Files
tasks/service/history/replication/dlq_handler.go
Alex ShtinandGitHub 8e8f1b0f7c Remove license header from every file (#7689)
## What changed?
<!-- Describe what has changed in this PR -->
Remove license header from every file. Because it is really hard to
follow in this PR here is the summary:
1. License header is removed from all `.go` and `.proto` files
:fireworks::fireworks:🎆.
2. `LICENSE` file in the root directory has only Temporal and Uber
copyrights.
3. 5 other `LICENSE` files added to the packages which have copyrights
different from Temporal and Uber: Datadog, Xargin, "Mat Ryer, Tyler
Bunnell and contributors".
4. `license_file` flag is removed from all code generation tools.
5. `copyright_file` flag is removed from `go:generate mockgen`
directive.
6. All copyright related targets are removed from `Makefile`.
7. Updated Temporal copyright year to 2025 everywhere.

## Why?
<!-- Tell your future self why have you made these changes -->
I double checked with legal department that it is not needed to have
license header in every file. One file per repo is enough. I put all
copyrights to the root `LICENSE` file and removed header from all other
files. Also updated tools and `Makefile`.
2025-05-01 18:50:21 -07:00

371 lines
10 KiB
Go

//go:generate mockgen -package $GOPACKAGE -source $GOFILE -destination dlq_handler_mock.go
package replication
import (
"context"
"errors"
"fmt"
"sync"
historypb "go.temporal.io/api/history/v1"
"go.temporal.io/server/api/adminservice/v1"
enumsspb "go.temporal.io/server/api/enums/v1"
historyspb "go.temporal.io/server/api/history/v1"
replicationspb "go.temporal.io/server/api/replication/v1"
"go.temporal.io/server/client"
"go.temporal.io/server/common/definition"
"go.temporal.io/server/common/log"
"go.temporal.io/server/common/log/tag"
"go.temporal.io/server/common/namespace"
"go.temporal.io/server/common/persistence"
"go.temporal.io/server/common/xdc"
"go.temporal.io/server/service/history/consts"
"go.temporal.io/server/service/history/deletemanager"
historyi "go.temporal.io/server/service/history/interfaces"
"go.temporal.io/server/service/history/tasks"
wcache "go.temporal.io/server/service/history/workflow/cache"
)
type (
// DLQHandler is the interface handles replication DLQ messages
DLQHandler interface {
GetMessages(
ctx context.Context,
sourceCluster string,
lastMessageID int64,
pageSize int,
pageToken []byte,
) ([]*replicationspb.ReplicationTask, []*replicationspb.ReplicationTaskInfo, []byte, error)
PurgeMessages(
ctx context.Context,
sourceCluster string,
lastMessageID int64,
) error
MergeMessages(
ctx context.Context,
sourceCluster string,
lastMessageID int64,
pageSize int,
pageToken []byte,
) ([]byte, error)
}
dlqHandlerImpl struct {
taskExecutorsLock sync.Mutex
taskExecutors map[string]TaskExecutor
shard historyi.ShardContext
deleteManager deletemanager.DeleteManager
workflowCache wcache.Cache
resender xdc.NDCHistoryResender
taskExecutorProvider TaskExecutorProvider
logger log.Logger
}
)
func NewLazyDLQHandler(
shard historyi.ShardContext,
deleteManager deletemanager.DeleteManager,
workflowCache wcache.Cache,
clientBean client.Bean,
taskExecutorProvider TaskExecutorProvider,
) DLQHandler {
return newDLQHandler(
shard,
deleteManager,
workflowCache,
clientBean,
make(map[string]TaskExecutor),
taskExecutorProvider,
)
}
func newDLQHandler(
shard historyi.ShardContext,
deleteManager deletemanager.DeleteManager,
workflowCache wcache.Cache,
clientBean client.Bean,
taskExecutors map[string]TaskExecutor,
taskExecutorProvider TaskExecutorProvider,
) *dlqHandlerImpl {
