Asmonitor showing records are written only some of the nodes


#1

i triggered the following command latency -k writes

out of 30 nodes only 6 boxes ops/sec >0 .

In case of reads also same issue some node has more reads/sec compared to other

and in console objects are increasing in all boxes .

yesterday we added new nodes and some nodes went down from that time we are facing issues .

I think having load on only few nodes may cause node going down

could you please update on this


#2

Which client version are you using to write data?

Is your client able to connect to all nodes?

Are you seeing proxies on your cluster?


#3

we are using 3.6.3 version. yes i am able to connect to all nodes

what is mean by proxies ??


#4

3.6.3 is known to be unstable when nodes are added/removed. You need to upgrade to the latest server version.