The only thing that is supposed to use the cluster network are the OSDs. Not even the MONs access the cluster network. I am sure that if you have a need to make this work that you can find a way, but I don't know that one exists in the standard tool set.
You might try temporarily setting the /etc/hosts reference for vdicnode02 and vdicnode03 to the cluster network and use the proper hosts name in the ceph-deploy command. Ceph cluster operations do not use dns at all, so you could probably leave your /etc/hosts in this state. I don't know if it would work though. It's really not intended for any communication to happen on this subnet other than inter-OSD traffic.
On Thu, Jul 27, 2017 at 6:31 PM Oscar Segarra <oscar.segarra@xxxxxxxxx> wrote:
Sorry! I'd like to add that I want to use the cluster network for both purposes:ceph-deploy --username vdicceph new vdicnode01 --cluster-network 192.168.100.0/24 --public-network 192.168.100.0/24Thanks a lot_______________________________________________2017-07-28 0:29 GMT+02:00 Oscar Segarra <oscar.segarra@xxxxxxxxx>:Hi,¿Do you mean that for security reasons ceph-deploy can only be executed from the public interface?Looks extrange that one cannot decide what network use for ceph-deploy... I could have a dedicated network for ceph-deploy... :SThanks a lot2017-07-28 0:03 GMT+02:00 Roger Brown <rogerpbrown@xxxxxxxxx>:I could be wrong, but I think you cannot achieve this objective. If you declare a cluster network, OSDs will route heartbeat, object replication and recovery traffic over the cluster network. We prefer that the cluster network is NOT reachable from the public network or the Internet for added security. Therefore it will not work with ceph-deploy actions.On Thu, Jul 27, 2017 at 3:53 PM Oscar Segarra <oscar.segarra@xxxxxxxxx> wrote:_______________________________________________Hi,In my environment I have 3 hosts, every host has 2 network interfaces:public: 192.168.2.0/24cluster: 192.168.100.0/24The hostname "vdicnode01", "vdicnode02" and "vdicnode03" are resolved by public DNS through the public interface, that means the "ping vdicnode01" will resolve 192.168.2.1.In my environment the "admin" node is the first node vdicnode01 and I'd like all the deployment "ceph-deploy" and all osd traffic to go from the cluster network.1) To begin with, I create the cluster and I want all traffic to go from the cluster network:ceph-deploy --username vdicceph new vdicnode01 --cluster-network 192.168.100.0/24 --public-network 192.168.100.0/242) The problem comes when I have to launch my commands to the other hosts for example, from node vdicnode01 I execute:2.1) ceph-deploy --username vdicceph osd create vdicnode02:sdb--> Finishes Ok but communication goes through the public interface2.2) ceph-deploy --username vdicceph osd create vdicnode02.local:sdb--> vdicnode02.local is added manually in /etc/hosts (assigned a cluster IP)--> It raises some errors/warnings becase vdicnod02.local is not the real hostname. Some files are created with vdicnode02.local in the middle of the name of the file and some errors appear when starting up the osd service related to "file does not exist"2.3) ceph-deploy --username vdicceph osd create vdicnode02-priv:sdb--> vdicnode02-priv is added manually in /etc/hosts (assigned a cluster IP)--> It raises some errors/warnings becase vdicnod02-priv is not the real hostname. Some files are created with vdicnode02-priv in the middle of the name of the file and some errors appear when starting up the osd service related to "file does not exist"What would be the right way to achieve my objective?If is there any documentation I have not found, please redirect me...Thanks a lot for your help in advance.
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
_______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
