Re: [Basho Riak] Fail To Update Document Repeatly With Cluster of 5 Nodes

2017-02-11 Thread my hue
Dear John, I performed the test with the following scenario : * Old cluster of 5 nodes which nodes are not belong to the local network. I performed test with : - pw and pr = 0 Bucket Property : {"props":{"name":*"bucket-name"*,"active":true,*"allow_mult":true* ,"backend":"bitcask_mult","ba

Re: [Basho Riak] Fail To Update Document Repeatly With Cluster of 5 Nodes

2017-02-11 Thread John Daily
It definitely sounds as if your nodes are having problems talking to each other, as suspected. Putting them on a single network is a significant improvement. The default values for pw and pr are 0; Riak by default prefers to always be available, even if the cluster is in a degraded state. This