfSegmentsToConsiderPerMove`|The percentage of the total number of segments in the cluster that are considered every time a segment needs to be selected for a move. Druid orders servers by available capacity ascending (the least available capacity first) and then iterates over the servers. For each server, Druid iterates over the segments on the server, considering them for moving. The default config of 100% means that every segment on every server is a candidate to be moved. This should make sense for most small to medium-sized clusters. However, an admin may find it preferable to drop this value lower if they don't think that it is worthwhile to consider every single segment in the cluster each time it is looking for a segment to move.|100|
+|`percentOfSegmentsToConsiderPerMove`|Deprecated. Will eventually be phased out by the batched segment sampler. Only used if `useBatchedSegmentSampler == false`. The percentage of the total number of segments in the cluster that are considered every time a segment needs to be selected for a move. Druid orders servers by available capacity ascending (the least available capacity first) and then iterates over the servers. For each server, Druid iterates over the segments on the server, considering them for moving. The default config of 100% means that every segment on every server is a candidate to be moved. This should make sense for most small to medium-sized clusters. However, an admin may find it preferable to drop this value lower if they don't think that it is worthwhile to consider every single segment in the cluster each time it is looking for a segment to move.|100|
Review comment:
Can we document what the preferred alternative is here?
--
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.
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For additional commands, e-mail: commits-help@druid.apache.org
[GitHub] [druid] capistrant commented on a change in pull request #11960: clean up the balancing code around the batched vs deprecated way of sampling segments to balance
Posted by GitBox <gi...@apache.org>.
capistrant commented on a change in pull request #11960:
URL: https://github.com/apache/druid/pull/11960#discussion_r756042237
##########
File path: docs/configuration/index.md
##########
@@ -886,7 +886,7 @@ Issuing a GET request at the same URL will return the spec that is currently in
|`mergeSegmentsLimit`|The maximum number of segments that can be in a single [append task](../ingestion/tasks.md).|100|
|`maxSegmentsToMove`|The maximum number of segments that can be moved at any given time.|5|
|`useBatchedSegmentSampler`|Boolean flag for whether or not we should use the Reservoir Sampling with a reservoir of size k instead of fixed size 1 to pick segments to move. This option can be enabled to speed up segment balancing process, especially if there are huge number of segments in the cluster or if there are too many segments to move.|false|
-|`percentOfSegmentsToConsiderPerMove`|The percentage of the total number of segments in the cluster that are considered every time a segment needs to be selected for a move. Druid orders servers by available capacity ascending (the least available capacity first) and then iterates over the servers. For each server, Druid iterates over the segments on the server, considering them for moving. The default config of 100% means that every segment on every server is a candidate to be moved. This should make sense for most small to medium-sized clusters. However, an admin may find it preferable to drop this value lower if they don't think that it is worthwhile to consider every single segment in the cluster each time it is looking for a segment to move.|100|
+|`percentOfSegmentsToConsiderPerMove`|Deprecated. Will eventually be phased out by the batched segment sampler. Only used if `useBatchedSegmentSampler == false`. The percentage of the total number of segments in the cluster that are considered every time a segment needs to be selected for a move. Druid orders servers by available capacity ascending (the least available capacity first) and then iterates over the servers. For each server, Druid iterates over the segments on the server, considering them for moving. The default config of 100% means that every segment on every server is a candidate to be moved. This should make sense for most small to medium-sized clusters. However, an admin may find it preferable to drop this value lower if they don't think that it is worthwhile to consider every single segment in the cluster each time it is looking for a segment to move.|100|
Review comment:
yep, good call @a2l007
--
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.
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For additional commands, e-mail: commits-help@druid.apache.org
[GitHub] [druid] yuanlihan commented on pull request #11960: clean up the balancing code around the batched vs deprecated way of sampling segments to balance
Posted by GitBox <gi...@apache.org>.
yuanlihan commented on pull request #11960:
URL: https://github.com/apache/druid/pull/11960#issuecomment-975285154
> @yuanlihan could you take a look at this since you wrote the original code for the batched sampler! I hope my code makes sense, but please comment on any mistakes I may have made!
Hi @capistrant
It looks good to me. Thanks for making the code neat!
--
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.
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For additional commands, e-mail: commits-help@druid.apache.org