Awesome.   Is there a doc page on this? Some questions I have below, if we
don't have a doc page on this, could be helpful.

1. "Starts" at, that indicates if I set drill.exec.rpc.user.server.port to
say 50000, then instead of 31010, 31011, and 31012 (per the doc page) then
it would use 50000, 50001, 50002?
2.  What about the UDP Port it uses (46655 UDP Used for JGroups and
Infinispan. Needed for multi-node installation of Apache Drill.)?
3.  So this can be on per node basis, in that when it registers with the
cluster defined in drill-override.conf, it registers it's ports so the
other nodes are aware of the ports without any other sort of config? (I.e.
I can have bit one start on port 31010 by default, and then have bit two
start at 50000, etc)
4. Are there any odd issues that people have seen by using non-standard
ports or is it built to handle that fairly easily?


Thanks for the answers!

John


On Thu, Feb 11, 2016 at 1:37 AM, Jacques Nadeau <jacq...@dremio.com> wrote:

> You can change the ports that Drill starts at from 31010 to something else
> by setting:
>
> drill.exec.rpc.user.server.port
>
> You can change the HTTP port by setting:
>
> drill.exec.http.port
>
> Both of these are configured in drill-override.conf on a per node basis.
> You can also disable the web server on a particular node by setting
> drill.exec.http.enabled to false.
>
> --
> Jacques Nadeau
> CTO and Co-Founder, Dremio
>
> On Wed, Feb 10, 2016 at 8:54 AM, John Omernik <j...@omernik.com> wrote:
>
> > Using https://drill.apache.org/docs/ports-used-by-drill/ as  start, I
> was
> > wondering if it is possible in the drill-overide.conf to set these ports
> to
> > be a specific port?
> >
> > Do all ports on each drillbit have to be same from bit to bit in a
> specific
> > drill cluster?  I.e. if I set the ports, that will be the ports used on a
> > bits for that cluster? I suppose at least one port must be the same on
> each
> > bit per drill cluster, otherwise they wouldn't know how to find. Either
> > way, is it possible to set these to be different values?
> >
> > Thanks!
> >
> > John
> >
>

Reply via email to