Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

clientv3: fix lease "freezing" on unhealthy cluster #7023

Merged
merged 3 commits into from
Dec 16, 2016
Merged

clientv3: fix lease "freezing" on unhealthy cluster #7023

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
87b1d95
v3api, rpctypes: add ErrTimeoutDueToConnectionLost
Dec 16, 2016
46bd842
clientv3/integration: test lease grant/keepalive with/without failures
Dec 16, 2016
a375e91
clientv3: don't reset keepalive stream on grant failure
Dec 16, 2016
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
56 changes: 56 additions & 0 deletions clientv3/integration/lease_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,10 +17,12 @@ package integration
import (
"reflect"
"sort"
"sync"
"testing"
"time"

"github.com/coreos/etcd/clientv3"
"github.com/coreos/etcd/clientv3/concurrency"
"github.com/coreos/etcd/etcdserver/api/v3rpc/rpctypes"
"github.com/coreos/etcd/integration"
"github.com/coreos/etcd/pkg/testutil"
Expand Down Expand Up @@ -574,3 +576,57 @@ func TestLeaseKeepAliveLoopExit(t *testing.T) {
t.Fatalf("expected %T, got %v(%T)", clientv3.ErrKeepAliveHalted{}, err, err)
}
}

// TestV3LeaseFailureOverlap issues Grant and Keepalive requests to a cluster
// before, during, and after quorum loss to confirm Grant/Keepalive tolerates
// transient cluster failure.
func TestV3LeaseFailureOverlap(t *testing.T) {
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It will be nice to add some explanation to why we need this test in the file. so the future reader might have a context.

Other than that LGTM.

clus := integration.NewClusterV3(t, &integration.ClusterConfig{Size: 2})
defer clus.Terminate(t)

numReqs := 5
cli := clus.Client(0)

// bring up a session, tear it down
updown := func(i int) error {
sess, err := concurrency.NewSession(cli)
if err != nil {
return err
}
ch := make(chan struct{})
go func() {
defer close(ch)
sess.Close()
}()
select {
case <-ch:
case <-time.After(time.Minute / 4):
t.Fatalf("timeout %d", i)
}
return nil
}

var wg sync.WaitGroup
mkReqs := func(n int) {
wg.Add(numReqs)
for i := 0; i < numReqs; i++ {
go func() {
defer wg.Done()
err := updown(n)
if err == nil || err == rpctypes.ErrTimeoutDueToConnectionLost {
return
}
t.Fatal(err)
}()
}
}

mkReqs(1)
clus.Members[1].Stop(t)
mkReqs(2)
time.Sleep(time.Second)
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Q. Why do we wait here?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

otherwise the leader resumes too fast; couldn't trigger the failure without it

mkReqs(3)
clus.Members[1].Restart(t)
mkReqs(4)
wg.Wait()
}
3 changes: 0 additions & 3 deletions clientv3/lease.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -177,9 +177,6 @@ func (l *lessor) Grant(ctx context.Context, ttl int64) (*LeaseGrantResponse, err
if isHaltErr(cctx, err) {
return nil, toErr(cctx, err)
}
if nerr := l.newStream(); nerr != nil {
return nil, nerr
}
}
}

Expand Down
39 changes: 21 additions & 18 deletions etcdserver/api/v3rpc/rpctypes/error.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -52,12 +52,13 @@ var (
ErrGRPCPermissionNotGranted = grpc.Errorf(codes.FailedPrecondition, "etcdserver: permission is not granted to the role")
ErrGRPCAuthNotEnabled = grpc.Errorf(codes.FailedPrecondition, "etcdserver: authentication is not enabled")

ErrGRPCNoLeader = grpc.Errorf(codes.Unavailable, "etcdserver: no leader")
ErrGRPCNotCapable = grpc.Errorf(codes.Unavailable, "etcdserver: not capable")
ErrGRPCStopped = grpc.Errorf(codes.Unavailable, "etcdserver: server stopped")
ErrGRPCTimeout = grpc.Errorf(codes.Unavailable, "etcdserver: request timed out")
ErrGRPCTimeoutDueToLeaderFail = grpc.Errorf(codes.Unavailable, "etcdserver: request timed out, possibly due to previous leader failure")
ErrGRPCUnhealthy = grpc.Errorf(codes.Unavailable, "etcdserver: unhealthy cluster")
ErrGRPCNoLeader = grpc.Errorf(codes.Unavailable, "etcdserver: no leader")
ErrGRPCNotCapable = grpc.Errorf(codes.Unavailable, "etcdserver: not capable")
ErrGRPCStopped = grpc.Errorf(codes.Unavailable, "etcdserver: server stopped")
ErrGRPCTimeout = grpc.Errorf(codes.Unavailable, "etcdserver: request timed out")
ErrGRPCTimeoutDueToLeaderFail = grpc.Errorf(codes.Unavailable, "etcdserver: request timed out, possibly due to previous leader failure")
ErrGRPCTimeoutDueToConnectionLost = grpc.Errorf(codes.Unavailable, "etcdserver: request timed out, possibly due to connection lost")
ErrGRPCUnhealthy = grpc.Errorf(codes.Unavailable, "etcdserver: unhealthy cluster")

