The start-all scripts will pass in the explicit names from the files to the start of the process as guidance for what IP to report as.
On Mon, Aug 12, 2013 at 1:56 PM, Michael Berman <[email protected]> wrote: > It's just a local, one-node deployment for development and testing. Given > the assumption that it could be fixed with more explicit configuration, > does the warning suggest my splitting isn't working right, or is it working > but also warning? > > Also, the master/slaves files are just used for > start-all.sh/stop-all.sh/etc, right? Not for server identification? > > > On Mon, Aug 12, 2013 at 1:31 PM, Sean Busbey <[email protected]> wrote: > >> It sounds like your cluster configuration isn't using DNS properly. >> >> How many hosts? Are your masters/slaves files using IP addresses or names? >> >> >> On Mon, Aug 12, 2013 at 9:14 AM, Michael Berman <[email protected]>wrote: >> >>> I'm getting lot of warnings, "Got a split from a server we don't >>> recognize: localhost:9997". It looks like the server in my master's >>> tserverSet is called "127.0.0.1" not "localhost". Is this something I >>> should be worried about? >>> >> >> >> >> -- >> Sean >> > >
