Couchbase queries stop working after adding a 2nd node to the cluster

SQL++
, ,
1

I have a 1 node cluster that runs my queries fine (through the couchbase console and through my application). I was able to pretty easily add a 2nd node to the cluster (with all the services except FTS) and now I can’t run queries through my application. My login query takes 35ms in the couchbase console but that same query now causes a “An unknown N1QL error occured. This is usually related to an out-of-memory condition” every single time.

Looking at the logs on the couchbase console it doesn’t look like the RAM or CPU usage is doing anything crazy so I am really confused by this.

Any thoughts?

2

Is the same issue in couchbase console. Try restart the application or cluster. or kill cbq-engine which will be started automatically

3

No, the queries run very quickly in the couchbase console (~35ms). The exact query returns an “out of error” memory though through the application though (only when there’s 2 nodes in the cluster, 1 works fine).

4

Application using what type of SDK? Restart the application may help If the queries are working on console.

5

I am using Node.js and couchbase community version 4.5.1 I thought it may be a RAM issue but I gave each node plenty of RAM and it only fails when there’s 2 nodes, 1 works perfectly. Thoughts?

6

Try simple statement by connecting to each query node on 8093 port Using Query workbench or cbq shell or REST API
https://developer.couchbase.com/documentation/server/current/n1ql/n1ql-rest-api/exsuccessful.html

7

Oh wow, port 8093 was not open on my 2nd node so I went and opened it up and now it works. Thank you!