Thanks David, I've made some changes, can you take a look?
As for the second test (which I removed), the thought was that we could ensure that proper error handling was in place when the number of replicas exceeded the number of tservers. As for the number, since the number of tservers in the test could be changed, I wanted to be sure that this test would pass regardless of how many tservers are configured, so I chose the max value for a signed 16 bit int. Thanks again, Jordan On Tue, Sep 6, 2016 at 1:09 PM, David Alves <[email protected]> wrote: > Hi Jordan > > Thanks for contributing. > I've left some comments. > > Best > David > > On 9/6/16, Jordan Birdsell <[email protected]> wrote: > > Hey all, > > > > I've submitted a patch for KUDU-1593 > > <https://issues.apache.org/jira/browse/KUDU-1593>. This patch exposes > the > > num_replicas method for the KuduTableCreator class to the python api. > > Currently, all tables created in python are created with the default > number > > of replicas. > > > > Can someone please review? > > > > https://gerrit.cloudera.org/#/c/4315/ > > > > Thanks, > > Jordan Birdsell > > >
