Files
LocalAI/pkg/model/connection_evicting_client.go
T
Ettore Di Giacinto d26263f9c0 fix(distributed): let a worker's own refusal be evidence about its backend
A worker that refuses a stream has answered, and cluster.Dial keeps the three
tunnelproto sentinels out of the ErrNoRoute umbrella precisely so a consumer
can act on that. No consumer did. Since workers stopped listening, a backend
process that crashed on a healthy worker is no longer a dead listener's
codes.Unavailable: the worker refuses the stream with
ErrStreamTargetUnavailable, gRPC flattens it into Unavailable anyway, and
nodes.unroutable reported the whole thing as "this frontend has no route".
Every reap path then answered ProbeUnknown and left the row, so the replica
slot never freed and at the default MaxReplicasPerModel=1 the only cleanup
left was LRU eviction of models that were working.

isWorkerAnswer is exported as cluster.IsWorkerAnswer, so the errors the dialer
keeps out of the umbrella are by construction the errors the consumers treat
as the worker answering. nodes.unroutable and pkg/model's transportFailure
both use it; ConnectionEvictingClient, the site reached during inference, goes
through transportFailure rather than asking the transport directly. A reply
code this frontend does not recognise is still not an answer, so a newer
worker's vocabulary costs a retry and not a replica.

The reap guards keep the allow-list rather than requiring ErrNoRoute: an
unrecognised dial error must mean "no route", never "the backend is gone".

Also in this final pass over the branch:

- Docs: recommend upgrading FRONTENDS first, with the symptom of each order.
  Workers-first fails now that a 4xx registration is a verdict rather than an
  outage, so an old frontend's "address is required for backend workers" makes
  each restarted worker exit and drains the fleet a node per restart.
- Docs: LOCALAI_WORKER_TUNNEL=false is a fatal startup error, not a degraded
  mode, in both places that described it; and a frontend rollback needs every
  worker restarted, because re-registration force-clears the address columns.
- A replica with no advertised address now says so every five minutes and
  names the workers only it can reach, instead of one startup warning for a
  cost paid for the life of the process.
- callerRanOut's rule now holds at all three siblings, so an expired caller
  deadline stops reading as a broken tunnel; probeHealth's withdrawn reason
  for using the raw client is corrected; the dead DoOrCached is deleted and
  its coverage kept on DoOrCachedResult; sweepLeakedInFlight enumerates the
  outcomes that reach it.
- The peer route's self-declared id is recorded as a phase-3 deferral, in the
  handler, in the isolation claim it narrows, and in the operator docs.

Assisted-by: Claude Opus 5 [claude-code]
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
2026-09-02 00:59:36 +00:00

