[Rdo-list] Trying to install the RDO

Marius Cornea marius at remote-lab.net
Sun Nov 15 10:09:13 UTC 2015

On Sat, Nov 14, 2015 at 11:27 PM, Hassan, Ashraf
<ashraf.hassan at t-mobile.nl> wrote:
> Well first I am very very  very sorry, because I have many questions, but the link is not very clear to me, sorry again for my stupid questions.
>     I am trying to compose the file, but I have some questions not clear to me:
>     1- I understood that the new installation requires 3 controller node, what I see in the link you sent me earlier, is states that the undercloud node is different node from the controller nodes? Does it mean that I need 3 controller nodes + the undercloud node, or I need 2 controller node + the undercloud node?

Undercloud node is different from the overcloud controller nodes. You
require 3 overcloud controllers when you deploy a HA control plane. In
that case yes, you need the undercloud + 3 controller nodes deployed
by the undercloud.

>     2- I have a node which I will use as a storage HW, but this node can be connected to the PXE/Boot VLAN over a trunk interface, that mean it cannot be PXE booted (I cannot PXE boot over a trunk interface as far as I know), it means I need to install  it separately configure the VLANs on this interface, and then let it join the controller, is that possible?

PXE boot can work on a trunk interface as long as it's running over
the native vlan.

>     3- Finally for the Jason file itself here is a sample section:
>              {
>             "pm_type":"pxe_ipmitool",
>             "mac":[
>                 "00:17:a4:77:78:30"
>             ],
>             "cpu":"2",
>             "memory":"49152",
>             "disk":"300",
>             "arch":"x86_64",
>             "pm_user":"admin",
>             "pm_password":"password",
>             "pm_addr":""
>         },
>             a- I understand that the MAC address for the PXE boot interface for the different nodes in the overcloud is that correct?

Yes, that's correct.

>             b- In the link it says that pm_addr is " node BMC IP address ", but that is not clear to me, does it mean it is the IP of the "local_ip" in the undercloud.conf, which is always the IP of the eth1 of the undercloud node in the link you sent me earlier, or it is not the case?

This is the IPMI address of the server used for power operations.

>              c- In the link it say that pm_user, and pm_password are "node BMC credentials" , are these the credentials I need to configure manually per node, do they need to be the same for all nodes? Or they are already defined in the stackrc file?

Again, these are the IPMI credentials used for logging into pm_addr .
They need to be configured on the servers to be able to manage them.

> Thanks,
> Ashraf
> ********************************************************************************
> N.B.: op (de inhoud van) deze e-mail is een DISCLAIMER met belangrijke VOORBEHOUDEN van toepassing: zie http://www.t-mobile.nl/disclaimer <http://www.t-mobile.nl/disclaimer>
> This e-mail and its contents are subject to a DISCLAIMER with important RESERVATIONS: see http://www.t-mobile.nl/disclaimer <http://www.t-mobile.nl/disclaimer>
> ********************************************************************************

More information about the rdo-list mailing list