add getworkermetadata streaming rpc #27767

m-trieu · 2023-08-01T03:58:59Z

Adding a stream for GetWorkerMetadata

Does not include (the following will be in a seperate CL)

Storing the metadata
Using the metadata
Adding stream to GrpcWindmillServer (I will add tests to GRPCWindmillServer as well)

Updated the type parameters for the *ResponseObserver.java files
Add unit tests (should do this for all of the stream classes)
make GetWorkerMetadata RPC a BiDi Stream

r: @scwhittle @dustin12

Thank you for your contribution! Follow this checklist to help us incorporate your contribution quickly and easily:

Mention the appropriate issue in your description (for example: addresses #123), if applicable. This will automatically add a link to the pull request in the issue. If you would like the issue to automatically close on merging the pull request, comment fixes #<ISSUE NUMBER> instead.
Update CHANGES.md with noteworthy changes.
If this contribution is large, please file an Apache Individual Contributor License Agreement.

See the Contributor Guide for more tips on how to make review process smoother.

To check the build health, please visit https://github.com/apache/beam/blob/master/.test-infra/BUILD_STATUS.md

GitHub Actions Tests Status (on master branch)

See CI.md for more information about GitHub Actions CI or the workflows README to see a list of phrases to trigger workflows.

...org/apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcGetWorkerMetadataStream.java

github-actions · 2023-08-01T05:34:46Z

Checks are failing. Will not request review until checks are succeeding. If you'd like to override that behavior, comment assign set of reviewers

.../java/org/apache/beam/runners/dataflow/worker/windmill/ForwardingClientResponseObserver.java

scwhittle · 2023-08-02T10:46:21Z

...r/src/main/java/org/apache/beam/runners/dataflow/worker/windmill/AbstractWindmillStream.java

  private final AtomicLong startTimeMs;
-  private final AtomicLong lastSendTimeMs;
+  protected final AtomicLong lastSendTimeMs;


can this be reverted? don't see it accessed by subclass

scwhittle · 2023-08-02T11:11:13Z

...r/src/main/java/org/apache/beam/runners/dataflow/worker/windmill/AbstractWindmillStream.java

@@ -73,9 +73,9 @@ public abstract class AbstractWindmillStream<RequestT, ResponseT> implements Win
  private final Executor executor;
  private final BackOff backoff;
  // Indicates if the current stream in requestObserver is closed by calling close() method
-  private final AtomicBoolean streamClosed;
+  protected AtomicBoolean streamClosed;


can this be reverted?
It seems at least it should be final, but also setting it as part of stream callback doesn't match the comment

scwhittle · 2023-08-02T11:31:05Z

...org/apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcGetWorkerMetadataStream.java

+  }
+
+  /**
+   * GetWorkerMetadata is a server streaming API. There is no request observer since the client side


It seems that a lot of the changes here are necessary because this is not a bi-directional stream.

Beyond the code complexity of supporting this, I'm concerned about us not being able to send application-level heartbeat requests as we've had numerous issues in the past where sending was necessary to detect that the stream had died. See https://issues.apache.org/jira/browse/BEAM-10808 and #12688.

I'm thinking that we should change the metadata streaming rpc to be bi-directional instead of just server side streaming to make sure we don't hit similar issues.

...apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcGetWorkerMetadataStreamTest.java

...org/apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcGetWorkerMetadataStream.java

scwhittle · 2023-08-10T11:07:46Z

...r/src/main/java/org/apache/beam/runners/dataflow/worker/windmill/AbstractWindmillStream.java

@@ -130,16 +127,18 @@ private static long debugDuration(long nowMs, long startMs) {
  protected abstract void onNewStream();

  /** Returns whether there are any pending requests that should be retried on a stream break. */
-  protected abstract boolean hasPendingRequests();
+  protected boolean hasPendingRequests() {


revert? Since this is related to correctness a default seems dangerous since it makes it possible to skip implementing correctly.

scwhittle · 2023-08-10T11:10:04Z

...worker/src/main/java/org/apache/beam/runners/dataflow/worker/windmill/WindmillEndpoints.java

+
+@AutoValue
+public abstract class WindmillEndpoints {
+  public abstract ImmutableMap<String, String> globalDataServers();


Comment on what key and value are

scwhittle · 2023-08-10T11:10:26Z

...worker/src/main/java/org/apache/beam/runners/dataflow/worker/windmill/WindmillEndpoints.java

