Files
ProxMenux/oci/remote/oci_native_stack.py
T
MacRimi 1f9d2acef5 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.
2026-10-04 20:32:17 +02:00

237 lines
12 KiB
Python

#!/usr/bin/env python3
"""Instance recording adapter for the existing dedicated stack installers."""
import argparse
import copy
import ipaddress
import json
import os
from pathlib import Path
import re
import subprocess
import sys
import time
import oci_console
import oci_instances as instances
from oci_installation_state import command, image_from_archive, sha
import oci_stack_replay
from oci_ui import translate
def begin(root, primary, template, deployment, members, adapter):
ids = [int(m[1]) for m in members]
if primary not in ids or len(set(ids)) != len(ids):
raise ValueError(translate('Invalid stack members'))
resources = json.loads(command('pvesh', 'get', '/cluster/resources', '--type', 'vm', '--output-format', 'json'))
if not isinstance(resources, list):
raise ValueError(translate('Invalid Proxmox inventory'))
occupied = {int(r['vmid']) for r in resources}
for vmid in ids:
if vmid in occupied or not instances.release_orphan(root, vmid):
raise ValueError(translate('The VMID or its contract is already in use; it is not adopted'))
adapter_source = Path(adapter).read_text()
prepared = []
for name, vmid, reference, archive in members:
archive_path = archive if archive.startswith('/') else command('pvesm', 'path', archive).decode().strip()
image = image_from_archive(archive_path)
child = {'id': template['id'] + '-' + name,
'container_contract': {'image': {'reference': reference}}}
plan = {'deployment_kind': 'dedicated-stack-member', 'stack_managed': True,
'role': name, 'archive_volume': archive, 'archive_path': archive_path,
'image': image, 'rootfs_adaptation_replay_required': True,
'mounts': []}
if Path(adapter).name == 'install_immich_stack.sh' and name == 'machine-learning':
plan['machine_learning'] = copy.deepcopy(deployment.get('machine_learning', {'acceleration': 'cpu'}))
if Path(adapter).name == 'install_immich_stack.sh' and name == 'server':
# An update must know that the server transcodes with NVIDIA.
plan['video_transcoding'] = copy.deepcopy(deployment.get('video_transcoding', {'acceleration': 'cpu'}))
if Path(adapter).name in oci_stack_replay.FILES:
plan['replay_profile'] = {'adapter': Path(adapter).name, 'role': name}
if not oci_stack_replay.FILES[Path(adapter).name][name]:
plan['rootfs_replay'] = {'schema_version': 1, 'adapter': Path(adapter).name,
'role': name, 'files': []}
prepared.append((int(vmid), child, plan))
for vmid, child, plan in prepared:
instances.prepare(root, vmid, child, plan)
parent = instances.read(root, primary)
parent['native_stack_intent'] = {
'template': template, 'deployment': copy.deepcopy(deployment),
'members': [{'name': m[0], 'vmid': int(m[1])} for m in members],
'adapter': {'name': Path(adapter).name, 'sha256': sha(adapter_source.encode()),
'source': adapter_source},
}
parent['native_stack_intent']['deployment']['base_vmid'] = primary
instances.write(instances.location(root, primary), parent)
def create(root, args):
vmid = int(args[0])
record = instances.read(root, vmid)
if record['status'] != 'installing' or args[1] != record['deployment']['archive_volume']:
raise ValueError(translate('The container creation does not match the prepared instance'))
argv = list(args)
if '--description' in argv:
index = argv.index('--description')
del argv[index:index + 2]
argv += ['--description', instances.MARKER + record['installation_id']]
record['deployment']['create_arguments'] = argv
instances.write(instances.location(root, vmid), record)
# No inherited registry/network locks in long-lived Proxmox processes.
# Keep PVE extraction directories traversable inside its standard idmap.
code = subprocess.run(['pct', 'create', *argv], close_fds=True, umask=0o022).returncode
if code == 0:
# Every member keeps its console as its own log and opens a Proxmox
# console as a shell, set before the stack records its configuration
# so an update rebuilds them the same way.
oci_console.configure(vmid)
return code
def capture_rootfs(root, vmid):
record = instances.read(root, vmid)
profile = record['deployment'].get('replay_profile')
if not profile:
raise ValueError(translate('The stack member has no declared adaptation profile'))
if record['status'] != 'installing':
raise ValueError(translate('The rootfs capture only belongs to the running installation'))
mounts = []
for line in command('pct', 'config', str(vmid)).decode().splitlines():
key, sep, value = line.partition(': ')
if sep and key.startswith('mp') and key[2:].isdigit():
options = dict(p.split('=', 1) for p in value.split(',')[1:] if '=' in p)
mounts.append({'container_path': options['mp']})
record['deployment']['rootfs_replay'] = oci_stack_replay.capture(
Path('/var/lib/lxc') / str(vmid) / 'rootfs', profile['adapter'], profile['role'], mounts)
instances.write(instances.location(root, vmid), record)
PRIVATE_STACK_NETWORK = ipaddress.ip_network('10.77.0.0/16')
def access_address(vmid, wait=20):
"""The address a container is reached at from the local network. A member
of a multi-container application also has a leg on its private network,
and that address answers from the host only: it is used when the
container has no other."""
