szkoludasebastian commented on issue #23908:
URL: https://github.com/apache/pulsar/issues/23908#issuecomment-2667800937

   I performed more tests for pulsar client versions 3.3.3 and 3.3.4 to check 
which version is causing issue but failed to reproduce it. Then I checked again 
versions 4.0.0 and 4.0.1, however, it failed to reproduce for them as well. I 
noticed that right now I built images for our services from our latest master, 
so I checked git history and lately we reverted such change:
   ```java
   loadManagerClassName: 
"org.apache.pulsar.broker.loadbalance.extensions.ExtensibleLoadManagerImpl"
   loadBalancerLoadSheddingStrategy: 
"org.apache.pulsar.broker.loadbalance.extensions.scheduler.TransferShedder"
   ```
   we had it added in broker configmap. 
   It turned out that when I add this to our broker configmap, I am able to 
reproduce the problem. When I remove it, I can't reproduce it. So this is 
probably the cause of the problem. I also tested this on the latest pulsar 
client and server (4.0.2) and it behaves the same way.
   
   Here is our broker configmap (added these two configurations mentioned 
above):
   
   ```
   apiVersion: v1
   data:
     PULSAR_GC: |
       -XX:+UseG1GC -XX:MaxGCPauseMillis=10 
-Dio.netty.leakDetectionLevel=disabled -Dio.netty.recycler.linkCapacity=1024 
-XX:+ParallelRefProcEnabled -XX:+UnlockExperimentalVMOptions 
-XX:+DoEscapeAnalysis -XX:ParallelGCThreads=4 -XX:ConcGCThreads=4 
-XX:G1NewSizePercent=50 -XX:+DisableExplicitGC -XX:-ResizePLAB 
-XX:+ExitOnOutOfMemoryError -XX:+PerfDisableSharedMem -Djute.maxbuffer=15485760
     PULSAR_MEM: |
       -Xms512m -Xmx5120m -XX:MaxDirectMemorySize=5120m
     acknowledgmentAtBatchIndexLevelEnabled: "true"
     brokerDeduplicationEnabled: "false"
     brokerDeleteInactiveTopicsEnabled: "false"
     brokerServicePort: "6650"
     clusterName: integ-pulsar
     configurationStoreServers: integ-pulsar-zookeeper:2181
     defaultNumberOfNamespaceBundles: "16"
     dispatcherDispatchMessagesInSubscriptionThread: "false"
     exposeTopicLevelMetricsInPrometheus: "true"
     functionsWorkerEnabled: "false"
     loadBalancerLoadSheddingStrategy: 
org.apache.pulsar.broker.loadbalance.extensions.scheduler.TransferShedder
     loadBalancerNamespaceMaximumBundles: "512"
     loadManagerClassName: 
org.apache.pulsar.broker.loadbalance.extensions.ExtensibleLoadManagerImpl
     managedLedgerDefaultAckQuorum: "2"
     managedLedgerDefaultEnsembleSize: "3"
     managedLedgerDefaultWriteQuorum: "3"
     managedLedgerMaxLedgerRolloverTimeMinutes: "60"
     managedLedgerMinLedgerRolloverTimeMinutes: "1"
     managedLedgerOffloadDriver: aws-s3
     maxUnackedMessagesPerConsumer: "500000"
     maxUnackedMessagesPerSubscription: "2000000"
     numHttpServerThreads: "8"
     s3ManagedLedgerOffloadBucket: tiered-storage-us-east-1
     s3ManagedLedgerOffloadRegion: us-east-1
     s3ManagedLedgerOffloadRole: arn:aws:iam::xyz:role/str-integ-broker
     s3ManagedLedgerOffloadRoleSessionName: pulsar-tiered-storage-offload
     statusFilePath: /pulsar/status
     subscriptionExpirationTimeMinutes: "43800"
     subscriptionPatternMaxLength: "200"
     systemTopicEnabled: "true"
     topicLevelPoliciesEnabled: "true"
     webServicePort: "8080"
     zooKeeperSessionTimeoutMillis: "30000"
     zookeeperServers: integ-pulsar-zookeeper:2181
   ```


-- 
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: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to