ipc: promote startup socket-wait to a shared DialWait; use in all modules
The cold-start crash-loop wasn't mavweb-specific — mavpoll and mavcaldav also ipc.Dial + exit on failure, so they crash-looped until core booted too. Moved the retry into ipc.DialWait (capped backoff, bounded) and switched mavweb, mavpoll, mavcaldav to it. mavweb's local dialCoreWithRetry is gone. Test: server appears after DialWait starts → it waits and connects. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
@@ -58,7 +58,7 @@ func run(args []string) error {
|
||||
ctx, stop := signal.NotifyContext(context.Background(), syscall.SIGINT, syscall.SIGTERM)
|
||||
defer stop()
|
||||
|
||||
core, err := ipc.Dial(*socket)
|
||||
core, err := ipc.DialWait(*socket, 60*time.Second)
|
||||
if err != nil {
|
||||
return err
|
||||
}
|
||||
|
||||
+1
-1
@@ -69,7 +69,7 @@ func run(args []string) error {
|
||||
ctx, stop := signal.NotifyContext(context.Background(), syscall.SIGINT, syscall.SIGTERM)
|
||||
defer stop()
|
||||
|
||||
core, err := ipc.Dial(*socket)
|
||||
core, err := ipc.DialWait(*socket, 60*time.Second)
|
||||
if err != nil {
|
||||
return err
|
||||
}
|
||||
|
||||
+1
-31
@@ -87,7 +87,7 @@ func main() {
|
||||
|
||||
var core ipc.CoreAPI
|
||||
if *coreSock != "" {
|
||||
c, err := dialCoreWithRetry(*coreSock, 60*time.Second)
|
||||
c, err := ipc.DialWait(*coreSock, 60*time.Second)
|
||||
if err != nil {
|
||||
log.Fatalf("dial core %s: %v", *coreSock, err)
|
||||
}
|
||||
@@ -169,36 +169,6 @@ func main() {
|
||||
}
|
||||
}
|
||||
|
||||
// dialCoreWithRetry waits for core's IPC socket to appear before giving up.
|
||||
// Core loads models on boot and (esp. under compose) may come up after mavweb;
|
||||
// depends_on only orders container start, not socket readiness. Retrying with
|
||||
// capped backoff up to timeout stops mavweb crash-looping on a cold start. A
|
||||
// core still absent past the deadline is genuinely fatal. Mid-life core
|
||||
// restarts are handled separately by ipc.Client's redial-on-drop.
|
||||
func dialCoreWithRetry(path string, timeout time.Duration) (*ipc.Client, error) {
|
||||
deadline := time.Now().Add(timeout)
|
||||
delay := 200 * time.Millisecond
|
||||
for attempt := 1; ; attempt++ {
|
||||
c, err := ipc.Dial(path)
|
||||
if err == nil {
|
||||
if attempt > 1 {
|
||||
log.Printf("core socket %s ready after %d attempts", path, attempt)
|
||||
}
|
||||
return c, nil
|
||||
}
|
||||
if time.Now().After(deadline) {
|
||||
return nil, err
|
||||
}
|
||||
if attempt == 1 {
|
||||
log.Printf("waiting for core socket %s ...", path)
|
||||
}
|
||||
time.Sleep(delay)
|
||||
if delay < 2*time.Second {
|
||||
delay *= 2
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
func handleWS(w http.ResponseWriter, r *http.Request, voiceAddr string) {
|
||||
conn, err := websocket.Accept(w, r, &websocket.AcceptOptions{
|
||||
OriginPatterns: []string{"*"},
|
||||
|
||||
Reference in New Issue
Block a user