+public abstract class WindmillEndpoints {
+  public abstract ImmutableMap<String, String> globalDataServers();
+
+  public abstract ImmutableList<String> windmillServers();


add a comment about what string is

alternatively should it be some sort of endpoint object instead of a string?

i guess we can convert the windmill server string to an object let me find the format

scwhittle · 2023-08-10T11:12:29Z

...n/java/org/apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcCommitWorkStream.java

@@ -56,7 +56,8 @@ final class GrpcCommitWorkStream
  private final int streamingRpcBatchLimit;

  private GrpcCommitWorkStream(
-      CloudWindmillServiceV1Alpha1Grpc.CloudWindmillServiceV1Alpha1Stub stub,
+      Function<StreamObserver<StreamingCommitResponse>, StreamObserver<StreamingCommitWorkRequest>>
+          commitWorkRpc,


nit: commitWorkRpc is a confusing name to me for a function. What about createCommitWorkRpcFn or startCommitWorkRpcFn etc?

ditto for the other files

scwhittle · 2023-08-10T11:19:38Z

...org/apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcGetWorkerMetadataStream.java

+  @Override
+  protected void onResponse(WorkerMetadataResponse response) {
+    metadataVersion.getAndUpdate(
+        current -> updateMetadataVersion(current, response.getMetadataVersion()));


if we get a version before our current version this seems to leave metadataVersion at the more current version but then it goes ahead and returns the stale response instead of filtering it.

done. good catch! tried to use the Atomic* non blocking to do the filtering.

scwhittle · 2023-08-10T11:21:17Z

...ain/java/org/apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcWindmillServer.java

+    return stubList
+        .get(rand.nextInt(stubList.size()))
+        .withDeadlineAfter(
+            AbstractWindmillStream.DEFAULT_STREAM_RPC_DEADLINE_SECONDS, TimeUnit.SECONDS);


this deadline is not being applied to stub returned above if only 1 stub

...apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcGetWorkerMetadataStreamTest.java

scwhittle · 2023-08-10T11:33:27Z

...apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcGetWorkerMetadataStreamTest.java

+        "stale_global_data",
+        WorkerMetadataResponse.Endpoint.newBuilder().setEndpoint("staleGlobalData").build());
+
+    testStub.injectWorkerMetadata(


see comment in code. It seems like this test should fail with the current logic. It would also pass if this injection is somehow not reaching the stream so maybe some wiring is wrong?

it failed (i didnt rerun it after i removed the locking)
passes now that the logic is fixed

github-actions · 2023-08-11T06:34:55Z

Assigning reviewers. If you would like to opt out of this review, comment assign to next reviewer:

R: @lostluck added as fallback since no labels match configuration

Available commands:

stop reviewer notifications - opt out of the automated review tooling
remind me after tests pass - tag the comment author after tests pass
waiting on author - shift the attention set back to the author (any comment or push by the author will return the attention set to the reviewers)

The PR bot will only process comments in the main thread (not review comments).

scwhittle · 2023-08-11T10:42:23Z

...r/src/main/java/org/apache/beam/runners/dataflow/worker/windmill/AbstractWindmillStream.java

@@ -149,7 +146,7 @@ private StreamObserver<RequestT> requestObserver() {
  }

  /** Send a request to the server. */
-  protected final void send(RequestT request) {
+  protected void send(RequestT request) {


can this remain final? don't see new override

missed it in the revert done.

scwhittle · 2023-08-11T10:43:19Z

...r/src/main/java/org/apache/beam/runners/dataflow/worker/windmill/AbstractWindmillStream.java

   */
  protected abstract void startThrottleTimer();

-  private StreamObserver<RequestT> requestObserver() {
+  protected StreamObserver<RequestT> requestObserver() {


can this remain private? don't see new call site

scwhittle · 2023-08-11T10:44:49Z

...r/src/main/java/org/apache/beam/runners/dataflow/worker/windmill/AbstractWindmillStream.java

@@ -252,7 +249,7 @@ public final Instant startTime() {
    return new Instant(startTimeMs.get());
  }

-  private class ResponseObserver implements StreamObserver<ResponseT> {
+  public class ResponseObserver implements StreamObserver<ResponseT> {


ditto, seems like this could remain private?