errStringToError = map[string]error{
grpc.ErrorDesc(ErrGRPCEmptyKey): ErrGRPCEmptyKey,
Expand Down Expand Up @@ -91,12 +92,13 @@ var (
grpc.ErrorDesc(ErrGRPCPermissionNotGranted): ErrGRPCPermissionNotGranted,
grpc.ErrorDesc(ErrGRPCAuthNotEnabled): ErrGRPCAuthNotEnabled,

grpc.ErrorDesc(ErrGRPCNoLeader): ErrGRPCNoLeader,
grpc.ErrorDesc(ErrGRPCNotCapable): ErrGRPCNotCapable,
grpc.ErrorDesc(ErrGRPCStopped): ErrGRPCStopped,
grpc.ErrorDesc(ErrGRPCTimeout): ErrGRPCTimeout,
grpc.ErrorDesc(ErrGRPCTimeoutDueToLeaderFail): ErrGRPCTimeoutDueToLeaderFail,
grpc.ErrorDesc(ErrGRPCUnhealthy): ErrGRPCUnhealthy,
grpc.ErrorDesc(ErrGRPCNoLeader): ErrGRPCNoLeader,
grpc.ErrorDesc(ErrGRPCNotCapable): ErrGRPCNotCapable,
grpc.ErrorDesc(ErrGRPCStopped): ErrGRPCStopped,
grpc.ErrorDesc(ErrGRPCTimeout): ErrGRPCTimeout,
grpc.ErrorDesc(ErrGRPCTimeoutDueToLeaderFail): ErrGRPCTimeoutDueToLeaderFail,
grpc.ErrorDesc(ErrGRPCTimeoutDueToConnectionLost): ErrGRPCTimeoutDueToConnectionLost,
grpc.ErrorDesc(ErrGRPCUnhealthy): ErrGRPCUnhealthy,
}

// client-side error
Expand Down Expand Up @@ -131,12 +133,13 @@ var (
ErrPermissionNotGranted = Error(ErrGRPCPermissionNotGranted)
ErrAuthNotEnabled = Error(ErrGRPCAuthNotEnabled)

ErrNoLeader = Error(ErrGRPCNoLeader)
ErrNotCapable = Error(ErrGRPCNotCapable)
ErrStopped = Error(ErrGRPCStopped)
ErrTimeout = Error(ErrGRPCTimeout)
ErrTimeoutDueToLeaderFail = Error(ErrGRPCTimeoutDueToLeaderFail)
ErrUnhealthy = Error(ErrGRPCUnhealthy)
ErrNoLeader = Error(ErrGRPCNoLeader)
ErrNotCapable = Error(ErrGRPCNotCapable)
ErrStopped = Error(ErrGRPCStopped)
ErrTimeout = Error(ErrGRPCTimeout)
ErrTimeoutDueToLeaderFail = Error(ErrGRPCTimeoutDueToLeaderFail)
ErrTimeoutDueToConnectionLost = Error(ErrGRPCTimeoutDueToConnectionLost)
ErrUnhealthy = Error(ErrGRPCUnhealthy)
)

// EtcdError defines gRPC server errors.
Expand Down
2 changes: 2 additions & 0 deletions etcdserver/api/v3rpc/util.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -57,6 +57,8 @@ func togRPCError(err error) error {
return rpctypes.ErrGRPCTimeout
case etcdserver.ErrTimeoutDueToLeaderFail:
return rpctypes.ErrGRPCTimeoutDueToLeaderFail
case etcdserver.ErrTimeoutDueToConnectionLost:
return rpctypes.ErrGRPCTimeoutDueToConnectionLost
case etcdserver.ErrUnhealthy:
return rpctypes.ErrGRPCUnhealthy

Expand Down