I just set up a NSS cluster with OES2 Linux 64bits.

It works fine, if a node shuts down, the volume are corectly migrated on an other node. My problem is related to the NCP port (TCP 524) which is quite mandatory to map an NSS volume for a client.

1. I realize that if I restart ndsd on a node, my resources hosted by the node are not listing on 524 port anymore.
I need to migrate de volume on another node for it to be listing on again.

2. if ndsd stop working for any reason on a node, the cluster doesn't detect it and doesn't auto-migrate the volumes. Of course the node still works but the clients are not able anymore to join the volumes. I mean that's what I saw on my lab.

Is it a normal behavior ? and if it is, is there a way to make NCS handle this ?


Sbastien Pulvrin