Page MenuHomeSoftware Heritage

staging infra: New vlan
Closed, MigratedEdits Locked


Following documentation at [1], send an email to sesi [2] asking for it:

  • 1 new vlan
  • 1 public IP address (will be used for a SSH front-end/router to that VLAN)
  • open ports for that IP address: http, https, ssh
  • our hypervisors need access to that vlan:
      • (beaubourg)
      • (orsay)
      • (hypervisor3)
    • the ip needs access as well (vm pergamon).

This opened the ticket [3]




Event Timeline

ardumont changed the task status from Open to Work in Progress.Jul 3 2019, 1:33 PM
ardumont triaged this task as High priority.
ardumont created this task.
ardumont updated the task description. (Show Details)
ardumont updated the task description. (Show Details)
ardumont updated the task description. (Show Details)
ardumont changed the status of subtask T1874: staging infra: Setup gateway vm from Open to Work in Progress.Jul 30 2019, 4:33 PM
ardumont mentioned this in Unknown Object (Paste).Aug 1 2019, 3:41 PM

An email asking more details has been received or clarification.
P485 is the draft for the answer.


Most of the relevant commits use the address space.

That network is already being used for the communication between banco and its storage bay.
It would be wise to use a different IP prefix.

sesi asked for some more details about the hardware to propagate that vlan too.
I just replied, so hopefully, that should converge soon.

Also, ack @ftigeot for the address space clash. Thanks for the heads up.
That should change soon enough i think ;)


A priori, everything is done on sesi's side, as asked by @olasd, last ticket entries excerpt:

>>... peux-tu m'indiquer sur quel(s) équipement(s) le nouveau vlan doit être propagé ?
>> Le reste de la configuration a été effectué.
> ...
> deja, merci pour le travail effectue!
> Ce sont les machines suivantes:
> -
> -
> -
> Merci d'avance.

Les modifications sont faites, tiens moi informé si c'est bon de votre côté.
Bonne journée,

Remains to check everything is fine and get back to them so we can close the ticket.


Is this now done? If that's the case this ticked should be closed.

Is this now done? If that's the case this ticked should be closed.

The VLAN has been provisioned on our hypervisors, but not setup / in use yet.

I guess we'll do that when we go ahead and install the new hypervisor.

(the VLAN id for the staging vlan is 443).

I wanted to rename the bridges on the proxmox hosts to something clearer (like vmbr-staging) but it turns out that proxmox only supports bridges named /vmbr\d+/. Ugh.

So I guess the bridges will be named vmbr$<vlanid>, which means vmbr443 for staging.

I've added a bridge vmbr443 to all hypervisors.

I've checked that all hypervisors were able to ping one another over this bridge, using their ipv6 link local addresses.

sudo clush -w @hypervisors 'set -e; for addr in fe80::e872:32ff:feac:ef39%vmbr443 fe80::24a5:2aff:fee3:3552%vmbr443 fe80::c82e:c2ff:fe01:e77d%vmbr443 fe80::589c:d0ff:fe33:4daa%vmbr443; do ping -c3 $addr; done'

Now we just need to move the machines to this VLAN.

olasd claimed this task.

This was now just a matter of doing the clickity click on all hosts. They're now using the dedicated vlan.