m-trieu · 2023-08-11T21:26:21Z

waiting for changes to be finalized in windmill api proto before continuing to work on this @scwhittle

m-trieu · 2023-08-15T18:08:50Z

changes to add worker token support are independent, this review can proceed
thanks! @scwhittle

scwhittle · 2023-08-16T08:46:29Z

.../java/org/apache/beam/runners/dataflow/worker/windmill/ForwardingClientResponseObserver.java

 */
-final class ForwardingClientResponseObserver<ReqT, RespT>
-    implements ClientResponseObserver<RespT, ReqT> {
+final class ForwardingClientResponseObserver<ResponseT, RequestT>


Can you rebase? I think a lot of this has already been merged with your previous PR

even after rebase this is showing up hmmm when i look at master https://github.com/apache/beam/blame/master/runners/google-cloud-dataflow-java/worker/src/main/java/org/apache/beam/runners/dataflow/worker/windmill/ForwardingClientResponseObserver.java it still shows it as the old code?

scwhittle · 2023-08-16T08:48:50Z

...worker/src/main/java/org/apache/beam/runners/dataflow/worker/windmill/WindmillEndpoints.java

+   */
+  public abstract ImmutableList<Endpoint> windmillServers();
+
+  @AutoValue


just checking that Autovalue automatically creates equals/hash methods

confirmed it does https://chromium.googlesource.com/external/github.com/google/auto/+/auto-value-1.0/value/README.md#background

scwhittle · 2023-08-16T08:52:10Z

...org/apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcGetWorkerMetadataStream.java

+      Set<AbstractWindmillStream<?, ?>> streamRegistry,
+      int logEveryNStreamFailures,
+      JobHeader jobHeader,
+      AtomicLong metadataVersion,


integer? Can keep the AtomicLong internal for non-testing case to ensure it isn't modified externally

still AtomicLong here

whoops misread the comment as for test only. done.

scwhittle · 2023-08-16T09:08:26Z

...apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcGetWorkerMetadataStreamTest.java

+    TestGetWorkMetadataRequestObserver requestObserver =
+        new TestGetWorkMetadataRequestObserver(mockResponse);
+    GetWorkerMetadataTestStub testStub = new GetWorkerMetadataTestStub(requestObserver);
+    AtomicLong metadataVersion = new AtomicLong(0);


nit: use some non-0 version, 0 might be correct if just default initialized on accident instead of propagated for example.

(sorry if some of these are duplicated, I was putting comments on some other snapshot I think, not sure where they'll show up).

done. no worries!

scwhittle · 2023-08-16T09:08:43Z

...apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcGetWorkerMetadataStreamTest.java

+
+    @Override
+    public void onNext(WorkerMetadataRequest workerMetadataRequest) {
+      responseObserver.onNext(response);


what about removing this way to send responses and just having tests explicitly call injectWorkerMetadata?
I think it will be easier to see what tests are doing

scwhittle · 2023-08-16T09:09:14Z

runners/google-cloud-dataflow-java/worker/windmill/src/main/proto/windmill.proto

@@ -758,7 +758,8 @@ message WorkerMetadataResponse {
  // CommitWorkStream. Each response on this stream replaces the previous, and
  // connections to endpoints that are no longer present should be closed.
  message Endpoint {
-    optional string endpoint = 1;
+    optional string direct_endpoint = 1;
+    optional string worker_token = 2;


I think we want to remove direct_path_endpoints and just use work_endpoints

m-trieu · 2023-08-25T23:27:14Z

@scwhittle do you have any more comments for this pull request? Thank you!

scwhittle · 2023-08-28T09:40:07Z

...org/apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcGetWorkerMetadataStream.java

+  private final AtomicLong metadataVersion;
+  private final WorkerMetadataRequest workerMetadataRequest;
+  private final ThrottleTimer getWorkerMetadataThrottleTimer;
+  private final Consumer<WindmillEndpoints> serverMappingUpdater;


nit: serverMappingConsumer or serverMappingObserver?
Updater makes it sound a little like it is the other direction to me

scwhittle · 2023-08-28T09:40:24Z

...org/apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcGetWorkerMetadataStream.java

+      Set<AbstractWindmillStream<?, ?>> streamRegistry,
+      int logEveryNStreamFailures,
+      JobHeader jobHeader,
+      AtomicLong metadataVersion,


still AtomicLong here

scwhittle · 2023-08-28T09:43:12Z

...org/apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcGetWorkerMetadataStream.java

