Open Stack

Hi friends,

I finally found the problem. The thing was that my VirtualBox interfaces were created with a different IP than expected by fuel as default IP
addresses.

Here is what I did. Checked VirtualBox -> File -> Preferences like this:

4 virtual box networks

vboxnet0, 10.20.0.1, 255.255.255.0
vboxnet1, 192.168.0.1, 255.255.255.0
vboxnet2, 192.168.1.1, 255.255.255.0
vboxnet3, 172.16.0.1, 255.255.255.0, I put DHCP like on the fuel dashboard that is: Server 172.16.0.1
255.255.255.0, lower address 172.16.0.2, higher address 172.16.0.126


I also did like Fabrizio advice and completely turned off the Fedora NetworkManager with:

# systemctl stop NetworkManager.service
# systemctl disable NetworkManager.service

I verified its effectiveness with:

# nmcli dev status

That should return Error: NetworkManager is not running. After that I shutdown all the nodes.
I configured each node to have 4 network interfaces, all host only, one for each vboxnetY
adapter I have. 

After that, even though the deployment finished with an error, fuel could continue just after ubuntu installation
and successfully deploy openstack over Vbox.


Cheers and thanks for the help,
Pedro



> On Apr 24, 2015, at 18:24, Pedro Velho <pedro.velho at gmail.com> wrote:
> 
> Yes I did, network verification passes. Using the 10.20.0.x network I get conectivity the problem is with this interfaces created by puppet, 192.168.0.x,  there is no switch I am using fuel on virtualbox.
> 
> Thanks,
> Pedro
> Sent from smartphone
> 
> On 24/04/2015, at 17:40, Mike Scherbakov <mscherbakov at mirantis.com <mailto:mscherbakov at mirantis.com>> wrote:
> 
>> Likely you don't have connectivity?
>> If you used VLANs, check your switch / if you use correct ports...
>> 
>> There is a network verification in Fuel, it should pass before you start deployment. Did you try to run it?
>> 
>> On Fri, Apr 24, 2015 at 5:17 AM, Pedro Velho <pedro.velho at gmail.com <mailto:pedro.velho at gmail.com>> wrote:
>> Hi,
>> 
>> Sorry but this was not my issue. I completely deactivated the NM and still have the same problem.
>> 
>> # service NetworkManager stop
>> Redirecting to /bin/systemctl stop  NetworkManager.service
>> 
>> I discovered the /etc/hosts was generated by puppet and has the good values.
>> The problem is that those 192.168.x.x interfaces cannot see each other on deployed nodes. Meanwhile I did a dirty hack
>> I changed the /etc/hosts file just after puppet created them. Nevertheless I would
>> like to understand why these 192.x.x.x interfaces cannot see each other.
>> 
>> 
>> Any help is welcome,
>> Pedro
>> 
>> 
>> 
>> 
>> 
>> 
>> On Thu, Apr 23, 2015 at 5:24 PM, Fabrizio Soppelsa <fsoppelsa at mirantis.com <mailto:fsoppelsa at mirantis.com>> wrote:
>> Hi Pedro,
>> there is a known issue when deploying HA Fuel clusters on Fedora: the interference of the NetworkManager service.
>> Check if this applies to your case https://review.openstack.org/#/c/156188/11/pages/virtualbox/0300-install-virtualbox.rst <https://review.openstack.org/#/c/156188/11/pages/virtualbox/0300-install-virtualbox.rst>
>> 
>> HTH,
>> Fabrizio
>> 
>> 
>> 
>> On 04/23/2015 05:11 PM, Pedro Velho wrote:
>>> Hi everybody,
>>> 
>>> I am new around here so sorry if I am doing something wrong posting my question on the wrong track. The thing is that I am fighting against the clock with this project and I stuck on deploying OpenStack with fuel. 
>>> 
>>> I am installing openstack on a virtualbox environment using fuel-community-6.0.iso, the fuel version is the last one from the public git repository, checkout and rebased this morning. My host is a Fedora 21 with kernel 3.19.3, and virtualbox 4.3.26. 
>>> 
>>> I configure a simple deployment with 5 machines: 2 compute, 2 storage ceph OSD, and 1 controller. When deploying openstack, the installation of ubuntu runs smoothly, the problem happens during openstack installation. What exactly happens is an error like below on all compute and storage nodes:
>>> 
>>> 2015-04-23 13:52:33ERR
>>>  (/Stage[main]/Ceph::Conf/Exec[ceph-deploy config pull]/returns) change from notrun to 0 failed: ceph-deploy --overwrite-conf config pull node-1 returned 1 instead of one of [0]
>>> This error means that there is no ssh route to node-1. I manually checked on the node in question and doing a simple ssh to node-1 won' t work, like this:
>>> 
>>> root at node-4:~# ssh -v node-1
>>> ...
>>> debug1: connect to address 192.168.0.1
>>> ...
>>> ssh: connect to host node-1 port 22: No route to host
>>> 
>>> Indeed it makes sense that ceph-deploy does not reach node-1, it is solving the name incorrectly
>>> on the ssh command. However if I manually try to resolve node-1 for a name I get a correct IP:
>>> 
>>> root at node-4:~# nslookup node-1
>>> ...
>>> Address: 10.20.0.4
>>> 
>>> So there is clearly a problem in the way that node-1 is being resolved. Could you please help me getting this to work. Any help would be very much appreciated.
>>> 
>>> Thanks in advance,
>>> Pedro
>>> 
>>> 
>>> 
>>> 
>>> _______________________________________________
>>> Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack <http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack>
>>> Post to     : openstack at lists.openstack.org <mailto:openstack at lists.openstack.org>
>>> Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack <http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack>
>> 
>> 
>> _______________________________________________
>> Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack <http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack>
>> Post to     : openstack at lists.openstack.org <mailto:openstack at lists.openstack.org>
>> Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack <http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack>
>> 
>> 
>> 
>> _______________________________________________
>> Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack <http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack>
>> Post to     : openstack at lists.openstack.org <mailto:openstack at lists.openstack.org>
>> Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack <http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack>
>> 
>> 
>> 
>> 
>> -- 
>> Mike Scherbakov
>> #mihgen
>> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack/attachments/20150427/414a194e/attachment.html>