def addresses():
result = subprocess.run(['lxc-info', '-n', str(vmid), '-iH'], capture_output=True, text=True, timeout=5)
found = []
for line in result.stdout.splitlines():
try:
found.append(ipaddress.IPv4Address(line.strip()))
except ValueError:
continue
return found
def outside(found):
return next((str(address) for address in found if address not in PRIVATE_STACK_NETWORK), '')
found = addresses()
config = subprocess.run(['pct', 'config', str(vmid)], capture_output=True, text=True, timeout=30).stdout
legs = re.findall(r'^net[0-9]+: .*?(?:^|,)ip=([^,\s]+)', config, re.MULTILINE)
# A leg outside the private network may still be waiting for its lease.
expected = any(leg == 'dhcp' or (leg[:1].isdigit() and ipaddress.ip_interface(leg).ip not in PRIVATE_STACK_NETWORK)
for leg in legs)
deadline = time.monotonic() + (wait if expected else 0)
while not outside(found) and time.monotonic() < deadline:
time.sleep(2)
found = addresses()
return outside(found) or (str(found[0]) if found else '')
def finalize(root, primary):
parent = instances.read(root, primary)
intent = parent['native_stack_intent']
services = []
for member in intent['members']:
vmid = member['vmid']
record = instances.read(root, vmid)
plan = record['deployment']
from oci_description import render
presentation = copy.deepcopy(record['template'])
if vmid == primary:
stack_ui = intent['template'].get('catalog_ui') or {}
presentation['catalog_ui'] = {**stack_ui, **(presentation.get('catalog_ui') or {})}
if not (presentation.get('first_run') or {}).get('endpoints'):
presentation['first_run'] = intent['template'].get('first_run') or {}
ip = access_address(vmid)
description = render(presentation, plan['image']['manifest_digest'],
record['installation_id'], ip)
subprocess.run(['pct', 'set', str(vmid), '--description', description], check=True)
instances.finish(root, vmid, plan['archive_path'], plan['image']['manifest_digest'])
record = instances.read(root, vmid)
plan = record['deployment']
# Store raw config: repeated LXC directives must not be collapsed.
plan['native_config'] = record['observed']['config']
if plan.get('rootfs_replay'):
try:
plan['member_replay_projection'] = oci_stack_replay.normalize(record)
plan.pop('member_replay_projection_error', None)
except (ValueError, KeyError):
# Recording an unsupported projection must not roll back a
# healthy installation; it remains blocked for management.
plan.pop('member_replay_projection', None)
plan['member_replay_projection_error'] = 'native-config-requires-reviewed-conversion'
mounts = []
for line in plan['native_config'].splitlines():
key, sep, value = line.partition(': ')
if sep and key.startswith('mp') and key[2:].isdigit():
parts = value.split(',')
options = dict(p.split('=', 1) for p in parts[1:] if '=' in p)
mounts.append({'container_path': options['mp'], 'source': parts[0],
'type': 'host-bind' if parts[0].startswith('/') else 'managed-volume',
'backup': options.get('backup') == '1',
'read_only': options.get('ro') == '1', 'existing_volume': True})
plan['mounts'] = mounts
services.append({'name': member['name'], 'vmid': vmid, 'deployment': plan})
path = instances.location(root, primary).parent / 'stack-assembly.json'
if not path.exists():
instances.save_assembly(root, primary, intent['template'], intent['deployment'], services)
instances.resume_assembly(root, primary)
def main():
parser = argparse.ArgumentParser(description=__doc__)
sub = parser.add_subparsers(dest='action', required=True)
p = sub.add_parser('begin')
p.add_argument('primary', type=int)
for option in ('template', 'deployment', 'adapter'):
p.add_argument('--' + option, required=True)
p.add_argument('--member', action='append', nargs=4, required=True)
p = sub.add_parser('create')
p.add_argument('arguments', nargs=argparse.REMAINDER)
for action in ('finalize', 'failed'):
p = sub.add_parser(action)
p.add_argument('primary', type=int)
p = sub.add_parser('capture-rootfs')
p.add_argument('vmid', type=int)
args = parser.parse_args()
if os.geteuid() != 0:
parser.error(translate('Root privileges are required'))
try:
with instances.locked(instances.ROOT):
if args.action == 'begin':
begin(instances.ROOT, args.primary, json.loads(Path(args.template).read_text()),
json.loads(Path(args.deployment).read_text()), args.member, args.adapter)
elif args.action == 'create':
return create(instances.ROOT, args.arguments)
elif args.action == 'finalize':
finalize(instances.ROOT, args.primary)
elif args.action == 'capture-rootfs':
capture_rootfs(instances.ROOT, args.vmid)
else:
parent = instances.read(instances.ROOT, args.primary)
for member in parent['native_stack_intent']['members']:
record = instances.read(instances.ROOT, member['vmid'])
record['status'] = 'failed'
instances.write(instances.location(instances.ROOT, member['vmid']), record)
return 0
except (OSError, ValueError, KeyError, RuntimeError, subprocess.TimeoutExpired) as exc:
print(f"ERROR: {translate('The stack registry is incomplete; review the private contracts.')} ({exc})",
file=sys.stderr)
return 1
if __name__ == '__main__':
raise SystemExit(main())