karthick-rn commented on issue #284: Support (optional) HA capabilities for Hadoop & Accumulo URL: https://github.com/apache/fluo-muchos/pull/284#issuecomment-533525431 > @karthick-rn thanks for the contribution. I have only taken a quick pass over this and plan to look at it in more detail later. One thing I noticed is there is a good bit of duplication for the hadoop config templates and I was curious what your thoughts on this were. Hi @keith-turner, Support for optional HA was implemented keeping in mind not to disturb the existing ‘hadoop’ role but add an additional role for ‘hadoop-ha’ making both these roles independent of each other. Based on the ‘hdfs_ha’ flag selected from muchos.props, the respective role gets invoked. By default, this flag is set to ‘False’ so the existing ‘hadoop’ role is invoked. When set to ‘True’, ‘hadoop-ha’ role is invoked. In regards to your point on duplication: The configuration files (hdfs-site.xml, core-site.xml, etc..) under templates remain the same for both ‘hadoop’ & ‘hadoop-ha’ however the content of these config files may differ. For instance, under ‘hadoop-ha’ hdfs-site.xml includes Journal quorum, shared edits directory, automatic failover mechanism, etc.. that are mandatory for HA cluster and these configurations are not required for a standalone cluster. Similarly under ‘hadoop’ hdfs-site.xml includes Secondary Namenode which is not required for HA. Happy to take any further questions/comments or jump on a call to discuss further. Thanks!
---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
