In T1475#29341, @DorianWinty wrote:

Due to various problems with vimbadmin,
we will provision mailbox ourself with salt

pefs module has been rebuilt, /opt/python ruamel yaml clib fix updated

Userland ready for update, rebooting server to new kernel

In T1932#29348, @DorianWinty wrote:

misssing php-gettext extention
(installed by myself)

misssing php-gettext extention
(installed by myself)

vimbadmin not maintained anymore

Due to various problems with vimbadmin,
we will provision mailbox ourself with salt

As a follow-up, for shellserver and webserver-alkane roles, we need to bump s/php82/php83 for main package name.

• pprint isn't used anymore
• need to run black
• description needs to be updated
• TTL can be much shorter if we automate this procedure

sudo kill -1 $(cat /var/run/vault.pid)

Error management should be done to check if that pids exist or return an error code.

Needs hvac and pyyaml as packages to be installed on the server, Complector doesn't currently have hvac, only pyyaml.

Decreasing priority as we tested a page successfully on WindRiver.

Currently, the responsibility seems to be to the nginx unit in the main webserver role, not on webserver-core:

000-fallback.conf exists on WindRiver but as an empty file, so without any server block

Did you know? Debian Sid still has 8.2 in the core repo.

Next steps:

Consolidating datacube volumes:

I've asked on a brainstorming channel if someone is interested we take at Online some RPN storage, I didn't receive any answer, so I get nobody is looking for a solution to store a lot of Gb of documents here, hence the priority decrease.

Done for WindRiver, our current development server.

Alcali is still alive.

I've added documentation to https://agora.nasqueron.org/Operations_grimoire/FreeBSD#PostgreSQL without real conviction.

Fixed. Form works again.

I'm rebuilding PostgreSQL 15.4 with XML support on db-A-001.

Jan 14 23:13:52 db-A-001 postgres[35351]: [7-1] 2024-01-14 23:13:52.659 UTC [35351] ERROR: unsupported XML feature

When upgraded to FreeBSD 14, we lost this.

When clicking to Save Progress:

Added RHEL question.

In T1924#28790, @dereckson wrote:

router-001 update would cut network connections between Ysul, WindRiver, CloudHugger and IntraNought VMs.

I think that's actually acceptable, as production services can reach web-001/db-*/ directly through their dedicated network card.

Impact:

• IRC bots wouldn't be able to reach MySQL or Vault as they're still on Ysul
• Services still on Ysul can't reach
• Development servers can't reach production services, or Dwellers

Another issue is we lose connections to all those machines, as router-001 is used to route traffic to them.

Probably best to ship router-002 under FreeBSD 14 so we can prepare to minimize this impact if we need more hypervisors in the future.

router-001 is out of scope as long as T1939 is implemented

router-001 update would cut network connections between Ysul, WindRiver, CloudHugger and IntraNought VMs.