Queue: Make WorkerPools and Queues flushable (#10001)

* Make WorkerPools and Queues flushable

Adds Flush methods to Queues and the WorkerPool
Further abstracts the WorkerPool
Adds a final step to Flush the queues in the defer from PrintCurrentTest
Fixes an issue with Settings inheritance in queues

Signed-off-by: Andrew Thornton <art27@cantab.net>

* Change to for loop

* Add IsEmpty and begin just making the queues composed WorkerPools

* subsume workerpool into the queues and create a flushable interface

* Add manager command

* Move flushall to queue.Manager and add to testlogger

* As per @guillep2k

* as per @guillep2k

* Just make queues all implement flushable and clean up the wrapped queue flushes

* cope with no timeout

Co-authored-by: Lauris BH <lauris@nix.lv>
This commit is contained in:
zeripath 2020-01-29 01:01:06 +00:00 committed by GitHub
parent 7c84dbca4f
commit c01221e70f
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
27 changed files with 1030 additions and 327 deletions

92
cmd/manager.go Normal file
View File

@ -0,0 +1,92 @@
// Copyright 2020 The Gitea Authors. All rights reserved.
// Use of this source code is governed by a MIT-style
// license that can be found in the LICENSE file.
package cmd
import (
"fmt"
"net/http"
"os"
"time"
"code.gitea.io/gitea/modules/private"
"github.com/urfave/cli"
)
var (
// CmdManager represents the manager command
CmdManager = cli.Command{
Name: "manager",
Usage: "Manage the running gitea process",
Description: "This is a command for managing the running gitea process",
Subcommands: []cli.Command{
subcmdShutdown,
subcmdRestart,
subcmdFlushQueues,
},
}
subcmdShutdown = cli.Command{
Name: "shutdown",
Usage: "Gracefully shutdown the running process",
Action: runShutdown,
}
subcmdRestart = cli.Command{
Name: "restart",
Usage: "Gracefully restart the running process - (not implemented for windows servers)",
Action: runRestart,
}
subcmdFlushQueues = cli.Command{
Name: "flush-queues",
Usage: "Flush queues in the running process",
Action: runFlushQueues,
Flags: []cli.Flag{
cli.DurationFlag{
Name: "timeout",
Value: 60 * time.Second,
Usage: "Timeout for the flushing process",
},
cli.BoolFlag{
Name: "non-blocking",
Usage: "Set to true to not wait for flush to complete before returning",
},
},
}
)
func runShutdown(c *cli.Context) error {
setup("manager", false)
statusCode, msg := private.Shutdown()
switch statusCode {
case http.StatusInternalServerError:
fail("InternalServerError", msg)
}
fmt.Fprintln(os.Stdout, msg)
return nil
}
func runRestart(c *cli.Context) error {
setup("manager", false)
statusCode, msg := private.Restart()
switch statusCode {
case http.StatusInternalServerError:
fail("InternalServerError", msg)
}
fmt.Fprintln(os.Stdout, msg)
return nil
}
func runFlushQueues(c *cli.Context) error {
setup("manager", false)
statusCode, msg := private.FlushQueues(c.Duration("timeout"), c.Bool("non-blocking"))
switch statusCode {
case http.StatusInternalServerError:
fail("InternalServerError", msg)
}
fmt.Fprintln(os.Stdout, msg)
return nil
}

View File

@ -5,6 +5,7 @@
package integrations
import (
"context"
"encoding/json"
"fmt"
"os"
@ -12,8 +13,10 @@ import (
"strings"
"sync"
"testing"
"time"
"code.gitea.io/gitea/modules/log"
"code.gitea.io/gitea/modules/queue"
)
var prefix string
@ -98,6 +101,9 @@ func PrintCurrentTest(t testing.TB, skip ...int) func() {
}
writerCloser.setT(&t)
return func() {
if err := queue.GetManager().FlushAll(context.Background(), 20*time.Second); err != nil {
t.Errorf("Flushing queues failed with error %v", err)
}
_ = writerCloser.Close()
}
}

