Re: problems with ceph.init / etc/hosts and ceph.conf

[Wido den Hollander <wido@widodh.nl>]

On 06/20/2012 11:20 AM, Stefan Priebe - Profihost AG wrote:
> Am 20.06.2012 11:13, schrieb Wido den Hollander:
>> On 06/20/2012 10:59 AM, Stefan Priebe - Profihost AG wrote:
>>> Hello list,
>>>
>>> i've some problems to find the correct combination of ceph.conf,
>>> /etc/hosts for the init script to work correctly.
>>>
>>> My systems are named:
>>> ssdstor000i
>>> ssdstor001i
>>> ssdstor002i
>>>
>> What does hostname | cut -d . -f 1 say?
> It says ssdstor000i as the machine defaults to it's public IP.
>
> Whouldn't it make sense to go through all aliases set in /etc/hosts
> which point to a local IP?

Ceph is not really intended to work with public or private IP's.

The idea is that you have a Layer 2 or Layer 3 network over which all 
the daemons and clients connect over a routable network, peer-to-peer, 
no NAT in it's way.

You can however have a seperate "osd" network where the OSD's 
interconnect for their replication traffic, but that has nothing to do 
with the hostnames.

You can take a look at the configuration options "public_addr" and 
"cluster_addr" for these.

Besides that, Ceph would run on a "private" network. If you have KVM 
hosts these will have a WAN connection over which they (and their VM's) 
communicate with the outside world and over a dedicated NIC the host 
communicates with the Ceph cluster.

I'd recommend sticking with 1 IP-Address for each OSD, keep it simple.

Wido

>
> Stefan
> --
> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html