user should not be able to create a xdcr replication on a mixed cluster ( 1.8/2.0) or on a cluster where nodes are down

Aleksey Kondratenko added a comment - 17/Jul/12 2:37 PM Farshid, when filing bug like that please include better description of why you think this is needed. We do have task of disabling CAPI-related stuff in mixed clusters. And for disabling xdcr creation when nodes are down I don't see any good reason for that.

Farshid Ghods added a comment - 17/Jul/12 2:46 PM yes. i field a bug so that i dont forget about this issue indeed. I will assign this to Ketaki so that she adds more information

Farshid Ghods added a comment - 17/Jul/12 2:49 PM Ketaki, can you please add this to the test scenario and also run few tests now to see if user is able to create a replication stream when bucket is not ready or cluster is in a nn-healthy condition .   Aliaksey, creating a replication seems like a cluster wide operation like bucket creation that should fail or disallowed if some nodes are not ready ( as i understand in bucket management . bucket create will succeed now but the status remain yellow when some nodes are down and i dont think that s a good approach of letting users create an entity when some nodes are not even available . if we want to take the same approach for xdcr then we can remove the "on a cluster where nodes are down" from the subject

Abhinav Dangeti added a comment - 22/Aug/12 7:21 PM Clusters:- 3 : 4 Source: 2.0.0-1623-rel nodes (3) Destination: 2.0.0-1623-rel nodes (2) + 1.8.1-942-rel nodes (2) Scenario1:     - Source: Master-2.0, added 2 more 2.0 nodes :: Added successfully Destination: Master-2.0, added 1 more 2.0 node, couldn't add 1.8 nodes :: Error: Version mismatch     - Move to scenario2.. Scenario2:     - Source: <SAME AS BEFORE>     - Destination: Master-1.8, added 1 more 1.8 node, added 2 2.0 nodes :: Added successfully     - Back on the source:         - Create a cluster reference to destination (any node).         - Upon Create replication:               - Error seen (Unable to create replication):: Unexpected server error, request logged. The request logged on the UI was: Server error during processing: ["web request failed", {path,"/controller/createReplication"}, {type,exit}, {what, {{badmatch,false}, {gen_server,call, [remote_clusters_info, {get_remote_bucket, [{uuid, <<"c53ca067c7940f561ae0fb75bc7217ad">>}, {name,"10.3.2.54"}, {hostname,"10.3.2.54"}, {username,"Administrator"}, {password,"password"}], "bucket",true,30000}, infinity]}}}, {trace, [{gen_server,call,3}, {menelaus_web_create_replication, validate_new_replication_params_check_from_bucket, 5}, {menelaus_web_create_replication, handle_create_replication,1}, {menelaus_web,loop,3}, {mochiweb_http,headers,5}, {proc_lib,init_p_do_apply,3}]}]

Aliaksey Artamonau added a comment - 23/Aug/12 6:00 AM I don't think that just saying "Unexpected server error" is the best way to handle this. Can you please attach diags for this scenario?

Aliaksey Artamonau added a comment - 23/Aug/12 6:02 AM Found a separate ticket for this.