{"error":"error","reason":"internal_server_error"} happened when Clustering

[John1Q84]

Hi, I'm very new at NoSQL DB and I have great interests in bigcouch for clustering. BTW, when I set up 5 bigcouch servers using AWS EC2, I face error message like, "root@due1cdb004:/opt/bigcouch/etc# curl -X PUT http://localhost:5984/sharddb
{"error":"error","reason":"internal_server_error"}"

I totally have no idea what I have been wrong.. I just follow steps on cloudant guide from website.

First, I tried to set two instances as one cluster by curl command like, < $ curl -X PUT http://localhost:5986/nodes/due1cdb003@cdb003p -d{} >. It; looks fine. Then, tried to create database 'albums' by , < $ curl -X PUT http://localhost:5984/albums>. And I saw error message next..
what am I wrong?

[John1Q84]

and the log looks like,

[Fri, 07 Dec 2012 10:50:37 GMT] [error] [emulator] [--------] Error in process <0.6323.2> on node 'due1cdb001@cdb001p' with exit value: {{rexi_DOWN,noconnect},[{mem3_rep,rexi_call,2},{mem3_rep,replicate_batch,1},{mem3_rep,go,3},{mem3_rep,go,2}]}

[wohali]

"$ curl -X PUT http://localhost:5986/nodes/due1cdb003@cdb003p -d{}"

This is your problem. Use a fully qualified domain name that resolves to the other machine's instance after the at-sign (@). The {rexi_DOWN,noconnect} error means the node @cbd001p cannot reach the other machine.

[John1Q84]

Thanks wohali for your answer.

But, I still have question with same error log after I revised EC2 public DNS instead of 'cdb001p'.
The log almost same.. hmm...

the error is "'[email protected]' with exit value: {{rexi_DOWN,noconnect},[{mem3_rep,rexi_call,2},{mem3_rep,replicate_batch,1},{mem3_rep,go,3},{mem3_rep,go,2}]}"

what's wrong with me...Orz(it is emoticon means 'down on my knees').....

[wohali]

Check firewall rules between your nodes. Most likely you do not have the various ports required open between them. See http://stackoverflow.com/questions/8459949/bigcouch-cluster-connection-issue for details.

[John1Q84]

Fortunately, I found one more problem that I made. I was connecting two machines by both side..(according to the post, it should be one side).
But symptom is same..they are still not connected.
Of course I checked the port 4369 to communicate one to the other.(Actually they are in the same security group of AWS, that means the machines could be connected by all ports. and I checked by telnet with port.)
hmm... I think I missed very tiny points... I am still looking..
Thanks anyway wohali :)

[vinaysaini]

try to do telnet to the other bigcouch server from main server to see if ports are open using command
telnet fqdn 5986

where fqdn is the fully qualified domain name of the other bigcouch server.it should give output as connected.