YARN-11235. Refactor Policy Code and Define getReservationHomeSubcluster #4656
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Contributor
Author
|
@goiri This is the first pr after YARN-5871 Split. Please help to review the code, thank you very much! |
|
🎊 +1 overall
This message was automatically generated. |
goiri
reviewed
Jul 28, 2022
| public SubClusterId getReservationHomeSubcluster(ReservationSubmissionRequest request) | ||
| throws YarnException { | ||
| if (request == null) { | ||
| throw new FederationPolicyException( |
| nodeRequest = rr; | ||
| } catch (YarnException e) { | ||
| LOG.error("Cannot resolve node : {}", e.getLocalizedMessage()); | ||
| LOG.error("Cannot resolve node.", e); |
Member
There was a problem hiding this comment.
I prefer not having the full exception. Thoughts?
Contributor
Author
There was a problem hiding this comment.
Thanks for your suggestion, I will modify the code!
| } | ||
| protected SubClusterId chooseSubCluster( | ||
| String queue, Map<SubClusterId, SubClusterInfo> preSelectSubClusters) throws YarnException { | ||
| if(preSelectSubClusters == null || preSelectSubClusters.size() == 0) { |
| // changes dynamically (and this would unfairly spread the load to | ||
| // sub-clusters adjacent to an inactive one), hence we need to count/scan | ||
| // the list and based on weight pick the next sub-cluster. | ||
| Map<SubClusterIdInfo, Float> weights = |
Contributor
Author
There was a problem hiding this comment.
I will modify the code.
| } | ||
| if (entry.getKey() != null | ||
| && activeSubclusters.containsKey(entry.getKey().toId())) { | ||
| if (entry.getKey() != null && preSelectSubClusters.containsKey(entry.getKey().toId())) { |
Contributor
Author
There was a problem hiding this comment.
Thank you for your suggestion.
| if (pickedIndex == -1) { | ||
| throw new FederationPolicyException( | ||
| "No positive weight found on active subclusters"); | ||
| "No positive weight found on active subClusters."); |
Member
There was a problem hiding this comment.
i would leave the old capitalization
| import org.apache.hadoop.yarn.server.federation.store.records.SubClusterState; | ||
| import org.apache.hadoop.yarn.server.federation.store.FederationStateStore; | ||
| import org.apache.hadoop.yarn.server.federation.store.impl.MemoryFederationStateStore; | ||
| import org.apache.hadoop.yarn.server.federation.store.records.*; |
|
🎊 +1 overall
This message was automatically generated. |
Contributor
Author
|
@goiri Please help to review the code again, thank you very much! |
goiri
reviewed
Jul 29, 2022
| try { | ||
| configuration = federationFacade.getPolicyConfiguration(copyQueue); | ||
| } catch (YarnException e) { | ||
| LOG.warn("There is no policy configured for the queue: {}, " + |
Member
There was a problem hiding this comment.
We can fit the string into a single line.
Contributor
Author
There was a problem hiding this comment.
Thanks for your help reviewing the code, I will modify the code.
| // cached) | ||
| if (configuration == null) { | ||
| final String policyKey = YarnConfiguration.DEFAULT_FEDERATION_POLICY_KEY; | ||
| LOG.warn("There is no policies configured for queue: {} " + |
Member
There was a problem hiding this comment.
We can rearrange so the string is a single line.
|
|
||
| // if the configuration has changed since last loaded, reinit the policy | ||
| // based on current configuration | ||
| if (!cachedConfiguration.containsKey(copyQueue) |
| // if the configuration has changed since last loaded, reinit the policy | ||
| // based on current configuration | ||
| if (!cachedConfiguration.containsKey(copyQueue) | ||
| || !cachedConfiguration.get(copyQueue).equals(configuration)) { |
Member
There was a problem hiding this comment.
I would extract to make it more readable.
| nodeRequest = rr; | ||
| } catch (YarnException e) { | ||
| LOG.error("Cannot resolve node : {}", e.getLocalizedMessage()); | ||
| LOG.error("Cannot resolve node."); |
Member
There was a problem hiding this comment.
Doesn't it make sense to keep output the message at least?
Contributor
Author
There was a problem hiding this comment.
I will fix it and use e.getMessage() .
| } catch (YarnException e) { | ||
| LOG.error("Validating resource requests failed, Falling back to " | ||
| + "WeightedRandomRouterPolicy placement: " + e.getMessage()); | ||
| + "WeightedRandomRouterPolicy placement.", e); |
| /** | ||
| * This implements a policy that interprets "weights" as a ordered list of | ||
| * preferences among sub-clusters. Highest weight among active subclusters is | ||
| * preferences among sub-clusters. Highest weight among active subClusters is |
Member
There was a problem hiding this comment.
We keep changing the capitalization.
I think "Subcluster" is fine; no need to capitalize the C.
|
💔 -1 overall
This message was automatically generated. |
|
🎊 +1 overall
This message was automatically generated. |
Contributor
Author
|
@goiri Please help to review the code again, thank you very much! |
goiri
reviewed
Aug 1, 2022
| Map<SubClusterId, SubClusterInfo> filteredSubClusters = prefilterSubClusters( | ||
| appContext.getReservationID(), getActiveSubclusters()); | ||
|
|
||
| FederationPolicyUtils.validateSubClusterAvailability( |
Member
There was a problem hiding this comment.
