Closed yveskerwyn closed 8 years ago
Can we also define the project...
We install them in the 0-complexity repo right? This has impact on the install script of @maxux
This is not a good idea for me. By convention, we used the repository name to be the same as the environment name. I think few blocks of code use this logic and this would break some code. Not only the installer.
A better way for me would use an organisation (I'm not sur about the name) for environments. Like gig-environments/be-scale-1
for exemple, and not keep them on 0-complexity's organisation.
We put them on 0-complexity
the first time because we didn't know where to put them with private option available. Now github's private repository policies changed, we can review something better ?
I'm in to move the env repos into a new organization
so what was discussed this morning is that we will have all our environments available on the gig-projects repo.
For every customer there is a proj_repo
an environment GIT repo needs to run on the master of the specific environment. In this repo we should have all the available information of that specific environment including:
So for our internal environments we indeed should make a repo called eg. proj_greenitglobe_env
For each environment owned by GIG we should have a separate folder with the name of that environment.
In the readme.md
file of that folder there should be a link to the environment repo which is located on the master of that environment.
@FastGeert @hofkensj
We all agree on that?
If proj_greenitglobe_env is an organisation and not a repository like you say, okay, otherwise that means that all environment will use the same repo and use specific folder on it ?
proj_gig_env should be a repo of the gig-projects organisation, We have a repo per customer called proj_customername, in this repo we should store the environment information of that customer in a dedicated environment folder.
The aim is to handle our environments like we should do for a customer project. Like this we test the full process.
That mean that all environment from the same customer got the information from others environment... Is that not a security issue ?
ok yes, as discussed, this would not be how we want this....
we have agreed on creating a repo per environment on the gig-projects organisation
repo naming convention should be env$customername$environmentname
the $environmentname has the following naming convention $countrycode - $G8type - $environmentnumber
$G8type can be scale = for scaleout environments (separate CPU 1U nodes (1 motherboard per server) which are scale able, storage and CPU all in one unit)
conv = converged environment (multiple nodes in one system eg. 4+ nodes in one physical box with shared backplane. Storage + multiple CPU nodes in one box)
stor = for environments which are using +2 storage nodes with a minimum CPU config (5 nodes)
G8 = for our G8 configs (CPU + separate storage on 70 disk nodes)
When there needs to be a cockpit deployed we need to create another separate repo for that operator. naming convention should be
cockpit$customername$purpose
$customername = the reseller of the operator capacity, or "gig" for internal use $purpose = a short useful name which describes the cockpit function
eg.
@FastGeert @hofkensj
pleas comment, if no comment we will add in the agile process
I've updated the Agile guidelines with this information, it's part of my last pull request.
Thanks for reviewing the pull request: https://github.com/Incubaid/dev_process/pull/38/files
now is moved to become cockpit_... repo's
cockpit and env is not the same, or is it @despiegk ?
@yveskerwyn env_ is for g8 installed by ays7 (which does not relate to a cockpit (yet))
So: be-scale-1 -> env_be-scale-1 be-scale-2 -> env_be-scale-2 ...