added proxmox lxc detection to script and fall back to --endpoint-mode dnsrr

[haouarihk]

Description

As the title suggest
This script adds compatibility support for proxmox lxcs

It also spit out a warning when the user is running on a proxmox lxc

What does it do.

• It checks for an eviroment variable string "container=lxc" inside /proc/1/environ to know if its in a proxmox lxc.
• If its not in a proxmox lxc, it does not change anything.
• If its in a proxmox lxc, it adds the --endpoint-mode dnsrr to the docker service command.

[x] Tested on a proxmox Lxc
[x] Tested on a VM

[grigoreo-fox]

I just encountered connection errors after installation because the PostgreSQL and Redis services also need --endpoint-mode dnsrr in LXC environments. The dokploy container couldn't connect to either database, throwing ECONNREFUSED errors for both ports 5432 (PostgreSQL) and 6379 (Redis).

fixed with

docker service update --endpoint-mode dnsrr dokploy-postgres
docker service update --endpoint-mode dnsrr dokploy-redis
docker service update --force dokploy

[grigoreo-fox]

Anyway every new application deployed still defaults to VIP endpoint mode, even when running in LXC environments.

For example, I just deployed n8n, and it was created with vip mode, resulting in a "bad gateway" error because Traefik can't reach the service.