Skip to content

Add retry logic in pulsar consumer unsubscribe #15284

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

Merged
sre-ci-robot merged 1 commit into from
Jan 19, 2022
Merged

Add retry logic in pulsar consumer unsubscribe #15284

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
6 changes: 0 additions & 6 deletions internal/datanode/data_node_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -143,8 +143,6 @@ func TestDataNode(t *testing.T) {
err = node1.Start()
assert.Nil(t, err)
defer func() {
// TODO: wait for reconnecting to Pulsar, delete sleep after Seek wouldn't lead to disconnect with Pulsar
time.Sleep(200 * time.Millisecond)
err := node1.Stop()
assert.Nil(t, err)
}()
Expand Down Expand Up @@ -338,8 +336,6 @@ func TestDataNode(t *testing.T) {
if i <= 2 {
err = node.flowgraphManager.addAndStart(node, &datapb.VchannelInfo{CollectionID: 1, ChannelName: test.dmChannelName})
assert.Nil(t, err)
// TODO: wait for reconnecting to Pulsar, delete sleep after Seek wouldn't lead to disconnect with Pulsar
time.Sleep(200 * time.Millisecond)
vchanNameCh <- test.dmChannelName
}
}
Expand Down Expand Up @@ -413,8 +409,6 @@ func TestWatchChannel(t *testing.T) {
exist := node.flowgraphManager.exist(ch)
assert.True(t, exist)

// TODO: wait for reconnecting to Pulsar, delete sleep after Seek wouldn't lead to disconnect with Pulsar
time.Sleep(200 * time.Millisecond)
err = kv.RemoveWithPrefix(fmt.Sprintf("%s/%d", Params.DataNodeCfg.ChannelWatchSubPath, node.NodeID))
assert.Nil(t, err)
//TODO there is not way to sync Release done, use sleep for now
Expand Down
7 changes: 0 additions & 7 deletions internal/datanode/flow_graph_manager_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,7 +19,6 @@ package datanode
import (
"context"
"testing"
"time"

"github.com/milvus-io/milvus/internal/proto/datapb"
"github.com/milvus-io/milvus/internal/proto/internalpb"
Expand All @@ -46,8 +45,6 @@ func TestFlowGraphManager(t *testing.T) {

fm := newFlowgraphManager()
defer func() {
// TODO: wait for reconnecting to Pulsar, delete sleep after Seek wouldn't lead to disconnect with Pulsar
time.Sleep(200 * time.Millisecond)
fm.dropAll()
}()
t.Run("Test addAndStart", func(t *testing.T) {
Expand All @@ -62,8 +59,6 @@ func TestFlowGraphManager(t *testing.T) {
assert.NoError(t, err)
assert.True(t, fm.exist(vchanName))

// TODO: wait for reconnecting to Pulsar, delete sleep after Seek wouldn't lead to disconnect with Pulsar
time.Sleep(200 * time.Millisecond)
fm.dropAll()
})

Expand All @@ -79,8 +74,6 @@ func TestFlowGraphManager(t *testing.T) {
assert.NoError(t, err)
assert.True(t, fm.exist(vchanName))

// TODO: wait for reconnecting to Pulsar, delete sleep after Seek wouldn't lead to disconnect with Pulsar
time.Sleep(200 * time.Millisecond)
fm.release(vchanName)

assert.False(t, fm.exist(vchanName))
Expand Down
10 changes: 9 additions & 1 deletion internal/util/mqclient/pulsar_consumer.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,11 +17,15 @@
package mqclient

import (
"context"
"sync"
"time"
"unsafe"

"github.com/apache/pulsar-client-go/pulsar"
"github.com/milvus-io/milvus/internal/log"
"github.com/milvus-io/milvus/internal/util/retry"
"go.uber.org/zap"
)

// PulsarConsumer consumes from pulsar
Expand Down Expand Up @@ -108,8 +112,12 @@ func (pc *PulsarConsumer) Close() {
pc.closeOnce.Do(func() {
defer pc.c.Close()
// Unsubscribe for the consumer
err := pc.c.Unsubscribe()
err := retry.Do(context.Background(), func() error {
//TODO need to check error retryable
return pc.c.Unsubscribe()
}, retry.MaxSleepTime(50*time.Millisecond), retry.Attempts(3))
if err != nil {
log.Error("failed to unsubscribe", zap.String("subscription", pc.Subscription()), zap.Error(err))
panic(err)
}
close(pc.closeCh)
Expand Down