Merge pull request #228624 from sreekzz/patch-150

Freshness MS Date change

Author: prmerger-automator[bot]

articles/hdinsight/hdinsight-troubleshoot-yarn.md

@@ -3,7 +3,7 @@ title: Troubleshoot YARN in Azure HDInsight
 description: Get answers to common questions about working with Apache Hadoop YARN and Azure HDInsight.
 ms.service: hdinsight
 ms.topic: troubleshooting
-ms.date: 08/15/2019
+ms.date: 02/27/2023
 ---
 
 # Troubleshoot Apache Hadoop YARN by using Azure HDInsight
@@ -50,15 +50,15 @@ In this example, two existing queues (**default** and **thriftsvr**) both are ch
 
 These changes are visible immediately on the YARN Scheduler UI.
 
-### Additional reading
+### Further reading
 
 - [Apache Hadoop YARN CapacityScheduler](https://hadoop.apache.org/docs/r2.7.2/hadoop-yarn/hadoop-yarn-site/CapacityScheduler.html)
 
 ## How do I download YARN logs from a cluster?
 
 ### Resolution steps
 
-1. Connect to the HDInsight cluster by using a Secure Shell (SSH) client. For more information, see [Additional reading](#additional-reading-2).
+1. Connect to the HDInsight cluster by using a Secure Shell (SSH) client. For more information, see [Further reading](#additional-reading-2).
 
 1. To list all the application IDs of the YARN applications that are currently running, run the following command:
 
@@ -131,22 +131,22 @@ These changes are visible immediately on the YARN Scheduler UI.
 
 ### Yarn UI isn't loading
 
-If your YARN UI isn't loading or is unreachable, and it returns "HTTP Error 502.3 - Bad Gateway," it highly indicates your ResourceManager service is unhealthy. To mitigate the issue, follow these steps:
+If your YARN UI isn't loading or is unreachable, and it returns "HTTP Error 502.3 - Bad Gateway," it highly indicates your Resource Manager service is unhealthy. To mitigate the issue, follow these steps:
 
-1. Go to **Ambari UI** > **YARN** > **SUMMARY** and check to see if only the active ResourceManager is in the **Started** state. If not, try to mitigate by restarting the unhealthy or stopped ResourceManager.
-2. If step 1 doesn't resolve the issue, SSH the active ResourceManager head node and check the garbage collection status using `jstat -gcutil <ResourceManager pid> 1000 100`. If you see the **FGCT** increase significantly in just a few seconds, it indicates ResourceManager is busy in *Full GC*, and is unable to process the other requests.
-3. Go to **Ambari UI** > **YARN** > **CONFIGS** > **Advanced** and increase `ResourceManager java heap size`.
+1. Go to **Ambari UI** > **YARN** > **SUMMARY** and check to see if only the active Resource Manager is in the **Started** state. If not, try to mitigate by restarting the unhealthy or stopped Resource Manager.
+2. If step 1 doesn't resolve the issue, SSH the active Resource Manager head node and check the garbage collection status using `jstat -gcutil <Resource Manager pid> 1000 100`. If you see the **FGCT** increase significantly in just a few seconds, it indicates Resource Manager is busy in *Full GC*, and is unable to process the other requests.
+3. Go to **Ambari UI** > **YARN** > **CONFIGS** > **Advanced** and increase `Resource Manager java heap size`.
 4. Restart required services in Ambari UI.
 
 ### Both resource managers are in standby
 
-1. Check ResourceManager log to see if below similar error exists.
+1. Check Resource Manager log to see if similar error exists. 
 ```
 Service RMActiveServices failed in state STARTED; cause: org.apache.hadoop.service.ServiceStateException: com.google.protobuf.InvalidProtocolBufferException: Could not obtain block: BP-452067264-10.0.0.16-1608006815288:blk_1074235266_494491 file=/yarn/node-labels/nodelabel.mirror
 ```
 2. If the error exists, check to see if some files are under replication or if there are missing blocks in the HDFS. You can run `hdfs fsck hdfs://mycluster/`
 
-3. Run `hdfs fsck hdfs://mycluster/ -delete` to forcefully clean up the HDFS and to get rid of the standby RM issue. Alternatively, run [PatchYarnNodeLabel](https://hdiconfigactions.blob.core.windows.net/hadoopcorepatchingscripts/PatchYarnNodeLabel.sh) on one of headnodes to patch the cluster.
+3. Run `hdfs fsck hdfs://mycluster/ -delete` too forcefully cleanup the HDFS and to get rid of the standby RM issue. Alternatively, run [PatchYarnNodeLabel](https://hdiconfigactions.blob.core.windows.net/hadoopcorepatchingscripts/PatchYarnNodeLabel.sh) on one of headnodes to patch the cluster.
 
 ## Next steps