We can use this OVH IP, it's stable enough now.

DNS records have been updated:

2001:470:1f13:9e1:0:cac:7e:1 (cac:7e for cache, cac:7e:1 for cache #1).

During WindRiver reprovisioning, the network unit behaved correctly.

So, to get routing back:

To make it works:

igb0 is the right target interface

Salt deployment of /usr/ports as Git repository done.

Taking db-B-001 and web-001.

Removing TC2 as not used anymore for accounts creation.

This is still an issue for accounts created with Salt.

Not sure of the current benefit to use TC2.

Consolidation recent works occurred at T1828 to further unify nginx configurations.

Some context about this staging repository:

[Moved to T1620]

This is still needed for acme.sh if we want to provision different *.nasqueron.org certificates on different servers.

Really blocked by T1602 if we want to have this on any server without copying private keys around.

The problem is somewhat fixed by two things:

That's an issue for Chrome to report upstream.

Local cache is OK.

Salt SELinux module issue

SELinux context was the default for anything created under /var, which we didn't allow and aren't interested to allow for nginx.

Authoritative DNS

DNS change

For RHEL 8, we probably only need docker-processes, but roles/paas-docker/devel refers dive too.

Those two are indeed authoritative:

Done at the console, but DNS records aren't instantaneously updated, and I'm not even sure those are authoritative (they are for poneytelecom.eu):

2024-10 plan:

Choose to only support FreeBSD:14:amd64 ABI or to use /${ABI}/ in path