Bug 1523598

+++ This bug was initially created as a clone of Bug #1414794 +++

Description of problem:
As of now user cannot prepare gluster nodes from cockpit UI for scaling. User has to do the following tasks manually on each node and it has to be added to the cluster one after the other.

   1)  Install all the required packages in case of RHEL deployments,configuring other components in the host like,
adding the required channels , setting up NTP, setting up glusterfs slice, etc

    3) Add all the three nodes one after the other to UI

    4) Prepare bricks from UI on all the three nodes.

    5) Now create an arbiter / replicate volume from UI.
    
    6) since virt group has only few options, user has to manually go and add other options which are required.

provide a way for the user to perform above actions from cockpit ui.

Version-Release number of selected component (if applicable):
cockpit-ovirt-dashboard-0.10.7-0.0.3.el7ev.noarch

How reproducible:


Steps to Reproduce:
1. 
2.
3.

Actual results:
As of now user cannot prepare gluster nodes for scaling from cockpit UI.

Expected results:
There should be a provision for the user to scale and prepare gluster nodes from cockpit UI. 

Additional info:

--- Additional comment from SATHEESARAN on 2017-01-20 01:18:56 EST ---

(In reply to RamaKasturi from comment #0)
> Description of problem:
> As of now user cannot prepare gluster nodes from cockpit UI for scaling.
> User has to do the following tasks manually on each node and it has to be
> added to the cluster one after the other.
> 
>    1)  Install all the required packages in case of RHEL
> deployments,configuring other components in the host like,
> adding the required channels , setting up NTP, setting up glusterfs slice,
> etc
> 
>     3) Add all the three nodes one after the other to UI
> 
>     4) Prepare bricks from UI on all the three nodes.
> 
>     5) Now create an arbiter / replicate volume from UI.
>     
>     6) since virt group has only few options, user has to manually go and
> add other options which are required.
> 
> provide a way for the user to perform above actions from cockpit ui.
> 

In the above mentioned steps, only step-1 & step-6 is not that easy for Grafton (HC) users.
Step-1 requires cockpit-gdeploy for sure. Step-3, 4, 5 could be done after the node is added in RHV UI.

Again step-6 requires some more help to set the required volume options on the volume.

--- Additional comment from SATHEESARAN on 2017-01-20 06:15:03 EST ---

Pasting the mail conversation so that it gives more context to this bug:
[kasturi]
Following is the only way Cu has to follow to prepare and add 3 more nodes to the cluster.

1) Install three nodes with RHEL7.3/RHV-H.

2) Install all the required packages in case of RHEL deployments. (vdsm, gluster,gluster-nagios-addons).

Package installation is not required, channel subscription should suffice.
You are right, vdsm and gluster packages should get pulled in as part of host deploy. But i think, we might need to install gluster-nagios-addons. 

[sas]
​>I will define this step-2 as 'Preparing the host' which not only includes >package installation, but also configuring other components in the host like,
>adding the required channels ( in case of RHEL server ), setting up NTP, >setting up glusterfs slice, etc
>This stuffs should be setup by cockpit-gdeploy on that node, without new >volume creation

[Ramesh]
This make sense. It will be better if we can do all these things in Cockpit. But I am not sure how to do that now. Current gdeploy wizard will not work. We need to come up with some better approach for node expansion in cockpit. Let me come with some idea for this. Feel free to suggest if you have any better idea.

Regards,
Ramesh