Hi
I currently get some timeout errors on my couchcluster.
Cluster setup:
Couchbase 4.0.0 Community edition
4 Servers
1 Interface for cluster replication
1 Interface for client access
I bought the server on an external server hoster (hetzner) because of this I’m not able to do a lot of network configuration.
To get better performance on the internal cluster replication I let my provider cable the second network interface on each server to a dedicated switch.
To get this working I added the servers by dns name and entered in the hosts file on each server the internal IP for the other servers.
Here a litte picture of the network setup:
I see them on the client side (node.js sdk) and in the cluster communication.
On the client side I get sometimes the errorcode 23 which means according to the docs that the request has timeouted.
On the server side I see a lot of errors that tells me that it coudn’t receive the indexstatus:
index_status_keeper_worker<0.363.0>:index_rest:get_json:45]Request to http://127.0.0.1:9102/getIndexStatus failed: {error,timeout}
and that one server cannot reach the current master:
Haven’t heard from a higher priority node or a master, so I’m taking over.
Resultion?
Can the jumbo frames help me?
My idea is to activate the jumbo frames on the second network interface to speed up the replication