[ https://issues.apache.org/jira/browse/HDDS-324?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16576792#comment-16576792 ]
Xiaoyu Yao edited comment on HDDS-324 at 8/10/18 8:30 PM: ---------------------------------------------------------- Thanks [~msingh] for working on this. The latest patch look good to me. Just have few minor comments: *Pom.xml* Can you elaborate on the change for upgrade ratis? Does the new version support ReplicationLevel.*_ALL?_* *PipelineID.java* Line66: NIT: "pid" is confusing, can we change it to "pipelineId" *Pipeline.java* Line 250: NIT: "name:" -> "id:" *XceiverServerRatis.java* Line 298: can we make this a configurable option for choosing different ReplicationLevel, such as ALL, MAJOR? Line 308: this is a very good fix so that we don't use a random client id for each request. *ContainerInfo.java* Line 240: NIT: pipelineName => pipelineID *RatisManagerImpl.java* Line 90: NIT: "name" -> "id" *StandaloneManagerImpl.java* Line 89: NIT: "name" -> "id" *TestCloseContainerhandler.Java* Line 23: unused imports TestContainerReportHandler.java Line 111: need to fix after HDDS-245. was (Author: xyao): Thanks [~msingh] for working on this. The latest patch look good to me. Just have few minor comments: *Pom.xml* Can you elaborate on the change for upgrade ratis? Does the new version support ReplicationLevel.*_ALL?_* *PipelineID.java* Line66: NIT: "pid" is confusing, can we change it to "pipelineId" *Pipeline.java* Line 250: NIT: "name:" -> "id:" *XceiverServerRatis.java* Line 298: can we make this a configurable option for choosing different ReplicationLevel, such as ALL, MAJOR? Line 308: this is a very good fix so that we don't use a random client id for each request. *ContainerInfo.java* Line 240: NIT: pipelineName => pipelineID *RatisManagerImpl.java* Line 90: NIT: "name" -> "id" *StandaloneManagerImpl.java* Line 89: NIT: "name" -> "id" *TestCloseContainerhandler.Java* Line 23: unused imports > Use pipeline name as Ratis groupID to allow datanode to report pipeline info > ---------------------------------------------------------------------------- > > Key: HDDS-324 > URL: https://issues.apache.org/jira/browse/HDDS-324 > Project: Hadoop Distributed Data Store > Issue Type: Bug > Components: Ozone Datanode > Affects Versions: 0.2.1 > Reporter: Mukul Kumar Singh > Assignee: Mukul Kumar Singh > Priority: Major > Fix For: 0.2.1 > > Attachments: HDDS-324.001.patch, HDDS-324.002.patch, > HDDS-324.003.patch, HDDS-324.004.patch, HDDS-324.005.patch, HDDS-324.006.patch > > > Currently Ozone creates a random pipeline id for every pipeline where a > pipeline consist of 3 nodes in a ratis ring. Ratis on the other hand uses the > notion of RaftGroupID which is a unique id for the nodes in a ratis ring. > When a datanode sends information to SCM, the pipeline for the node is > currently identified using dn2PipelineMap. With correct use of RaftGroupID, > we can eliminate the use of dn2PipelineMap. -- This message was sent by Atlassian JIRA (v7.6.3#76005) --------------------------------------------------------------------- To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org