Glad you got it working. Ops are hard :) -n
On Tue, Mar 10, 2015 at 4:56 AM, Michele Giusto <[email protected]> wrote: > Nick Dimiduk <ndimiduk@...> writes: > > > > > In AWS you should use internal host names, not external host names. > The > > reason being is that when a host resolves it's own name, it is > resolved to > > the internal name, which means that's how it is referenced when > entries are > > added to meta. If you use internal host names, /etc/hosts entries are > not > > required. > > > > -n > > > > On Mon, Mar 9, 2015 at 2:37 AM, Michele Giusto <m.giusto@...> wrote: > > > > > Hi Ted, hi Nick, > > > thanks for helping. I was not that confident that the problem was > the > > > hostname, however that was it. The node was recognized with its FQDN > > > because when we added the new node we forgot to modify /etc/hosts > > > neither on the old nodes nor on the new one. Now I have fixed all > > > /etc/hosts and it is working. > > > > > > Any idea on the reason why we were getting that problem? At the end > of > > > the day, all the nodes were communicating as expected (DNS is up, > > > /etc/hosts settings are probably superfluous in my case) and other > > > services (hdfs, map-reduce, Impala, ...) were not reporting problem. > > > > > > > > > Thanks again, > > > Michele > > > > > > > > > > > > > > > > > > > Hi Nick, > thanks for answering, we are using internal host names so things should > have worked also before my edits to /etc/hosts, I suppose. However I > have also found another wrong configuration on Impala and fixed it, now > everything is working and probably the HBase problem was not the real > reason behind my never ending queries. > > > Bye, > Michele > > > >
