Hi everyone, I’d like to start a thread to discuss merging the HDDS-48 branch to trunk. The ContainerIO work refactors the HDDS Datanode IO path to enforce clean separation between the Container management and the Storage layers.
Note: HDDS/Ozone code is not compiled by default in trunk. The 'hdds' maven profile must be enabled to compile the branch payload. The merge payload includes the following key improvements: 1. Support multiple container types on the datanode. 2. Adopt a new disk layout for the containers that supports future upgrades. 3. Support volume Choosing policy for container data locations. 4. Changed the format of the .container file to a human-readable format (yaml) Below are the links for design documents attached to HDDS-48. https://issues.apache.org/jira/secure/attachment/12923107/ContainerIO-StorageManagement-DesignDoc.pdf https://issues.apache.org/jira/secure/attachment/12923108/HDDS DataNode Disk Layout.pdf The branch is ready to merge. Over the next week we will clean up the unused classes, fix old integration tests and continue testing the changes. Thanks to Hanisha Koneru, Arpit Agarwal, Anu Engineer, Jitendra Pandey, Xiaoyu Yao, Ajay Kumar, Mukul Kumar Singh, Marton Elek and Shashikant Banerjee for their contributions in design, development and code reviews. Thanks, Bharat --------------------------------------------------------------------- To unsubscribe, e-mail: hdfs-dev-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-dev-h...@hadoop.apache.org