if taskExecutors == nil {
panic("Failed to initialize replication DLQ handler due to nil task executors")
}
return &dlqHandlerImpl{
shard: shard,
deleteManager: deleteManager,
workflowCache: workflowCache,
resender: xdc.NewNDCHistoryResender(
shard.GetNamespaceRegistry(),
clientBean,
func(
ctx context.Context,
sourceClusterName string,
namespaceId namespace.ID,
workflowId string,
runId string,
events [][]*historypb.HistoryEvent,
versionHistory []*historyspb.VersionHistoryItem,
) error {
engine, err := shard.GetEngine(ctx)
if err != nil {
return err
}
err = engine.ReplicateHistoryEvents(
ctx,
definition.WorkflowKey{
NamespaceID: namespaceId.String(),
WorkflowID: workflowId,
RunID: runId,
},
nil,
versionHistory,
events,
nil,
"",
)
if errors.Is(err, consts.ErrDuplicate) {
return nil
}
return err
},
shard.GetPayloadSerializer(),
shard.GetConfig().StandbyTaskReReplicationContextTimeout,
shard.GetLogger(),
nil,
),
taskExecutors: taskExecutors,
taskExecutorProvider: taskExecutorProvider,
logger: shard.GetLogger(),
}
}
func (r *dlqHandlerImpl) GetMessages(
ctx context.Context,
sourceCluster string,
lastMessageID int64,
pageSize int,
pageToken []byte,
) ([]*replicationspb.ReplicationTask, []*replicationspb.ReplicationTaskInfo, []byte, error) {
taskList, taskInfoList, _, token, err := r.readMessagesWithAckLevel(
ctx,
sourceCluster,
lastMessageID,
pageSize,
pageToken,
)
return taskList, taskInfoList, token, err
}
func (r *dlqHandlerImpl) PurgeMessages(
ctx context.Context,
sourceCluster string,
lastMessageID int64,
) error {
ackLevel := r.shard.GetReplicatorDLQAckLevel(sourceCluster)
err := r.shard.GetExecutionManager().RangeDeleteReplicationTaskFromDLQ(
ctx,
&persistence.RangeDeleteReplicationTaskFromDLQRequest{
RangeCompleteHistoryTasksRequest: persistence.RangeCompleteHistoryTasksRequest{
ShardID: r.shard.GetShardID(),
TaskCategory: tasks.CategoryReplication,
InclusiveMinTaskKey: tasks.NewImmediateKey(ackLevel + 1),
ExclusiveMaxTaskKey: tasks.NewImmediateKey(lastMessageID + 1),
},
SourceClusterName: sourceCluster,
},
)
if err != nil {
return err
}
if err = r.shard.UpdateReplicatorDLQAckLevel(
sourceCluster,
lastMessageID,
); err != nil {
r.logger.Error("Failed to purge history replication message", tag.Error(err))
// The update ack level should not block the call. Ignore the error.
}
return nil
}
func (r *dlqHandlerImpl) MergeMessages(
ctx context.Context,
sourceCluster string,
lastMessageID int64,
pageSize int,
pageToken []byte,
) ([]byte, error) {
replicationTasks, _, ackLevel, token, err := r.readMessagesWithAckLevel(
ctx,
sourceCluster,
lastMessageID,
pageSize,
pageToken,
)
if err != nil {
return nil, err
}
taskExecutor, err := r.getOrCreateTaskExecutor(sourceCluster)
if err != nil {
return nil, err
}
for _, task := range replicationTasks {
if err := taskExecutor.Execute(
ctx,
task,
true,
); err != nil {
return nil, err
}
}
err = r.shard.GetExecutionManager().RangeDeleteReplicationTaskFromDLQ(
ctx,
&persistence.RangeDeleteReplicationTaskFromDLQRequest{
RangeCompleteHistoryTasksRequest: persistence.RangeCompleteHistoryTasksRequest{
ShardID: r.shard.GetShardID(),
TaskCategory: tasks.CategoryReplication,
InclusiveMinTaskKey: tasks.NewImmediateKey(ackLevel + 1),
ExclusiveMaxTaskKey: tasks.NewImmediateKey(lastMessageID + 1),
},
SourceClusterName: sourceCluster,
},
)
if err != nil {
return nil, err
}
if err = r.shard.UpdateReplicatorDLQAckLevel(
sourceCluster,
lastMessageID,
); err != nil {
r.logger.Error("Failed to purge history replication message", tag.Error(err))
// The update ack level should not block the call. Ignore the error.
