[client] Use ctx.Err() instead of gRPC codes.Canceled to detect shutdown (#6019)

Detecting shutdown by inspecting the gRPC status code conflates a local
context cancellation with a server- or proxy-sent codes.Canceled. When
the latter occurs (e.g. an intermediary proxy resets the stream), the
retry loop silently terminates and the client never reconnects.

Switch to ctx.Err() in the signal Receive loop and management Sync/Job
handlers, and stop matching gRPC Canceled/DeadlineExceeded in the flow
client's isContextDone helper. With this change, a server-sent Canceled
is treated as a transient error and the backoff retry loop continues.
This commit is contained in:
Zoltan Papp
2026-05-04 11:59:25 +02:00
committed by GitHub
parent a21f6ecb0a
commit a547fc74ed
3 changed files with 21 additions and 35 deletions
+6 -9
View File
@@ -13,11 +13,9 @@ import (
"github.com/cenkalti/backoff/v4"
log "github.com/sirupsen/logrus"
"google.golang.org/grpc"
"google.golang.org/grpc/codes"
"google.golang.org/grpc/credentials"
"google.golang.org/grpc/credentials/insecure"
"google.golang.org/grpc/keepalive"
"google.golang.org/grpc/status"
nbgrpc "github.com/netbirdio/netbird/client/grpc"
"github.com/netbirdio/netbird/flow/proto"
@@ -301,12 +299,11 @@ func defaultBackoff(ctx context.Context, interval time.Duration) backoff.BackOff
}, ctx)
}
// isContextDone reports whether the local context has been canceled or has
// exceeded its deadline. It deliberately does not inspect gRPC status codes:
// a server- or proxy-sent codes.Canceled / codes.DeadlineExceeded must not
// short-circuit our retry loop, since retrying is the correct response when
// the local context is still alive.
func isContextDone(err error) bool {
if errors.Is(err, context.Canceled) || errors.Is(err, context.DeadlineExceeded) {
return true
}
if s, ok := status.FromError(err); ok {
return s.Code() == codes.Canceled || s.Code() == codes.DeadlineExceeded
}
return false
return errors.Is(err, context.Canceled) || errors.Is(err, context.DeadlineExceeded)
}