What steps will reproduce the problem? 1. I have set up 1 master directory node and 3 volume nodes on 4 different servers 2. When I execute an assign command the Unknown Replication Type error appears

What is the expected output? What do you see instead? {{{ curl http://10.0.0.70:9333/dir/assign?replication=002 }}} produces {"error":"Unknown Replication Type:002"}

What version of the product are you using? On what operating system? curl http://10.0.0.70:9333/dir/status {"Topology":{"DataCenters":[{"Free":12,"Max":12,"Racks":[{"DataNodes":[{"Free":4,"Max":4,"PublicUrl":"10.0.0.75:8080","Url":"10.0.0.75:8080","Volumes":0},{"Free":4,"Max":4,"PublicUrl":"10.0.0.76:8080","Url":"10.0.0.76:8080","Volumes":0},{"Free":4,"Max":4,"PublicUrl":"10.0.0.77:8080","Url":"10.0.0.77:8080","Volumes":0}],"Free":12,"Max":12}]}],"Free":12,"Max":12,"layouts":null},"Version":"0.45"}

I'm using Centos 6.4

Please provide any additional information below.

The master node is started with this command: ./weed master -mdir="." -readTimeout=30 -defaultReplicationType="002" &

The volume nodes with: ./weed volume -dir="." -mserver="10.0.0.70:9333" -publicUrl="10.0.0.75:8080" -dataCenter="Groenekan" -rack="rack1" -max="4" -readTimeout=30 -ip="10.0.0.75" & ./weed volume -dir="." -mserver="10.0.0.70:9333" -publicUrl="10.0.0.76:8080" -dataCenter="Groenekan" -rack="rack1" -max="4" -readTimeout=30 -ip="10.0.0.76" & ./weed volume -dir="." -mserver="10.0.0.70:9333" -publicUrl="10.0.0.77:8080" -dataCenter="Groenekan" -rack="rack1" -max="4" -readTimeout=30 -ip="10.0.0.77" &

What am I doing wrong?

Your version is old. Can you please test with a more recent version?

Thanks for your reply.

I downloaded and used the most recent binary (I think): weed_0.45_linux_amd64.tar.gz To which version should I upgrade? However, I don't know how to compile Go... or is it just running a makefile?

The link is on the front page:

https://bintray.com/chrislusf/Weed-FS/weed

Thanks, I am on version 0.55 now.

My setup is the same.

The master node: ./weed master -mdir="." -defaultReplication="002" -ip="10.0.0.70" &

The volume nodes (one of them, the others have a different ip number of course): ./weed volume -dir="." -mserver="10.0.0.70:9333" -publicUrl="10.0.0.75:8080" -dataCenter="Groenekan" -rack="rack1" -max="4" -ip="10.0.0.75"

The result is that the volume node is looking for the master on localhost despite the -mserver parameter pointing to another ip:

I0421 21:31:18 21762 file_util.go:19] Folder . Permission: -rwxr-xr-x I0421 21:31:18 21762 store.go:237] Store started on dir: . with 0 volumes max 4 I0421 21:31:18 21762 volume_server.go:72] store joined at 10.0.0.70:9333 I0421 21:31:18 21762 volume.go:81] Start Weed volume server 0.55 at http://10.0.0.75:8080 I0421 21:31:18 21762 http_util.go:27] Post http://localhost:9333/dir/join: dial tcp 127.0.0.1:9333: connection refused

Comment deleted

Did you try "-peers" before? See https://code.google.com/p/weed-fs/issues/detail?id=70

Please delete file /tmp/conf

Removing the ./conf file did it. It's working now. Thanks for your help!