feat(oci): recover applications after a Proxmox reinstall, cluster records and AMD GPU profiles

The installation record of an OCI application travels with its container:
a copy inside the container and another in /etc/pve, written together
with the one kept on the host. A container restored on a newly installed
Proxmox, restored with another ID or moved to another node of a cluster
is recognised and registered again, with its private network, hookscript,
Rclone mount, host firewall rule and NVIDIA runtime. The Monitor offers
the same recovery from the Updates tab.

AMD GPUs are offered by generation. A GPU the ROCm image supports takes
the profile as it is; one of a supported family (Radeon 680M, 780M) is an
experimental option that asks for confirmation and is never proposed; an
older one is not offered. The GPU is checked with a real inference before
the installation accepts it. Recreate changes what runs recognition in an
installed Immich, between the CPU and a GPU of the host.

Updates:
- A failed update that is restored and checked removes its temporary
  container and the disks of the failed attempt.
- Every container volume is part of the backups, so Jellyfin, Plex and
  Hugo update with their default installation.
- An image published with a Docker-format manifest is recognised by its
  layers and build time and updates.
- The Proxmox notes of a multi-container application link to its LAN
  address.
This commit is contained in:
MacRimi
2026-10-04 20:32:17 +02:00
parent 7e9f16fdb3
commit 1f9d2acef5
67 changed files with 4464 additions and 126 deletions
@@ -61,6 +61,7 @@
"rows": [
["Intel/AMD DRM", "<code>/dev/dri/renderD*</code> y <code>/sys/class/drm/NODO/device/vendor</code>", "Un dispositivo de caracteres con fabricante <code>0x8086</code> (Intel) o <code>0x1002</code> (AMD)"],
["OpenCL AMD", "El render node, más <code>/dev/kfd</code> cuando el perfil lo necesita", "Existencia, tipo, fabricante, permisos y compatibilidad declarada"],
["AMD ROCm", "El nodo de render, <code>/dev/kfd</code> y la generación que declara el driver de cómputo en <code>/sys/class/kfd/kfd/topology/nodes/*/properties</code>", "Una generación para la que la imagen ROCm de la aplicación trae kernels, las GPU RDNA2 que incluye y las posteriores, se ofrece tal cual. Las demás GPU de esas familias, como las Radeon 680M y 780M, se ofrecen como experimentales y nunca son la opción propuesta: ROCm no las soporta oficialmente, la aplicación las presenta como la generación de su familia (<code>HSA_OVERRIDE_GFX_VERSION</code>) y el instalador pide confirmación antes de descargar la imagen, que es más grande. Una GPU anterior, como los gráficos integrados de un Ryzen 5000U, no se ofrece e Immich deja el reconocimiento en la CPU"],
["NVIDIA", "<code>nvidia-smi</code> y <code>nvidia-container-cli</code>", "GPU, UUID, bus PCI, versión del driver, Toolkit, nodos <code>/dev/nvidia*</code>, binarios y librerías"],
["Coral PCIe/M.2", "<code>/dev/apex_N</code> y su enlace en <code>/sys/dev/char/MAJOR:MINOR</code>", "Nodo de caracteres, major/minor, propietario, GID y permisos"],
["USB y serie", "<code>/dev/ttyUSB*</code>, <code>/dev/ttyACM*</code> o <code>/dev/bus/usb/BBB/DDD</code>", "Nodo de caracteres; en USB también fabricante, producto y número de serie cuando sysfs los publica"],