Yes, you should ensure that admin credentials, couchdb/secret and erlang cookie are the same on all three nodes. Otherwise they won't connect each other. Also, you have to interact with /_cluster_setup with curl now - Fauxton update in on the way. dev/run script just automatize all these routines.
-- ,,,^..^,,, On Thu, Apr 23, 2015 at 12:32 PM, Antoine Duchâteau <[email protected]> wrote: > No. They happen while starting the cluster by hand (bin/couchdb from > /opt/couchdb directory). We are currently testing couchdb on a cluster of > three different machines. > > I just checked in the dev/run script and I see that it calls > /_cluster_setup on the cluster while specifying the admin user and > password. Should I do that manually when I change the admin password in > local.ini. > > Thanks, > A. > > Alexander Shorin <[email protected]> > 23 avril 2015 11:25 > Hi Antoine, > > Are these errors happened when you run CouchDB 2.0 with dev/run > script? Because for now this is the easiest way to setup a local > cluster to play around with. > -- > ,,,^..^,,, > > Antoine Duchâteau <[email protected]> > 23 avril 2015 10:41 > Hi List, > > I am trying to setup CouchDB 2.0, based on the develop branch of the git > for some time and I run into a series of problems: > > 1. When I try to access the config from Fauxton, I get An Error occurred: > Database does not exist. > 2. When I try to end the admin party using Fauxton, I get Could not > create admin. Database does not exist. > 3. When I set the admin directly inside the etc/local.ini, the cluster > loses access to the shards and I get a : > > 10:38:15.324 [error] req_err(2569857535) internal_server_error : No DB > shards could be opened. > [<<"fabric_util:get_shard/4 L177">>,<<"fabric:get_security/2 > L141">>,<<"chttpd_auth_request:db_authorization_check/1 > L79">>,<<"chttpd_auth_request:authorize_request/1 > L19">>,<<"chttpd:handle_request/1 L210">>,<<"mochiweb_http:headers/5 > L93">>,<<"proc_lib:init_p_do_apply/3 L239">>] > > Removing the password from the local.ini doesn't fix the issue. In fact, I > have not been able to recover the access to the shards at this point. > > Should I start with another branch or is there a way to manually fix the > config ? > > Thanks in advance, > Antoine > > >
