dearwolf/LocalAI
1
0
Fork 0
mirror of https://github.com/mudler/LocalAI.git synced 2025-05-20 02:24:59 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 21

chore(refactor): drop duplicated shutdown logics (#3589)

Browse source 
* chore(refactor): drop duplicated shutdown logics

- Handle locking in Shutdown and CheckModelIsLoaded in a more go-idiomatic way
- Drop duplicated code and re-organize shutdown code

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

* fix: drop leftover

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

* chore: improve logging and add missing locks

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

---------

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
This commit is contained in:
Ettore Di Giacinto 2024-09-17 16:51:40 +02:00 • committed by GitHub
parent eee1fb2c75
commit a53392f919
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
5 changed files with 32 additions and 38 deletions
Download patch file Download diff file Expand all files Collapse all files

16
pkg/model/initializers.go
View file

@ -320,7 +320,7 @@ func (ml *ModelLoader) grpcModel(backend string, o *Options) func(string, string
} else {
grpcProcess := backendPath(o.assetDir, backend)
if err := utils.VerifyPath(grpcProcess, o.assetDir); err != nil {
return nil, fmt.Errorf("grpc process not found in assetdir: %s", err.Error())
return nil, fmt.Errorf("refering to a backend not in asset dir: %s", err.Error())
}
if autoDetect {
@ -332,7 +332,7 @@ func (ml *ModelLoader) grpcModel(backend string, o *Options) func(string, string
// Check if the file exists
if _, err := os.Stat(grpcProcess); os.IsNotExist(err) {
return nil, fmt.Errorf("grpc process not found: %s. some backends(stablediffusion, tts) require LocalAI compiled with GO_TAGS", grpcProcess)
return nil, fmt.Errorf("backend not found: %s", grpcProcess)
}
serverAddress, err := getFreeAddress()
@ -355,6 +355,8 @@ func (ml *ModelLoader) grpcModel(backend string, o *Options) func(string, string
client = NewModel(serverAddress)
}
log.Debug().Msgf("Wait for the service to start up")
// Wait for the service to start up
ready := false
for i := 0; i < o.grpcAttempts; i++ {
@ -413,10 +415,8 @@ func (ml *ModelLoader) BackendLoader(opts ...Option) (client grpc.Backend, err e
}
if o.singleActiveBackend {
ml.mu.Lock()
log.Debug().Msgf("Stopping all backends except '%s'", o.model)
err := ml.StopAllExcept(o.model)
ml.mu.Unlock()
err := ml.StopGRPC(allExcept(o.model))
if err != nil {
log.Error().Err(err).Str("keptModel", o.model).Msg("error while shutting down all backends except for the keptModel")
return nil, err
@ -444,13 +444,10 @@ func (ml *ModelLoader) BackendLoader(opts ...Option) (client grpc.Backend, err e
func (ml *ModelLoader) GreedyLoader(opts ...Option) (grpc.Backend, error) {
o := NewOptions(opts...)
ml.mu.Lock()
// Return earlier if we have a model already loaded
// (avoid looping through all the backends)
if m := ml.CheckIsLoaded(o.model); m != nil {
log.Debug().Msgf("Model '%s' already loaded", o.model)
ml.mu.Unlock()
return m.GRPC(o.parallelRequests, ml.wd), nil
}
@ -458,12 +455,11 @@ func (ml *ModelLoader) GreedyLoader(opts ...Option) (grpc.Backend, error) {
// If we can have only one backend active, kill all the others (except external backends)
if o.singleActiveBackend {
log.Debug().Msgf("Stopping all backends except '%s'", o.model)
err := ml.StopAllExcept(o.model)
err := ml.StopGRPC(allExcept(o.model))
if err != nil {
log.Error().Err(err).Str("keptModel", o.model).Msg("error while shutting down all backends except for the keptModel - greedyloader continuing")
}
}
ml.mu.Unlock()
var err error