+    long currentMetadataVersion = metadataVersion.get();
+    long updatedMetadataVersion =
+        metadataVersion.updateAndGet(
+            current -> updateMetadataVersion(current, response.getMetadataVersion()));


how about inlining the non-logging part of updateMetadataVersion here, and moving the logging below to else? Seems clearer than logging in the separate method.

scwhittle · 2023-08-28T09:44:48Z

...org/apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcGetWorkerMetadataStream.java

+            current -> updateMetadataVersion(current, response.getMetadataVersion()));
+
+    if (updatedMetadataVersion > currentMetadataVersion) {
+      serverMappingUpdater.accept(WindmillEndpoints.from(response));


Do we need the atomic? I think that there is only one stream calling onResponse and thus those should be ordered. If we did have concurrent callers, this could be racy because you could have
current version is 0
T1: atomically updates to 1
T2: atomically updates to 2, calls accept
T1: resumes and calls accept
then the consumer has last observed the payload for 1.

So either:

remove the atomic if it isn't needed and document the thread-safety

or ensure that the ordering is maintained (ie could use synchronized and update version and call accept under it)

scwhittle · 2023-08-28T09:46:38Z

...org/apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcGetWorkerMetadataStream.java

+  @Override
+  protected void appendSpecificHtml(PrintWriter writer) {
+    writer.format(
+        "GetWorkerMetadataStream: version=[%d] , job_header=[%s]",


could show the latest response? could be useful.

scwhittle · 2023-08-28T09:47:42Z

...ain/java/org/apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcWindmillServer.java

+        (stubList.size() == 1 ? stubList.get(0) : stubList.get(rand.nextInt(stubList.size())));
+
+    return stub.withDeadlineAfter(
+        AbstractWindmillStream.DEFAULT_STREAM_RPC_DEADLINE_SECONDS, TimeUnit.SECONDS);


see other cl about withDeadlineAfter

either use some other mechanism or document that result of stub() should just be used and not cached.

could name stubWithAbsoluteDeadlineSet to make it obvious to callers instead of a comment.

done, decided to keep it the same and change the name for simplicity

Looking again, the AbstractWindmillStream are used for longer than the deadline, they reconnect grpc streams internally. So I think you need to use the observer approach and not a fixed deadline.

scwhittle · 2023-08-28T09:50:23Z

...worker/src/main/java/org/apache/beam/runners/dataflow/worker/windmill/WindmillEndpoints.java

+      return endpointBuilder.build();
+    }
+
+    public abstract Optional<String> directEndpoint();


is there a better way to represent ipv6 address than String?

lets use Inet6Address, i will update

scwhittle · 2023-09-01T08:47:30Z

It looks like you added a merge commit, instead could you rebase your change?

scwhittle · 2023-09-01T08:51:18Z

...worker/src/main/java/org/apache/beam/runners/dataflow/worker/windmill/WindmillEndpoints.java

+
+      return directEndpointAddress == null
+          ? Optional.empty()
+          : Optional.of((Inet6Address) directEndpointAddress);


this cast seems wrong if the above instanceof fails.
seems we should return empty in that case

since null will be false for instanceof you could just do a single check using that.

scwhittle · 2023-09-01T08:59:08Z

...org/apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcGetWorkerMetadataStream.java

+          metadataVersion);
+    }
+
+    this.latestResponse = response;


don't update lastResponse in the case it is stale

scwhittle · 2023-09-01T08:59:59Z

...org/apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcGetWorkerMetadataStream.java

+  protected void onResponse(WorkerMetadataResponse response) {
+    if (response.getMetadataVersion() > metadataVersion) {
+      metadataVersion = response.getMetadataVersion();
+      serverMappingConsumer.accept(WindmillEndpoints.from(response));


if you add sychronization for html rendering (see below), it would be good to keep the accept call out of the synchronized block.

From AbstractStream header it sounds like you should not synchronize on this but some other object:

Synchronization on this is used to synchronize the gRpc stream state and internal data

structures. Since grpc channel operations may block, synchronization on this stream may also

block. This is generally not a problem since streams are used in a single-threaded manner.