View File

@ -69,6 +69,7 @@ arguments - which can alternatively be run by running the subcommand web.`
cmd.CmdKeys,
cmd.CmdConvert,
cmd.CmdDoctor,
cmd.CmdManager,
}
// Now adjust these commands to add our global configuration options

View File

@ -110,28 +110,19 @@ func (g *Manager) handleSignals(ctx context.Context) {
case sig := <-signalChannel:
switch sig {
case syscall.SIGHUP:
if setting.GracefulRestartable {
log.Info("PID: %d. Received SIGHUP. Forking...", pid)
err := g.doFork()
if err != nil && err.Error() != "another process already forked. Ignoring this one" {
log.Error("Error whilst forking from PID: %d : %v", pid, err)
}
} else {
log.Info("PID: %d. Received SIGHUP. Not set restartable. Shutting down...", pid)
g.doShutdown()
}
log.Info("PID: %d. Received SIGHUP. Attempting GracefulShutdown...", pid)
g.DoGracefulShutdown()
case syscall.SIGUSR1:
log.Info("PID %d. Received SIGUSR1.", pid)
case syscall.SIGUSR2:
log.Warn("PID %d. Received SIGUSR2. Hammering...", pid)
g.doHammerTime(0 * time.Second)
g.DoImmediateHammer()
case syscall.SIGINT:
log.Warn("PID %d. Received SIGINT. Shutting down...", pid)
g.doShutdown()
g.DoGracefulShutdown()
case syscall.SIGTERM:
log.Warn("PID %d. Received SIGTERM. Shutting down...", pid)
g.doShutdown()
g.DoGracefulShutdown()
case syscall.SIGTSTP:
log.Info("PID %d. Received SIGTSTP.", pid)
default:
@ -139,7 +130,7 @@ func (g *Manager) handleSignals(ctx context.Context) {
}
case <-ctx.Done():
log.Warn("PID: %d. Background context for manager closed - %v - Shutting down...", pid, ctx.Err())
g.doShutdown()
g.DoGracefulShutdown()
}
}
}
@ -160,6 +151,31 @@ func (g *Manager) doFork() error {
return err
}
// DoGracefulRestart causes a graceful restart
func (g *Manager) DoGracefulRestart() {
if setting.GracefulRestartable {
log.Info("PID: %d. Forking...", os.Getpid())
err := g.doFork()
if err != nil && err.Error() != "another process already forked. Ignoring this one" {
log.Error("Error whilst forking from PID: %d : %v", os.Getpid(), err)
}
} else {
log.Info("PID: %d. Not set restartable. Shutting down...", os.Getpid())
g.doShutdown()
}
}
// DoImmediateHammer causes an immediate hammer
func (g *Manager) DoImmediateHammer() {
g.doHammerTime(0 * time.Second)
}
// DoGracefulShutdown causes a graceful shutdown
func (g *Manager) DoGracefulShutdown() {
g.doShutdown()
}
// RegisterServer registers the running of a listening server, in the case of unix this means that the parent process can now die.
// Any call to RegisterServer must be matched by a call to ServerDone
func (g *Manager) RegisterServer() {

