2
0
mirror of https://github.com/hibiken/asynq.git synced 2024-11-10 11:31:58 +08:00

Add timeout to worker goroutines when TERM signal is received

Wait for a certain amount of time to allow for worker goroutines to
finish. If the goroutines don't finish with the timeout duration,
processor will quit the goroutines and restore any unfinished tasks from
the in_progress queue back to the default queue.
This commit is contained in:
Ken Hibino 2019-12-15 21:00:09 -08:00
parent e5686894d3
commit 989b2b6d55
4 changed files with 75 additions and 122 deletions

View File

@ -4,6 +4,7 @@ import (
"testing" "testing"
"time" "time"
"github.com/go-redis/redis/v7"
"github.com/google/go-cmp/cmp" "github.com/google/go-cmp/cmp"
"github.com/hibiken/asynq/internal/rdb" "github.com/hibiken/asynq/internal/rdb"
) )
@ -74,7 +75,10 @@ func TestPoller(t *testing.T) {
} }
// initialize retry queue // initialize retry queue
for _, st := range tc.initRetry { for _, st := range tc.initRetry {
err := rdbClient.RetryLater(st.msg, st.processAt) err := r.ZAdd(retryQ, &redis.Z{
Member: mustMarshal(t, st.msg),
Score: float64(st.processAt.Unix()),
}).Err()
if err != nil { if err != nil {
t.Fatal(err) t.Fatal(err)
} }

View File

@ -3,6 +3,8 @@ package asynq
import ( import (
"fmt" "fmt"
"log" "log"
"math"
"math/rand"
"time" "time"
"github.com/hibiken/asynq/internal/rdb" "github.com/hibiken/asynq/internal/rdb"
@ -24,6 +26,9 @@ type processor struct {
// channel to communicate back to the long running "processor" goroutine. // channel to communicate back to the long running "processor" goroutine.
done chan struct{} done chan struct{}
// quit channel communicates to the in-flight worker goroutines to stop.
quit chan struct{}
} }
func newProcessor(r *rdb.RDB, numWorkers int, handler Handler) *processor { func newProcessor(r *rdb.RDB, numWorkers int, handler Handler) *processor {
@ -33,6 +38,7 @@ func newProcessor(r *rdb.RDB, numWorkers int, handler Handler) *processor {
dequeueTimeout: 5 * time.Second, dequeueTimeout: 5 * time.Second,
sema: make(chan struct{}, numWorkers), sema: make(chan struct{}, numWorkers),
done: make(chan struct{}), done: make(chan struct{}),
quit: make(chan struct{}),
} }
} }
@ -42,12 +48,16 @@ func (p *processor) terminate() {
// Signal the processor goroutine to stop processing tasks from the queue. // Signal the processor goroutine to stop processing tasks from the queue.
p.done <- struct{}{} p.done <- struct{}{}
// TODO(hibiken): Allow user to customize this timeout value.
const timeout = 8 * time.Second
time.AfterFunc(timeout, func() { close(p.quit) })
log.Println("[INFO] Waiting for all workers to finish...") log.Println("[INFO] Waiting for all workers to finish...")
// block until all workers have released the token // block until all workers have released the token
for i := 0; i < cap(p.sema); i++ { for i := 0; i < cap(p.sema); i++ {
p.sema <- struct{}{} p.sema <- struct{}{}
} }
log.Println("[INFO] All workers have finished.") log.Println("[INFO] All workers have finished.")
p.restore() // move any unfinished tasks back to the queue.
} }
func (p *processor) start() { func (p *processor) start() {
@ -80,22 +90,37 @@ func (p *processor) exec() {
return return
} }
task := &Task{Type: msg.Type, Payload: msg.Payload}
p.sema <- struct{}{} // acquire token p.sema <- struct{}{} // acquire token
go func(task *Task) { go func() {
// NOTE: This deferred anonymous function needs to take taskMessage as a value because defer func() { <-p.sema /* release token */ }()
// the message can be mutated by the time this function is called.
defer func(msg rdb.TaskMessage) { resCh := make(chan error)
if err := p.rdb.Done(&msg); err != nil { task := &Task{Type: msg.Type, Payload: msg.Payload}
log.Printf("[ERROR] could not mark task %+v as done: %v\n", msg, err) go func() {
resCh <- perform(p.handler, task)
}()
select {
case <-p.quit:
// time is up, quit this worker goroutine.
return
case resErr := <-resCh:
// Note: One of three things should happen.
// 1) Done -> Removes the message from InProgress
// 2) Retry -> Removes the message from InProgress & Adds the message to Retry
// 3) Kill -> Removes the message from InProgress & Adds the message to Dead
if resErr != nil {
if msg.Retried >= msg.Retry {
p.kill(msg, resErr.Error())
return
}
p.retry(msg, resErr.Error())
return
} }
<-p.sema // release token p.markAsDone(msg)
}(*msg) return
err := perform(p.handler, task)
if err != nil {
retryTask(p.rdb, msg, err)
} }
}(task) }()
} }
// restore moves all tasks from "in-progress" back to queue // restore moves all tasks from "in-progress" back to queue
@ -103,7 +128,30 @@ func (p *processor) exec() {
func (p *processor) restore() { func (p *processor) restore() {
err := p.rdb.RestoreUnfinished() err := p.rdb.RestoreUnfinished()
if err != nil { if err != nil {
log.Printf("[ERROR] could not restore unfinished tasks: %v\n", err) log.Printf("[ERROR] Could not restore unfinished tasks: %v\n", err)
}
}
func (p *processor) markAsDone(msg *rdb.TaskMessage) {
err := p.rdb.Done(msg)
if err != nil {
log.Printf("[ERROR] Could not remove task from InProgress queue: %v\n", err)
}
}
func (p *processor) retry(msg *rdb.TaskMessage, errMsg string) {
retryAt := time.Now().Add(delaySeconds(msg.Retried))
err := p.rdb.Retry(msg, retryAt, errMsg)
if err != nil {
log.Printf("[ERROR] Could not send task %+v to Retry queue: %v\n", msg, err)
}
}
func (p *processor) kill(msg *rdb.TaskMessage, errMsg string) {
log.Printf("[WARN] Retry exhausted for task(Type: %q, ID: %v)\n", msg.Type, msg.ID)
err := p.rdb.Kill(msg, errMsg)
if err != nil {
log.Printf("[ERROR] Could not send task %+v to Dead queue: %v\n", msg, err)
} }
} }
@ -118,3 +166,11 @@ func perform(h Handler, task *Task) (err error) {
}() }()
return h.ProcessTask(task) return h.ProcessTask(task)
} }
// delaySeconds returns a number seconds to delay before retrying.
// Formula taken from https://github.com/mperham/sidekiq.
func delaySeconds(count int) time.Duration {
r := rand.New(rand.NewSource(time.Now().UnixNano()))
s := int(math.Pow(float64(count), 4)) + 15 + (r.Intn(30) * (count + 1))
return time.Duration(s) * time.Second
}