However, some accessors used for status page and other debugging need to take care not to require

synchronizing on this.

i opted to use StampedLock https://docs.oracle.com/javase/8/docs/api/java/util/concurrent/locks/StampedLock.html cause it supports optimistic locking and allows for multiple readers.

scwhittle · 2023-09-01T09:00:02Z

...org/apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcGetWorkerMetadataStream.java

+  protected void appendSpecificHtml(PrintWriter writer) {
+    writer.format(
+        "GetWorkerMetadataStream: version=[%d] , job_header=[%s], latest_response=[%s]",
+        metadataVersion, workerMetadataRequest.getHeader(), latestResponse);


I think you need some synchronization for latestResponse etc since I believe the html rendering can be concurrent

done. added locking

scwhittle · 2023-09-01T09:03:53Z

...ain/java/org/apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcWindmillServer.java

+        (stubList.size() == 1 ? stubList.get(0) : stubList.get(rand.nextInt(stubList.size())));
+
+    return stub.withDeadlineAfter(
+        AbstractWindmillStream.DEFAULT_STREAM_RPC_DEADLINE_SECONDS, TimeUnit.SECONDS);


Looking again, the AbstractWindmillStream are used for longer than the deadline, they reconnect grpc streams internally. So I think you need to use the observer approach and not a fixed deadline.

scwhittle · 2023-09-01T09:04:27Z

runners/google-cloud-dataflow-java/worker/windmill/src/main/proto/windmill.proto

-  // DirectPath endpoints to be used by user workers for streaming engine jobs.
-  // DirectPath endpoints here are virtual IPv6 addresses of the windmill
-  // workers.
-  repeated Endpoint direct_path_endpoints = 4;


reserve the tag # so it is not reused

reserved 4;

I don't see a newer commit. You're likely not done addressing the other comments yet but just fyi in case you're waiting on me.

yep changes are still in the repo, was just using this to keep track of what i was finished with will tag you when finished thanks!

m-trieu · 2023-09-06T01:06:19Z

@scwhittle ready for another round of reviews!

i had the WindmillServiceAddress.java file in the next CL in the chain, but I moved it to this one since it fits with the updates.

scwhittle · 2023-09-06T10:14:34Z

...org/apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcGetWorkerMetadataStream.java

+  /** Guards access to metadataVersion and latestResponse. */
+  private final StampedLock metadataLock;
+
+  private long metadataVersion;


use GuardedBy annotations

scwhittle · 2023-09-06T10:18:48Z

...org/apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcGetWorkerMetadataStream.java

+            metadataVersion,
+            new ThrottleTimer(),
+            serverMappingUpdater,
+            lock);


what about just using @VisibleForTesting annotation instead of injecting lower-level things like the lock which seems like it adds boilerplate here and in the test as well as making it less obvious that it is just for testing.

or add a getMetadataVersion() method that the tests use

done, just reused the create method in the test.

scwhittle · 2023-09-06T10:21:08Z

...org/apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcGetWorkerMetadataStream.java

+   */
+  @Override
+  protected void onResponse(WorkerMetadataResponse response) {
+    long readStamp = metadataLock.tryOptimisticRead();


I don't think the complication of the StampedLock is beneficial here. Both the responses and status page are fairly infrequent and the logic in the critical section is minimal.

More complex locks can perform worse for such cases and is certainly more error-prone.

done. using an object and synchronizing on it for read/write access

scwhittle · 2023-09-06T11:47:00Z

...worker/src/main/java/org/apache/beam/runners/dataflow/worker/windmill/WindmillEndpoints.java

+      return Optional.ofNullable((Inet6Address) directEndpointAddress);
+    }
+
+    public abstract Optional<WindmillServiceAddress> directEndpoint();


nit: move public accessors above all the private methods?
add some comments about how these shoudl be used or refer to the proto file and add a comment there?

done. added comments and linked to proto and java files.

m-trieu · 2023-09-07T04:51:05Z

back to you @scwhittle thank you!

scwhittle

LGTM except for question about improving test with matchers. Feel free to ask someone else to review/merge

scwhittle · 2023-09-07T19:40:47Z

...apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcGetWorkerMetadataStreamTest.java

+
+    assertEquals(
+        GLOBAL_DATA_ENDPOINTS.size(), testWindmillEndpointsConsumer.globalDataEndpoints.size());
+    testWindmillEndpointsConsumer.globalDataEndpoints.forEach(


are there matchers you could use instead? This is where having someone more familiar with Java could help review.

