grpc: don't tear down the shared master connection on a caller's own timeout (#9775)

A Canceled/DeadlineExceeded from the caller's per-request context was
treated like a dead channel: it closed the shared cached ClientConn and
cancelled every other in-flight RPC on it with "the client connection is
closing". Under a burst of concurrent chunk assigns (e.g. a large S3
multipart upload) one slow assign hitting its 10s attempt timeout could
poison the connection for all the rest, cascading into a flood of 500s.

Thread the caller's context into shouldInvalidateConnection and only
invalidate on Canceled/DeadlineExceeded while that context is still live,
which isolates the genuine stale-channel signal (a peer restart behind a
k8s Service VIP). To carry the context, add a ctx parameter to the
existing WithGrpcClient, WithMasterClient, and WithMasterServerClient; the
master assign and volume-lookup paths pass their per-attempt context and
every other caller passes context.Background().
This commit is contained in:
Chris Lu authored and GitHub committed 2026-06-01 15:11:02 -07:00
1 parent dfa86b4313
commit 2386fa550a
31 files changed
+129 -63

No files matched your search

+4 -3
View File
@@ -1,6 +1,7 @@
package s3api
import (
"context"
"encoding/base64"
"errors"
"fmt"
@@ -24,7 +25,7 @@ func (s3a *S3ApiServer) WithFilerClient(streamingMode bool, fn func(filer_pb.Sea
// Fallback to direct connection if filerClient not initialized
// This should only happen during initialization or testing
return pb.WithGrpcClient(streamingMode, s3a.randomClientId, func(grpcConnection *grpc.ClientConn) error {
return pb.WithGrpcClient(context.Background(), streamingMode, s3a.randomClientId, func(grpcConnection *grpc.ClientConn) error {
client := filer_pb.NewSeaweedFilerClient(grpcConnection)
return fn(client)
}, s3a.getFilerAddress().ToGrpcAddress(), false, s3a.option.GrpcDialOption)
@@ -37,7 +38,7 @@ func (s3a *S3ApiServer) withFilerClientFailover(streamingMode bool, fn func(file
currentFiler := s3a.filerClient.GetCurrentFiler()
// Try current filer first (fast path)
err := pb.WithGrpcClient(streamingMode, s3a.randomClientId, func(grpcConnection *grpc.ClientConn) error {
err := pb.WithGrpcClient(context.Background(), streamingMode, s3a.randomClientId, func(grpcConnection *grpc.ClientConn) error {
client := filer_pb.NewSeaweedFilerClient(grpcConnection)
return fn(client)
}, currentFiler.ToGrpcAddress(), false, s3a.option.GrpcDialOption)
@@ -70,7 +71,7 @@ func (s3a *S3ApiServer) withFilerClientFailover(streamingMode bool, fn func(file
continue
}
err = pb.WithGrpcClient(streamingMode, s3a.randomClientId, func(grpcConnection *grpc.ClientConn) error {
err = pb.WithGrpcClient(context.Background(), streamingMode, s3a.randomClientId, func(grpcConnection *grpc.ClientConn) error {
client := filer_pb.NewSeaweedFilerClient(grpcConnection)
return fn(client)
}, filer.ToGrpcAddress(), false, s3a.option.GrpcDialOption)