163 lines
6.2 KiB
Go

package model
import (
"context"
"sync"
grpc "github.com/mudler/LocalAI/pkg/grpc"
pb "github.com/mudler/LocalAI/pkg/grpc/proto"
"github.com/mudler/xlog"
ggrpc "google.golang.org/grpc"
)
// ConnectionEvictingClient wraps a grpc.Backend. When any inference method
// fails with a connection error (server unreachable), it calls the evict
// callback to remove the model from the ModelLoader's cache. The error is
// still returned to the caller — the NEXT request will trigger rescheduling
// via SmartRouter.
type ConnectionEvictingClient struct {
grpc.WrappedBackend
modelID string
evict func()
once sync.Once
}
var _ grpc.BackendUnwrapper = (*ConnectionEvictingClient)(nil)
func newConnectionEvictingClient(inner grpc.Backend, modelID string, evict func()) grpc.Backend {
return &ConnectionEvictingClient{
WrappedBackend: grpc.WrappedBackend{Backend: inner},
modelID: modelID,
evict: evict,
}
}
func (c *ConnectionEvictingClient) checkErr(err error) {
if err == nil || !isConnectionError(err) {
return
}
// The fifth site of the same shape, and the one reached during INFERENCE
// rather than a health check. evict() runs ShutdownModel, which for a remote
// model sends backend.stop over NATS to every node holding it and deletes
// every replica row. In distributed mode the client underneath reaches the
// backend over the worker's tunnel, and a failure of THAT transport arrives
// as the same codes.Unavailable a dead backend produces; evicting on it
// stops a model that is loaded and serving, on a worker that is
// heartbeating. A locally spawned backend has no custom transport, so this
// reports nil and the behaviour there is exactly what it always was.
// transportFailure and not LastDialErrorOf: a refusal the WORKER wrote is
// the worker answering that it could not reach the process, which is what a
// crashed backend produces now that a worker listens on nothing. Treating
// that as a transport failure kept a genuinely dead model loaded and
// failing every request, which is the mirror image of the mistake this
// guard exists to prevent.
if dialErr := transportFailure(c.Backend); dialErr != nil {
xlog.Warn("Inference failed because the worker could not be reached; keeping the model",
"model", c.modelID, "error", dialErr)
return
}
c.once.Do(func() {
xlog.Warn("Connection error during inference, evicting model from cache",
"model", c.modelID, "error", err)
c.evict()
})
}
// --- Intercepted inference methods ---
func (c *ConnectionEvictingClient) Predict(ctx context.Context, in *pb.PredictOptions, opts ...ggrpc.CallOption) (*pb.Reply, error) {
reply, err := c.Backend.Predict(ctx, in, opts...)
c.checkErr(err)
return reply, err
}
func (c *ConnectionEvictingClient) PredictStream(ctx context.Context, in *pb.PredictOptions, f func(reply *pb.Reply), opts ...ggrpc.CallOption) error {
err := c.Backend.PredictStream(ctx, in, f, opts...)
c.checkErr(err)
return err
}
func (c *ConnectionEvictingClient) Embeddings(ctx context.Context, in *pb.PredictOptions, opts ...ggrpc.CallOption) (*pb.EmbeddingResult, error) {
result, err := c.Backend.Embeddings(ctx, in, opts...)
c.checkErr(err)
return result, err
}
func (c *ConnectionEvictingClient) GenerateImage(ctx context.Context, in *pb.GenerateImageRequest, opts ...ggrpc.CallOption) (*pb.Result, error) {
result, err := c.Backend.GenerateImage(ctx, in, opts...)
c.checkErr(err)
return result, err
}
func (c *ConnectionEvictingClient) GenerateVideo(ctx context.Context, in *pb.GenerateVideoRequest, opts ...ggrpc.CallOption) (*pb.Result, error) {
result, err := c.Backend.GenerateVideo(ctx, in, opts...)
c.checkErr(err)
return result, err
}
func (c *ConnectionEvictingClient) Generate3D(ctx context.Context, in *pb.Generate3DRequest, opts ...ggrpc.CallOption) (*pb.Result, error) {
result, err := c.Backend.Generate3D(ctx, in, opts...)
c.checkErr(err)
return result, err
}
func (c *ConnectionEvictingClient) TTS(ctx context.Context, in *pb.TTSRequest, opts ...ggrpc.CallOption) (*pb.Result, error) {
result, err := c.Backend.TTS(ctx, in, opts...)
c.checkErr(err)
return result, err
}
func (c *ConnectionEvictingClient) TTSStream(ctx context.Context, in *pb.TTSRequest, f func(reply *pb.Reply), opts ...ggrpc.CallOption) error {
err := c.Backend.TTSStream(ctx, in, f, opts...)
c.checkErr(err)
return err
}
func (c *ConnectionEvictingClient) SoundGeneration(ctx context.Context, in *pb.SoundGenerationRequest, opts ...ggrpc.CallOption) (*pb.Result, error) {
result, err := c.Backend.SoundGeneration(ctx, in, opts...)
c.checkErr(err)
return result, err
}
func (c *ConnectionEvictingClient) AudioTranscription(ctx context.Context, in *pb.TranscriptRequest, opts ...ggrpc.CallOption) (*pb.TranscriptResult, error) {
result, err := c.Backend.AudioTranscription(ctx, in, opts...)
c.checkErr(err)
return result, err
}
func (c *ConnectionEvictingClient) AudioTranscriptionStream(ctx context.Context, in *pb.TranscriptRequest, f func(chunk *pb.TranscriptStreamResponse), opts ...ggrpc.CallOption) error {
err := c.Backend.AudioTranscriptionStream(ctx, in, f, opts...)
c.checkErr(err)
return err
}
func (c *ConnectionEvictingClient) Detect(ctx context.Context, in *pb.DetectOptions, opts ...ggrpc.CallOption) (*pb.DetectResponse, error) {
result, err := c.Backend.Detect(ctx, in, opts...)
c.checkErr(err)
return result, err
}
func (c *ConnectionEvictingClient) Depth(ctx context.Context, in *pb.DepthRequest, opts ...ggrpc.CallOption) (*pb.DepthResponse, error) {
result, err := c.Backend.Depth(ctx, in, opts...)
c.checkErr(err)
return result, err
}
func (c *ConnectionEvictingClient) Rerank(ctx context.Context, in *pb.RerankRequest, opts ...ggrpc.CallOption) (*pb.RerankResult, error) {
result, err := c.Backend.Rerank(ctx, in, opts...)
c.checkErr(err)
return result, err
}
func (c *ConnectionEvictingClient) TokenClassify(ctx context.Context, in *pb.TokenClassifyRequest, opts ...ggrpc.CallOption) (*pb.TokenClassifyResponse, error) {
result, err := c.Backend.TokenClassify(ctx, in, opts...)
c.checkErr(err)
return result, err
}
func (c *ConnectionEvictingClient) Score(ctx context.Context, in *pb.ScoreRequest, opts ...ggrpc.CallOption) (*pb.ScoreResponse, error) {
result, err := c.Backend.Score(ctx, in, opts...)
c.checkErr(err)
return result, err
}