Details
-
Bug
-
Status: Closed
-
Major
-
Resolution: Fixed
-
6.6.0
-
Untriaged
-
1
-
Unknown
Description
Came across this while verifying MB-39278
Steps
Setup an EC2 instance and install couchbase in it
Enabled Alternate address on it and configure kv to run on 9000:
curl -v -X PUT -u Administrator:password \
> http://localhost:8091/node/controller/setupAlternateAddresses/external \
> -d hostname=ec2-3-237-24-39.compute-1.amazonaws.com \
> -d kv=9000 \
> -d n1ql=9050
Installed couchbase in a vagrant box
Set up remote link using alternate address of ec2 instance:
[root@node2-mad-hatter-testing-centos7 bin]# ./couchbase-cli analytics-link-setup -c localhost:8091 -u Administrator -p password --create --name testAltLink --dataverse Default --type couchbase --hostname ec2-3-237-24-39.compute-1.amazonaws.com --link-username Administrator --link-password password --encryption none
ERROR: Internal server error, please retry your request
Works fine if kv runs on default port in the ec2 instance
Note, this should be a 400 (i.e. Bad Request) as ec2-3-237-24-39.compute-1.amazonaws.com does not provide a mapping for the mgmt port, it is only a bug that this is returning an internal error.
Both the mgmt and kv ports (or mgmtSSL and kvSSL) ports must be specified for each alternate host to be an acceptable remote cluster target.
There is no functional outage due to this issue.