In D4165#103265, @ardumont wrote:

looks good to me.

i don't see vagrant in there but i gather that's what you said about making the network part in vagrant a noop or something.

WDYT to add a variable like profile::network::[activated|managed|whatever] to activate or not the network profile application ? It would avoid to introduce some vagrant specifics in the manifests

this one is easy ;)

In D4150#102910, @ardumont wrote:

If i'm understanding this correctly, this will allow us to generate self-signed certificates when we want to create a service in our stack that needs a certificate.
Just generate it with the script within (generate-certificate) and commit into this repository.
Then trigger back the vagrant provisision <vm-with-desired-service>.
Then everything should run smoothly within that provision step.
correct?

Yes exactly, it's correct. It remains only the icinga part to remove the last errors during the provisioning. I still haven't a simple way to do it as it uses a certificate named with the vm's fqdn and should be generated after the vm creation if we want it to be automatised.

Another question is i'm just wondering whether it should be named netbox-vagrant instead of netbox given what we have in the defaults.yaml [1]

[1] https://forge.softwareheritage.org/source/puppet-swh-site/browse/production/data/defaults.yaml$820-822

good remark. There must be a mistake somewhere on this override of this property as when I provision the vm locally, it searches for netbox. I will remove this declaration because it's not necessary and it will allow to remove a property for vagrant on the defaults.yaml file.

Refactor virtualbox images declarations

Remove useless empty lines

I failed to successfully execute mount on the container without the privileged option so I finally configured the swh-fuse job with this option.

In fact after others tests, only the device and the --privileged option are necessary as runnin in privileged mode completely disable seccomp.

I made some test locally, adding the options --privileged, --device /dev/fuse and --cap-add SYS_ADMIN is working :

The service is up and runnig at https://inventory.internal.softwareheritage.org
I will add the admin password on the credentials.

Uppdate to match the real server name

rebase

nice improvement, thanks

Keep the end of file untouched - second try

Keep the end of file untouched

Revert the netbox typo

Use redis's configuration properties in the configuration template

LGMT

lgtm

lgtm

A good test could be to generate the image, publish it to the public annex and upgrade the image on the Vagrantfile. Doing so, you will cover the complete image lifecycle ;).

lgtm

I also have the problem when I upgrade the plugin to 0.25.0.

rebase

rebase

nice improvment thanks

rebase

LGTM

• rebase
• add the vagrant configuration

rebase

Switch the default test branch to staging

• Add a script to avoid to commit each change before being able to test it (watch-vagrant-conf)
• Update the documentation accordingly
• Declare the admin-inventory server

rebase

Change the commit message according to the diff's feedback

• rebase
• Remove the now useless CASS_DRIVER_NO_CYTHON option

fix typo on the commit message

Use only one commit per diff