[ https://issues.apache.org/jira/browse/HADOOP-15774?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16623767#comment-16623767 ]
Steve Loughran commented on HADOOP-15774: ----------------------------------------- bq. Would it make sense to make the registry a standalone service instead of running in the RM? Being part of the YARN may bring some weird dependencies (HDFS->YARN), any ideas here? Should it be move to a separate place? Maybe parts? It can run standalone —its just a subclass of org.apache.hadoop.utils.AbstractService with its normal lifecycle. It's integrated in YARN so that * the RM can set things up with its kerberos permissions. * it can do cleanup after an app is shut down. Provided HDFS is set up with the permissions to edit the /services part of the tree, it could run the ZK "manager" service internally, registering itself on startup. This'd be easier if the HDFS NN went over to using that same Hadoop util.Service lifecycle, but if not, it could still wire it up. All you really need to keep running is ZK. > Discovery of HA servers > ----------------------- > > Key: HADOOP-15774 > URL: https://issues.apache.org/jira/browse/HADOOP-15774 > Project: Hadoop Common > Issue Type: New Feature > Reporter: Íñigo Goiri > Priority: Major > Attachments: Discovery Service.pdf > > > Currently, Hadoop relies on configuration files to specify the servers. > This requires maintaining these configuration files and propagating the > changes. > Hadoop should have a framework to provide discovery. > For example, in HDFS, we could define the Namenodes in a shared location and > the DNs would use the framework to find the Namenodes. -- This message was sent by Atlassian JIRA (v7.6.3#76005) --------------------------------------------------------------------- To unsubscribe, e-mail: common-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: common-issues-h...@hadoop.apache.org