+1 on re-fix. I don't think it was intentional. The command "hbase master stop" ought to stop it all by design.
On Sat, Mar 3, 2012 at 4:58 AM, Gregory Chanan <gcha...@cloudera.com> wrote: > Link: https://issues.apache.org/jira/browse/HBASE-5213 > > From my tests, calling "hbase master stop" or "HBaseAdmin.shutdown()" does > not seem to bring down backup masters. JD seemed to have fixed this in > HBASE-1448 (https://issues.apache.org/jira/browse/HBASE-1448), but it looks > like that code went away with the master rewrite. Does anyone have an > objection to me fixing using the same idea/code as 1448 (adding a "shutdown > node" in ZK on which all masters should listen)? Or perhaps the omission > was intentional? > > Thanks, > Greg -- Harsh J