fix(swarm): relocalize master-built config on worker before deploy
deploy --mode swarm was failing on every heterogeneous fleet: the master populates config.interface from its own box (detect_interface() → its default NIC), then ships that verbatim. The worker's deployer then calls get_host_ip(config.interface), hits 'ip addr show wlp6s0' on a VM whose NIC is enp0s3, and 500s. Fix: agent.executor._relocalize() runs on every swarm-mode deploy. Re-detects the worker's interface/subnet/gateway/host_ip locally and swaps them into the config before calling deployer.deploy(). When the worker's subnet doesn't match the master's, decky IPs are re-allocated from the worker's subnet via allocate_ips() so they're reachable. Unihost-mode configs are left untouched — they're already built against the local box and second-guessing them would be wrong. Validated against anti@192.168.1.13: master dispatched interface=wlp6s0, agent logged 'relocalized interface=enp0s3', deployer ran successfully, dry-run returned ok=deployed. 4 new tests cover both branches (matching-subnet preserves decky IPs; mismatch re-allocates), the end-to-end executor.deploy() path, and the unihost short-circuit.
This commit is contained in:
@@ -9,22 +9,74 @@ blocking) so the FastAPI event loop stays responsive.
|
||||
from __future__ import annotations
|
||||
|
||||
import asyncio
|
||||
from ipaddress import IPv4Network
|
||||
from typing import Any
|
||||
|
||||
from decnet.engine import deployer as _deployer
|
||||
from decnet.config import DecnetConfig, load_state, clear_state
|
||||
from decnet.logging import get_logger
|
||||
from decnet.network import (
|
||||
allocate_ips,
|
||||
detect_interface,
|
||||
detect_subnet,
|
||||
get_host_ip,
|
||||
)
|
||||
|
||||
log = get_logger("agent.executor")
|
||||
|
||||
|
||||
def _relocalize(config: DecnetConfig) -> DecnetConfig:
|
||||
"""Rewrite a master-built config to the worker's local network reality.
|
||||
|
||||
The master populates ``interface``/``subnet``/``gateway`` from its own
|
||||
box before dispatching, which blows up the deployer on any worker whose
|
||||
NIC name differs (common in heterogeneous fleets — master on ``wlp6s0``,
|
||||
worker on ``enp0s3``). We always re-detect locally; if the worker sits
|
||||
on a different subnet than the master, decky IPs are re-allocated from
|
||||
the worker's subnet so they're actually reachable.
|
||||
"""
|
||||
local_iface = detect_interface()
|
||||
local_subnet, local_gateway = detect_subnet(local_iface)
|
||||
local_host_ip = get_host_ip(local_iface)
|
||||
|
||||
updates: dict[str, Any] = {
|
||||
"interface": local_iface,
|
||||
"subnet": local_subnet,
|
||||
"gateway": local_gateway,
|
||||
}
|
||||
|
||||
master_net = IPv4Network(config.subnet, strict=False) if config.subnet else None
|
||||
local_net = IPv4Network(local_subnet, strict=False)
|
||||
if master_net is None or master_net != local_net:
|
||||
log.info(
|
||||
"agent.deploy subnet mismatch master=%s local=%s — re-allocating decky IPs",
|
||||
config.subnet, local_subnet,
|
||||
)
|
||||
fresh_ips = allocate_ips(
|
||||
subnet=local_subnet,
|
||||
gateway=local_gateway,
|
||||
host_ip=local_host_ip,
|
||||
count=len(config.deckies),
|
||||
)
|
||||
new_deckies = [d.model_copy(update={"ip": ip}) for d, ip in zip(config.deckies, fresh_ips)]
|
||||
updates["deckies"] = new_deckies
|
||||
|
||||
return config.model_copy(update=updates)
|
||||
|
||||
|
||||
async def deploy(config: DecnetConfig, dry_run: bool = False, no_cache: bool = False) -> None:
|
||||
"""Run the blocking deployer off-loop. The deployer itself calls
|
||||
save_state() internally once the compose file is materialised."""
|
||||
log.info(
|
||||
"agent.deploy mode=%s deckies=%d interface=%s",
|
||||
"agent.deploy mode=%s deckies=%d interface=%s (incoming)",
|
||||
config.mode, len(config.deckies), config.interface,
|
||||
)
|
||||
if config.mode == "swarm":
|
||||
config = _relocalize(config)
|
||||
log.info(
|
||||
"agent.deploy relocalized interface=%s subnet=%s gateway=%s",
|
||||
config.interface, config.subnet, config.gateway,
|
||||
)
|
||||
await asyncio.to_thread(_deployer.deploy, config, dry_run, no_cache, False)
|
||||
|
||||
|
||||
|
||||
Reference in New Issue
Block a user