Can we have validateSubClusterAvailability taking Collection so we don't need temp data structures?
| FederationStateStoreFacade fedFacade = | ||
| FederationStateStoreFacade.getInstance(); | ||
| YarnConfiguration conf = new YarnConfiguration(); | ||
| conf.set(YarnConfiguration.FEDERATION_CACHE_TIME_TO_LIVE_SECS, "0"); |
| public FederationStateStoreFacade getMemoryFacade() throws YarnException { | ||
|
|
||
| // setting up a store and its facade (with caching off) | ||
| FederationStateStoreFacade fedFacade = |
| import org.junit.Assert; | ||
| import org.junit.Test; | ||
|
|
||
|
|
|
|
||
| @Test | ||
| public void testNullReservationContext() throws Exception { | ||
| FederationRouterPolicy policy = ((FederationRouterPolicy) getPolicy()); |
| Map<SubClusterIdInfo, Float> routerWeights = new HashMap<>(); | ||
| Map<SubClusterIdInfo, Float> amrmWeights = new HashMap<>(); | ||
|
|
||
| long now = System.currentTimeMillis(); |
| .getHomeSubcluster(getApplicationSubmissionContext(), null); | ||
| fail(); | ||
| } catch (YarnException ex) { | ||
| Assert.assertTrue( |
Contributor
Author
There was a problem hiding this comment.
Thanks for your suggestion, I will modify the code.
| when(sci.getState()).thenReturn(SubClusterState.SC_RUNNING); | ||
| when(sci.getSubClusterId()).thenReturn(sc.toId()); | ||
| getActiveSubclusters().put(sc.toId(), sci); | ||
| long now = System.currentTimeMillis(); |
...src/test/java/org/apache/hadoop/yarn/server/federation/utils/FederationPoliciesTestUtil.java
Outdated
Show resolved
Hide resolved
|
🎊 +1 overall
This message was automatically generated. |
Contributor
Author
|
@goiri Please help to review the code again, thank you very much! I will follow up with YARN-11236. |
|
💔 -1 overall
This message was automatically generated. |
|
🎊 +1 overall
This message was automatically generated. |
goiri
reviewed
Aug 2, 2022
.../java/org/apache/hadoop/yarn/server/federation/store/records/GetSubClustersInfoResponse.java
Show resolved
Hide resolved
| public static GetSubClustersInfoResponse newInstance( | ||
| Collection<SubClusterInfo> subClusters) { | ||
| GetSubClustersInfoResponse subClusterInfos = | ||
| Records.newRecord(GetSubClustersInfoResponse.class); |
Member
There was a problem hiding this comment.
Indentation doesn't seem correct.
I think it even fits in one line.
.../test/java/org/apache/hadoop/yarn/server/federation/policies/BaseFederationPoliciesTest.java
Show resolved
Hide resolved
| when(sci.getSubClusterId()).thenReturn(sc.toId()); | ||
| getActiveSubclusters().put(sc.toId(), sci); | ||
| long now = Time.now(); | ||
| SubClusterInfo federationSubClusterInfo = |
Member
There was a problem hiding this comment.
Not sure the indetation is correct.
Let's do:
SubClusterInfo federationSubClusterInfo = SubClusterInfo.newInstance(
sc.toId(), "dns1:80", "dns1:81", "dns1:82", "dns1:83",
now - 1000, SubClusterState.SC_RUNNING, now - 2000, generateClusterMetricsInfo(i));
| when(sci.getSubClusterId()).thenReturn(sc.toId()); | ||
| getActiveSubclusters().put(sc.toId(), sci); | ||
| long now = Time.now(); | ||
| SubClusterInfo federationSubClusterInfo = |
| if (activeSubClusters == null) { | ||
| activeSubClusters = new HashMap<>(); | ||
| } | ||
| GetSubClustersInfoResponse response = |
|
💔 -1 overall
This message was automatically generated. |
|
🎊 +1 overall
This message was automatically generated. |
|
🎊 +1 overall
This message was automatically generated. |
Contributor
Author
|
@goiri Please help to review the code again, thank you very much! |
added 2 commits
August 3, 2022 09:24
This reverts commit 874c9c5.
|
🎊 +1 overall
This message was automatically generated. |
goiri
approved these changes
Aug 3, 2022
goiri
reviewed
Aug 3, 2022
|
|
||
| FederationPolicyUtils.validateSubClusterAvailability( | ||
| list, blackListSubClusters); | ||
| FederationPolicyUtils.validateSubClusterAvailability(list, blackListSubClusters); |
Member
There was a problem hiding this comment.
We could just do:
FederationPolicyUtils.validateSubClusterAvailability(activeSubclusters.keySet(), blackListSubClusters);
Contributor
Author
There was a problem hiding this comment.
Thanks for your suggestion, I will modify the code.
Contributor
Author
|
@goiri Please help to review the code again, thank you very much! |
|
🎊 +1 overall
This message was automatically generated. |
goiri
approved these changes
Aug 4, 2022
Contributor
Author
|
@goiri Thank you for helping to review the code, I will follow up with YARN-11236 after this pr is completed. |
HarshitGupta11
pushed a commit
to HarshitGupta11/hadoop
that referenced
this pull request
Nov 28, 2022
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
JIRA:YARN-11235. Refactor Policy Code and Define getReservationHomeSubcluster.