}
return token, nil
}
func (r *dlqHandlerImpl) readMessagesWithAckLevel(
ctx context.Context,
sourceCluster string,
lastMessageID int64,
pageSize int,
pageToken []byte,
) ([]*replicationspb.ReplicationTask, []*replicationspb.ReplicationTaskInfo, int64, []byte, error) {
ackLevel := r.shard.GetReplicatorDLQAckLevel(sourceCluster)
resp, err := r.shard.GetExecutionManager().GetReplicationTasksFromDLQ(ctx, &persistence.GetReplicationTasksFromDLQRequest{
GetHistoryTasksRequest: persistence.GetHistoryTasksRequest{
ShardID: r.shard.GetShardID(),
TaskCategory: tasks.CategoryReplication,
InclusiveMinTaskKey: tasks.NewImmediateKey(ackLevel + 1),
ExclusiveMaxTaskKey: tasks.NewImmediateKey(lastMessageID + 1),
BatchSize: pageSize,
NextPageToken: pageToken,
},
SourceClusterName: sourceCluster,
})
if err != nil {
return nil, nil, ackLevel, nil, err
}
pageToken = resp.NextPageToken
remoteAdminClient, err := r.shard.GetRemoteAdminClient(sourceCluster)
if err != nil {
return nil, nil, ackLevel, nil, err
}
taskInfo := make([]*replicationspb.ReplicationTaskInfo, 0, len(resp.Tasks))
for _, task := range resp.Tasks {
switch task := task.(type) {
case *tasks.SyncActivityTask:
taskInfo = append(taskInfo, &replicationspb.ReplicationTaskInfo{
NamespaceId: task.NamespaceID,
WorkflowId: task.WorkflowID,
RunId: task.RunID,
TaskType: enumsspb.TASK_TYPE_REPLICATION_SYNC_ACTIVITY,
TaskId: task.TaskID,
Version: task.GetVersion(),
FirstEventId: 0,
NextEventId: 0,
ScheduledEventId: task.ScheduledEventID,
})
case *tasks.HistoryReplicationTask:
taskInfo = append(taskInfo, &replicationspb.ReplicationTaskInfo{
NamespaceId: task.NamespaceID,
WorkflowId: task.WorkflowID,
RunId: task.RunID,
TaskType: enumsspb.TASK_TYPE_REPLICATION_HISTORY,
TaskId: task.TaskID,
Version: task.Version,
FirstEventId: task.FirstEventID,
NextEventId: task.NextEventID,
ScheduledEventId: 0,
})
case *tasks.SyncWorkflowStateTask:
taskInfo = append(taskInfo, &replicationspb.ReplicationTaskInfo{
NamespaceId: task.NamespaceID,
WorkflowId: task.WorkflowID,
RunId: task.RunID,
TaskType: enumsspb.TASK_TYPE_REPLICATION_SYNC_WORKFLOW_STATE,
TaskId: task.TaskID,
Version: task.Version,
FirstEventId: 0,
NextEventId: 0,
ScheduledEventId: 0,
})
case *tasks.SyncHSMTask:
taskInfo = append(taskInfo, &replicationspb.ReplicationTaskInfo{
NamespaceId: task.NamespaceID,
WorkflowId: task.WorkflowID,
RunId: task.RunID,
TaskType: enumsspb.TASK_TYPE_REPLICATION_SYNC_HSM,
TaskId: task.TaskID,
})
default:
panic(fmt.Sprintf("Unknown repication task type: %v", task))
}
}
if len(taskInfo) == 0 {
return nil, nil, ackLevel, pageToken, nil
}
dlqResponse, err := remoteAdminClient.GetDLQReplicationMessages(
ctx,
&adminservice.GetDLQReplicationMessagesRequest{
TaskInfos: taskInfo,
},
)
if err != nil {
return nil, nil, ackLevel, nil, err
}
return dlqResponse.ReplicationTasks, taskInfo, ackLevel, pageToken, nil
}
func (r *dlqHandlerImpl) getOrCreateTaskExecutor(clusterName string) (TaskExecutor, error) {
r.taskExecutorsLock.Lock()
defer r.taskExecutorsLock.Unlock()
if executor, ok := r.taskExecutors[clusterName]; ok {
return executor, nil
}
taskExecutor := r.taskExecutorProvider(TaskExecutorParams{
RemoteCluster: clusterName,
Shard: r.shard,
HistoryResender: r.resender,
DeleteManager: r.deleteManager,
WorkflowCache: r.workflowCache,
})
r.taskExecutors[clusterName] = taskExecutor
return taskExecutor, nil
}