You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@celeborn.apache.org by an...@apache.org on 2023/03/14 09:23:51 UTC
[incubator-celeborn] branch main updated: [CELEBORN-420] Add unavailablePeers to http request (#1345)
This is an automated email from the ASF dual-hosted git repository.
angerszhuuuu pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/incubator-celeborn.git
The following commit(s) were added to refs/heads/main by this push:
new 7d7279a9b [CELEBORN-420] Add unavailablePeers to http request (#1345)
7d7279a9b is described below
commit 7d7279a9bc8dd03f56f7da0f7756fc825b2fbbd1
Author: Angerszhuuuu <an...@gmail.com>
AuthorDate: Tue Mar 14 17:23:45 2023 +0800
[CELEBORN-420] Add unavailablePeers to http request (#1345)
* [CELEBORN-420] Add unavailablePeers to http request
---
.../scala/org/apache/celeborn/service/deploy/master/Master.scala | 2 ++
.../scala/org/apache/celeborn/server/common/HttpService.scala | 2 ++
.../apache/celeborn/server/common/http/HttpRequestHandler.scala | 2 ++
.../scala/org/apache/celeborn/service/deploy/worker/Worker.scala | 9 +++++++++
4 files changed, 15 insertions(+)
diff --git a/master/src/main/scala/org/apache/celeborn/service/deploy/master/Master.scala b/master/src/main/scala/org/apache/celeborn/service/deploy/master/Master.scala
index 5c0913cc3..1fc2a817f 100644
--- a/master/src/main/scala/org/apache/celeborn/service/deploy/master/Master.scala
+++ b/master/src/main/scala/org/apache/celeborn/service/deploy/master/Master.scala
@@ -775,6 +775,8 @@ private[celeborn] class Master(
override def listPartitionLocationInfo: String = throw new UnsupportedOperationException()
+ override def getUnavailablePeers: String = throw new UnsupportedOperationException()
+
private def requestGetWorkerInfos(endpoint: RpcEndpointRef): GetWorkerInfosResponse = {
try {
if (endpoint != null) {
diff --git a/service/src/main/scala/org/apache/celeborn/server/common/HttpService.scala b/service/src/main/scala/org/apache/celeborn/server/common/HttpService.scala
index 93f9b52be..b9adc578c 100644
--- a/service/src/main/scala/org/apache/celeborn/server/common/HttpService.scala
+++ b/service/src/main/scala/org/apache/celeborn/server/common/HttpService.scala
@@ -53,6 +53,8 @@ abstract class HttpService extends Service with Logging {
def listPartitionLocationInfo: String
+ def getUnavailablePeers: String
+
def startHttpServer(): Unit = {
val handlers =
if (metricsSystem.running) {
diff --git a/service/src/main/scala/org/apache/celeborn/server/common/http/HttpRequestHandler.scala b/service/src/main/scala/org/apache/celeborn/server/common/http/HttpRequestHandler.scala
index ac853d26d..3b55368f4 100644
--- a/service/src/main/scala/org/apache/celeborn/server/common/http/HttpRequestHandler.scala
+++ b/service/src/main/scala/org/apache/celeborn/server/common/http/HttpRequestHandler.scala
@@ -82,6 +82,8 @@ class HttpRequestHandler(
service.listTopDiskUseApps
case "/listPartitionLocationInfo" if service.serviceName == Service.WORKER =>
service.listPartitionLocationInfo
+ case "/unavailablePeers" if service.serviceName == Service.WORKER =>
+ service.getUnavailablePeers
case _ => INVALID
}
}
diff --git a/worker/src/main/scala/org/apache/celeborn/service/deploy/worker/Worker.scala b/worker/src/main/scala/org/apache/celeborn/service/deploy/worker/Worker.scala
index 7f0e8de49..ff1208e62 100644
--- a/worker/src/main/scala/org/apache/celeborn/service/deploy/worker/Worker.scala
+++ b/worker/src/main/scala/org/apache/celeborn/service/deploy/worker/Worker.scala
@@ -475,6 +475,15 @@ private[celeborn] class Worker(
partitionLocationInfo.toString
}
+ override def getUnavailablePeers: String = {
+ val sb = new StringBuilder
+ sb.append("==========Unavailable Peers of Worker==========\n")
+ unavailablePeers.asScala.foreach { case (peer, time) =>
+ sb.append(s"${peer.toUniqueId().padTo(55, "")}$time\n");
+ }
+ sb.toString()
+ }
+
@VisibleForTesting
def isRegistered(): Boolean = {
registered.get()