View File

@ -1,36 +0,0 @@
package asynq
import (
"log"
"math"
"math/rand"
"time"
"github.com/hibiken/asynq/internal/rdb"
)
func retryTask(r *rdb.RDB, msg *rdb.TaskMessage, err error) {
msg.ErrorMsg = err.Error()
if msg.Retried >= msg.Retry {
log.Printf("[WARN] Retry exhausted for task(Type: %q, ID: %v)\n", msg.Type, msg.ID)
if err := r.Kill(msg); err != nil {
log.Printf("[ERROR] Could not add task %+v to 'dead'\n", err)
}
return
}
retryAt := time.Now().Add(delaySeconds((msg.Retried)))
log.Printf("[INFO] Retrying task(Type: %q, ID: %v) in %v\n", msg.Type, msg.ID, retryAt.Sub(time.Now()))
msg.Retried++
if err := r.RetryLater(msg, retryAt); err != nil {
log.Printf("[ERROR] Could not add msg %+v to 'retry': %v\n", msg, err)
return
}
}
// delaySeconds returns a number seconds to delay before retrying.
// Formula taken from https://github.com/mperham/sidekiq.
func delaySeconds(count int) time.Duration {
r := rand.New(rand.NewSource(time.Now().UnixNano()))
s := int(math.Pow(float64(count), 4)) + 15 + (r.Intn(30) * (count + 1))
return time.Duration(s) * time.Second
}

View File

@ -1,71 +0,0 @@
package asynq
import (
"fmt"
"testing"
"github.com/google/go-cmp/cmp"
"github.com/rs/xid"
"github.com/hibiken/asynq/internal/rdb"
)
func TestRetry(t *testing.T) {
r := setup(t)
rdbClient := rdb.NewRDB(r)
errMsg := "email server not responding"
// t1 is a task with max-retry count reached.
t1 := &rdb.TaskMessage{Type: "send_email", Retry: 10, Retried: 10, Queue: "default", ID: xid.New()}
// t2 is t1 with updated error message.
t2 := *t1
t2.ErrorMsg = errMsg
// t3 is a task which hasn't reached max-retry count.
t3 := &rdb.TaskMessage{Type: "send_email", Retry: 10, Retried: 5, Queue: "default", ID: xid.New()}
// t4 is t3 after retry.
t4 := *t3
t4.Retried++
t4.ErrorMsg = errMsg
tests := []struct {
desc string // test case description
msg *rdb.TaskMessage // task to retry
err error // error that caused retry
wantDead []*rdb.TaskMessage // state "dead" queue should be in
wantRetry []*rdb.TaskMessage // state "retry" queue should be in
}{
{
desc: "With retry exhausted task",
msg: t1,
err: fmt.Errorf(errMsg),
wantDead: []*rdb.TaskMessage{&t2},
wantRetry: []*rdb.TaskMessage{},
},
{
desc: "With retry-able task",
msg: t3,
err: fmt.Errorf(errMsg),
wantDead: []*rdb.TaskMessage{},
wantRetry: []*rdb.TaskMessage{&t4},
},
}
for _, tc := range tests {
// clean up db before each test case.
if err := r.FlushDB().Err(); err != nil {
t.Fatal(err)
}
retryTask(rdbClient, tc.msg, tc.err)
deadQueue := r.ZRange(deadQ, 0, -1).Val()
gotDead := mustUnmarshalSlice(t, deadQueue)
if diff := cmp.Diff(tc.wantDead, gotDead, sortMsgOpt); diff != "" {
t.Errorf("%s;\nmismatch found in %q after retryTask(); (-want, +got)\n%s", tc.desc, deadQ, diff)
}
retryQueue := r.ZRange(retryQ, 0, -1).Val()
gotRetry := mustUnmarshalSlice(t, retryQueue)
if diff := cmp.Diff(tc.wantRetry, gotRetry, sortMsgOpt); diff != "" {
t.Errorf("%s;\nmismatch found in %q after retryTask(); (-want, +got)\n%s", tc.desc, deadQ, diff)
}
}
}