popov/neuron-tai
1
0
Fork 0
Code Issues 42 Pull Requests Actions Packages Projects Releases Wiki Activity

fix: reconcile merged runtime behavior

Browse Source
This commit is contained in:
Dobromir Popov
2026-07-17 14:29:09 +03:00
parent 6aced6a005
commit f0197bfa83
3 changed files with 45 additions and 58 deletions
Download Patch File Download Diff File Expand all files Collapse all files

50
packages/node/meshnet_node/doctor.py
View File

@@ -36,7 +36,6 @@ from .capability import (
CapabilityReport,
build_capability_report,
)
from .native_backend import NativeWorkerBackendAdapter
from .recipe_manifest import (
DEFAULT_RECIPE_ID,
Recipe,
@@ -450,9 +449,11 @@ def _validate_recipe(
category: str | None = None
error: BaseException | None = None
diagnostics: list[str] = []
detail: dict = {}
try:
backend = load_backend(selection, recipe)
probe_forward(backend)
detail = probe_forward(backend)
except DoctorError as exc:
category, error = exc.category, exc
diagnostics = [str(exc), exc.hint]
@@ -463,48 +464,23 @@ def _validate_recipe(
duration_ms = int((time.monotonic() - started) * 1000)
device = _backend_device(backend, selection)
# Only the native adapter has an authoritative immutable GGUF report and
# deployment pin. The Transformers path deliberately remains dark: a
# model/config fingerprint is not an exact ArtifactIdentity.
identity = backend.identity if isinstance(backend, NativeWorkerBackendAdapter) else None
model_id = selection.model_id if identity is None else identity.artifact.artifact_id
shard_start = selection.shard_start if identity is None else identity.shard_start
shard_end = selection.shard_end if identity is None else identity.shard_end - 1
recipe_id = recipe.id if identity is None else identity.recipe.recipe_id
recipe_version = recipe.version if identity is None else identity.recipe.recipe_version
catalogue_version = (
manifest.catalogue_version if identity is None else identity.recipe.catalogue_version
)
backend_id = recipe.backend_id if identity is None else identity.recipe.backend_id
quantization = (
selection.quantization if identity is None else identity.recipe.weight_quantization
)
runtime = _runtime_versions()
model_config = _model_config(backend)
revision = None
if identity is not None:
revision = identity.artifact.revision
model_config = "sha256:" + identity.artifact.architecture_digest
runtime = {**runtime, "native_runtime": identity.recipe.runtime_version}
report = build_capability_report(
model_id=model_id,
shard_start=shard_start,
shard_end=shard_end,
recipe_id=recipe_id,
recipe_version=recipe_version,
catalogue_version=catalogue_version,
backend_id=backend_id,
model_id=selection.model_id,
shard_start=selection.shard_start,
shard_end=selection.shard_end,
recipe_id=recipe.id,
recipe_version=recipe.version,
catalogue_version=manifest.catalogue_version,
backend_id=recipe.backend_id,
device=device,
device_name=_backend_device_name(device),
quantization=quantization,
runtime=runtime,
revision=revision,
model_config=model_config,
quantization=selection.quantization,
runtime=_runtime_versions(),
model_config=_model_config(backend),
status=STATUS_FAILED if category else STATUS_PASSED,
duration_ms=duration_ms,
diagnostics=[d for d in diagnostics if d] or None,
validated_at=clock(),
identity=identity,
)
if category:
return RecipeResult(

18
packages/node/meshnet_node/torch_server.py
View File

@@ -1576,8 +1576,7 @@ class TorchNodeServer:
if drop_directive is not None:
model_id = str(drop_directive.get("model") or "")
removed = self._backends.pop(model_id, None)
if removed is None:
return None
if removed is not None:
if self._backends:
self._backend = next(iter(self._backends.values()))
self._tracker_mode = self._backend.shard_start == 0
@@ -1588,7 +1587,13 @@ class TorchNodeServer:
self._server.backends = dict(self._backends)
self._server.backend = self._backend
self._server.tracker_mode = self._tracker_mode
return {"action": "DROP_SHARD", "model": model_id}
# A DROP_SHARD may be paired with a subsequent LOAD_SHARD in the
# same heartbeat; continue so the replacement is applied atomically.
if not any(
directive.get("action") in {"ADD_SHARD", "LOAD_SHARD"}
for directive in directives
):
return {"action": "DROP_SHARD", "model": model_id} if removed is not None else None
add_directive = next(
(directive for directive in reversed(directives) if directive.get("action") == "ADD_SHARD"),
None,
@@ -1602,8 +1607,11 @@ class TorchNodeServer:
return None
shard_start = int(directive["shard_start"])
shard_end = int(directive["shard_end"])
quantization = str(directive.get("quantization") or self._backend.quantization)
model_id = str(directive.get("model") or self._backend.model_id)
quantization = str(
directive.get("quantization")
or (self._backend.quantization if self._backend is not None else "auto")
)
model_id = str(directive.get("model") or (self._backend.model_id if self._backend is not None else ""))
replacing = directive.get("action") == "LOAD_SHARD"
if not replacing and len(self._backends) >= self._max_loaded_shards:
print(

13
packages/tracker/meshnet_tracker/server.py
View File

@@ -3295,12 +3295,15 @@ class _TrackerHandler(http.server.BaseHTTPRequestHandler):
node.hf_repo or node.model
for node in alive
if node.model is not None
# The same model can be registered under its HF repository while
# the catalogue exposes its short preset id. Do not emit a second
# repo-keyed entry when either node identifier resolves to a preset.
and _resolve_model_preset(
server.model_presets, node.hf_repo or node.model,
# Explicit HF repositories are emitted as stable identifiers even
# when they also resolve to a short-name preset; clients may use
# either identifier when selecting a model.
and (
node.hf_repo is not None
or _resolve_model_preset(
server.model_presets, node.model,
)[1] is None
)
and node.shard_start is not None
and node.shard_end is not None
and node.num_layers is not None