The main issue seemed to be /etc/hosts on the nodes, after that the most issues have gone away. I did ceph-deploy install (without specifiying the distribution, so I assume stable is installed) on all 3 nodes the same time which worked without any problems (I use Debian wheezy). The OSDs on wheezy pointed journal to a uuid-path in devfs which didn't exist on wheezy so I had to correct it to a existing path to the partition. Perhaps this is a debian issue and perhaps this could be solved by naming the path in the deploy-command even if it's on the same hdd which I haven't tested so far. This took the main time debugging because this error doesn't show up clearly in the logs. Regards, Roman On Fri, 24 May 2013 15:43:15 -0700, John Wilkins <john.wilkins@xxxxxxxxxxx> wrote: > I ran into a few issues like this when writing the procedure up. One > problem with gatherkeys had to do with network/firewall settings. > Another problem had to do with being unable to find the mds-bootstrap > when using Bobtail. The OSDs did work for me. With Cuttlefish, the > bootstrap keys issue went away for me. > > You did install Ceph on each of those hosts--with our without > ceph-deploy--correct? You'll need Ceph on each of them. I notice you > hadn't mentioned that in your writeup. > > http://ceph.com/docs/master/rados/deployment/ceph-deploy-install/#install > > You can specify multiple hosts for install and uninstall. > > > Regards, > > > John _______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
 |