found them adding the dependency for the gradle file. I will update the tests to reflect and then submit

using google.Truth

m-trieu · 2023-09-11T23:37:18Z

@scwhittle thanks for reviewing!
@lostluck I had to add a dependency to the gradle file to use Google Truth assertions in a unit test and wanted to make sure i did that correctly. Other than that, ready to merge thank you!

github-actions bot added runners dataflow labels Aug 1, 2023

m-trieu commented Aug 1, 2023

View reviewed changes

...org/apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcGetWorkerMetadataStream.java Show resolved Hide resolved

m-trieu force-pushed the mtrieu-metadata-stream branch from 9e6fc7e to 523e487 Compare August 1, 2023 04:22

scwhittle requested changes Aug 2, 2023

View reviewed changes

m-trieu force-pushed the mtrieu-metadata-stream branch from 523e487 to 2cb875f Compare August 4, 2023 02:54

m-trieu commented Aug 4, 2023

View reviewed changes

...apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcGetWorkerMetadataStreamTest.java Show resolved Hide resolved

m-trieu commented Aug 4, 2023

View reviewed changes

...org/apache/beam/runners/dataflow/worker/windmill/grpcclient/GrpcGetWorkerMetadataStream.java Outdated Show resolved Hide resolved

m-trieu force-pushed the mtrieu-metadata-stream branch 2 times, most recently from 07ad823 to 32ba5c2 Compare August 4, 2023 17:15

scwhittle requested changes Aug 10, 2023

View reviewed changes

m-trieu force-pushed the mtrieu-metadata-stream branch from 32ba5c2 to 061e846 Compare August 11, 2023 05:49

github-actions bot added the Next Action: Reviewers label Aug 11, 2023

scwhittle requested changes Aug 11, 2023

View reviewed changes

m-trieu force-pushed the mtrieu-metadata-stream branch from 061e846 to 711830e Compare August 11, 2023 21:24

m-trieu force-pushed the mtrieu-metadata-stream branch 3 times, most recently from 147b6c8 to 99b89a0 Compare August 16, 2023 01:07

scwhittle requested changes Aug 16, 2023

View reviewed changes

m-trieu force-pushed the mtrieu-metadata-stream branch 2 times, most recently from ab7544e to b4a7a05 Compare August 23, 2023 16:37

scwhittle requested changes Aug 28, 2023

View reviewed changes

m-trieu force-pushed the mtrieu-metadata-stream branch 3 times, most recently from 632aef6 to 5728ff5 Compare August 29, 2023 08:08

scwhittle requested changes Sep 1, 2023

View reviewed changes

m-trieu force-pushed the mtrieu-metadata-stream branch from 94b56e1 to 7c2bf12 Compare September 6, 2023 01:05

m-trieu force-pushed the mtrieu-metadata-stream branch from 7c2bf12 to d97c48c Compare September 6, 2023 01:10

scwhittle requested changes Sep 6, 2023

View reviewed changes

scwhittle reviewed Sep 6, 2023

View reviewed changes

m-trieu force-pushed the mtrieu-metadata-stream branch 2 times, most recently from 81fe06a to f92c3e6 Compare September 7, 2023 04:48

m-trieu force-pushed the mtrieu-metadata-stream branch from f92c3e6 to 8c3b580 Compare September 7, 2023 05:11

scwhittle approved these changes Sep 7, 2023

View reviewed changes

add getworkermetadata streaming rpc

3152fac

m-trieu force-pushed the mtrieu-metadata-stream branch from 8c3b580 to 3152fac Compare September 11, 2023 23:34

lostluck merged commit 9b8644d into apache:master Sep 12, 2023

m-trieu deleted the mtrieu-metadata-stream branch June 10, 2024 15:41

add getworkermetadata streaming rpc #27767

add getworkermetadata streaming rpc #27767

Conversation

m-trieu commented Aug 1, 2023 • edited Loading

GitHub Actions Tests Status (on master branch)

github-actions bot commented Aug 1, 2023

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

m-trieu Aug 10, 2023 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

github-actions bot commented Aug 11, 2023

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

m-trieu commented Aug 11, 2023

m-trieu commented Aug 15, 2023

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

m-trieu commented Aug 25, 2023

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

scwhittle commented Sep 1, 2023

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

m-trieu commented Aug 1, 2023 •

edited

Loading

m-trieu Aug 10, 2023 •

edited

Loading