> On 18 Aug 2016, at 14:57, Junping Du <j...@hortonworks.com> wrote: > > I think Allen's previous comments are very misleading. > In my understanding, only incompatible API (RPC, CLIs, WebService, etc.) > shouldn't land on branch-2, but other incompatible behaviors (logs, > audit-log, daemon's restart, etc.) should get flexible for landing. > Otherwise, how could 52 issues ( https://s.apache.org/xJk5) marked with > incompatible-changes could get landed on branch-2 after 2.2.0 release? Most > of them are already released. > > Thanks, > > Junping
Don't get AW started on compatiblity; it'll only upset him. One thing he does care about is the ability of programs to consume the output of commands and logs —and for that even the output of commands and logs need to continue to be parseable https://hadoop.apache.org/docs/r2.7.1/hadoop-project-dist/hadoop-common/Compatibility.html#Command_Line_Interface_CLI " Changing the path of a command, removing or renaming command line options, the order of arguments, or the command return code and output break compatibility and may adversely affect users." I believe Allen is particularly concerned that a minor point release is going in as incompatible, on the basis the audit log output will change —that's the log that is explicitly designed for machine processing, hooking up to flume & kafka, etc. As example, Spotify spoke at a Hadoop Summit conference about how they used it to identify files which hadn't been used for a long time; inferring an atime attribute from the access history. What has changed in the output? --------------------------------------------------------------------- To unsubscribe, e-mail: mapreduce-dev-unsubscr...@hadoop.apache.org For additional commands, e-mail: mapreduce-dev-h...@hadoop.apache.org
