Crowbar vs OpenCrowbar

[S Nivas]

Hi All,

         I am new to Crowbar. Could some one help me in understanding the difference between Crowbar and Opencrowbar? I am looking for nodes api in crowbar(not opencrowbar)

to add nodes manually (same as the one here in opencrowbar documentation). Is that possible in Crowbar?

Thanks in advance!

[NivaS]

I found the answer doe difference between Crowbar and OpenCrowbar. Could any one help me with this query,

I am looking for nodes api in crowbar(not opencrowbar) to add nodes manually (same as the one here in opencrowbar documentation). Is that possible in Crowbar?

[Adam Spiers]

NivaS <agsn...@gmail.com> wrote:
> On Friday, 22 April 2016 09:56:54 UTC+5:30, S Nivas wrote:
> >
> > Hi All,
> > I am new to Crowbar. Could some one help me in understanding the

> > difference between Crowbar <https://github.com/crowbar>and Opencrowbar
> > <https://github.com/opencrowbar>?

Crowbar is an active project maintained by SUSE with the primary focus
of deploying OpenStack:

https://youtu.be/SpLbDlTmVFE

OpenCrowbar was a rewrite of Crowbar, but it is now a dead project,
superseded by Digital Rebar:

http://rebar.digital/

(and Rob, if you are reading this, it would be helpful to kill the
OpenCrowbar organization to avoid it causing further confusion!)

> > I am looking for nodes api in crowbar(not opencrowbar)
> > to add nodes manually (same as the one here in opencrowbar documentation

> > <https://github.com/lhrc-mikeyp/opencrowbar-docs/blob/master/doc/development-guides/api/node.md>).

> > Is that possible in Crowbar?

Yes it is.

> I am looking for nodes api in crowbar(not opencrowbar) to add nodes
> manually (same as the one here in opencrowbar documentation

> <https://github.com/lhrc-mikeyp/opencrowbar-docs/blob/master/doc/development-guides/api/node.md>).

> Is that possible in Crowbar?

Crowbar has a similar nodes API. But which version of Crowbar are you
using, and what are you actually trying to achieve with it?

[NivaS]

Thanks Adam.

I have installed suse openstack cloud 6 and using the crowbar that comes with it. Could you provide me the pointers to those node apis ?

My objective is to skip the discovery process and to register the nodes. Assume you have the details of all baremetal nodes(Mac,ip., .etc) available, how to register them with crowbar?

Thanks,
Nivas

[Dirk Müller]

Hi Nivas,

Does the documentation help?

https://www.suse.com/documentation/suse-openstack-cloud-6/book_cloud_deploy/data/sec_depl_inst_nodes_install_external.html

Thanks,
Dirk

[NivaS]

Thanks dirk.

Yes, I am following this method currently.

I would like to know if there are any explicit REST APIs /CLIs to do that.

[Dirk Müller]

Hi,

I wonder, how is crowbar_register different from a CLI? What is missing from it?

There is no direct REST call for this because the registration involves an initial chef-client run (at least collection of system information and upload to the node database), which means it is a multiple step process including installation of additional packages and registration against the chef server.

Are you saying that you want to register a node without executing code on that note?

Greetings,
Dirk

[NivaS]

>>I wonder, how is crowbar_register different from a CLI? What is missing from it?

Here the mechanism is pull. You need to get that script in every node and execute it. Push would be better like you execute a command at server side.

>>There is no direct REST call for this because the registration involves an initial chef-client run (at least collection of system information and upload to the node database), which means it is a multiple step process including installation of additional packages and registration against the chef server.

If you have all the data already available, you can skip that step right?. If you go through the link I referred in opencrowbar, there you can add a node with just name and Mac address..then update it more details.

>>Are you saying that you want to register a node without executing code on that note?

Yes, I am looking for such thing . The registration related thing happens at one node(admin node). I agree that even after such thing, the crowbar needs to contact every registered node to do certain standard configurations as in crowbar_register script

Thanks,
Nivas

[Adam Spiers]

Hi again,

Sorry for the slow reply - we were busy with the OpenStack summit in Austin.

What is the use case for this? Crowbar will need to execute code on
the node as soon as it is allocated, so what's the benefit of avoiding
that early during registration?

Regards,
Adam

[NivaS]

Thanks Adam.

What is the use case for this? 

Suppose you have list of nodes which already has OS and you want to include those nodes in to Crowbar. you need to run crowbar_register in each and every node. It would be better to register those nodes from the crowbar admin node itself

Crowbar will need to execute code on  the node as soon as it is allocated, so what's the benefit of avoiding  that early during registration? 

No benefit. But we are in need to add a list of machines into crowbar which already has os

Thanks,

Nivas

[Adam Spiers]

NivaS <agsn...@gmail.com> wrote:
> Thanks Adam.
>
> What is the use case for this?
>
> Suppose you have list of nodes which already has OS and you want to include
> those nodes in to Crowbar. you need to run crowbar_register in each and
> every node. It would be better to register those nodes from the crowbar
> admin node itself
>
> Crowbar will need to execute code on the node as soon as it is allocated,
> so what's the benefit of avoiding that early during registration?
>
> No benefit. But we are in need to add a list of machines into crowbar which
> already has os

I am confused - if there is no benefit to avoiding running
crowbar_register, then why not just use that approach?