View File

@ -43,6 +43,7 @@ type Manager struct {
runningServerWaitGroup sync.WaitGroup
createServerWaitGroup sync.WaitGroup
terminateWaitGroup sync.WaitGroup
shutdownRequested chan struct{}
}
func newGracefulManager(ctx context.Context) *Manager {
@ -62,6 +63,7 @@ func (g *Manager) start() {
g.shutdown = make(chan struct{})
g.hammer = make(chan struct{})
g.done = make(chan struct{})
g.shutdownRequested = make(chan struct{})
// Set the running state
g.setState(stateRunning)
@ -107,7 +109,10 @@ loop:
for {
select {
case <-g.ctx.Done():
g.doShutdown()
g.DoGracefulShutdown()
waitTime += setting.GracefulHammerTime
break loop
case <-g.shutdownRequested:
waitTime += setting.GracefulHammerTime
break loop
case change := <-changes:
@ -115,12 +120,12 @@ loop:
case svc.Interrogate:
status <- change.CurrentStatus
case svc.Stop, svc.Shutdown:
g.doShutdown()
g.DoGracefulShutdown()
waitTime += setting.GracefulHammerTime
break loop
case hammerCode:
g.doShutdown()
g.doHammerTime(0 * time.Second)
g.DoGracefulShutdown()
g.DoImmediateHammer()
break loop
default:
log.Debug("Unexpected control request: %v", change.Cmd)
@ -140,7 +145,7 @@ hammerLoop:
case svc.Interrogate:
status <- change.CurrentStatus
case svc.Stop, svc.Shutdown, hammerCmd:
g.doHammerTime(0 * time.Second)
g.DoImmediateHammer()
break hammerLoop
default:
log.Debug("Unexpected control request: %v", change.Cmd)
@ -152,6 +157,24 @@ hammerLoop:
return false, 0
}
// DoImmediateHammer causes an immediate hammer
func (g *Manager) DoImmediateHammer() {
g.doHammerTime(0 * time.Second)
}
// DoGracefulShutdown causes a graceful shutdown
func (g *Manager) DoGracefulShutdown() {
g.lock.Lock()
select {
case <-g.shutdownRequested:
g.lock.Unlock()
default:
close(g.shutdownRequested)
g.lock.Unlock()
g.doShutdown()
}
}
// RegisterServer registers the running of a listening server.
// Any call to RegisterServer must be matched by a call to ServerDone
func (g *Manager) RegisterServer() {

View File

@ -0,0 +1,83 @@
// Copyright 2020 The Gitea Authors. All rights reserved.
// Use of this source code is governed by a MIT-style
// license that can be found in the LICENSE file.
package private
import (
"encoding/json"
"fmt"
"net/http"
"time"
"code.gitea.io/gitea/modules/setting"
)
// Shutdown calls the internal shutdown function
func Shutdown() (int, string) {
reqURL := setting.LocalURL + fmt.Sprintf("api/internal/manager/shutdown")
req := newInternalRequest(reqURL, "POST")
resp, err := req.Response()
if err != nil {
return http.StatusInternalServerError, fmt.Sprintf("Unable to contact gitea: %v", err.Error())
}
defer resp.Body.Close()
if resp.StatusCode != http.StatusOK {
return resp.StatusCode, decodeJSONError(resp).Err
}
return http.StatusOK, "Shutting down"
}
// Restart calls the internal restart function
func Restart() (int, string) {
reqURL := setting.LocalURL + fmt.Sprintf("api/internal/manager/restart")
req := newInternalRequest(reqURL, "POST")
resp, err := req.Response()
if err != nil {
return http.StatusInternalServerError, fmt.Sprintf("Unable to contact gitea: %v", err.Error())
}
defer resp.Body.Close()
if resp.StatusCode != http.StatusOK {
return resp.StatusCode, decodeJSONError(resp).Err
}
return http.StatusOK, "Restarting"
}
// FlushOptions represents the options for the flush call
type FlushOptions struct {
Timeout time.Duration
NonBlocking bool
}
// FlushQueues calls the internal flush-queues function
func FlushQueues(timeout time.Duration, nonBlocking bool) (int, string) {
reqURL := setting.LocalURL + fmt.Sprintf("api/internal/manager/flush-queues")
req := newInternalRequest(reqURL, "POST")
if timeout > 0 {
req.SetTimeout(timeout+10*time.Second, timeout+10*time.Second)
}
req = req.Header("Content-Type", "application/json")
jsonBytes, _ := json.Marshal(FlushOptions{
Timeout: timeout,
NonBlocking: nonBlocking,
})
req.Body(jsonBytes)
resp, err := req.Response()
if err != nil {
return http.StatusInternalServerError, fmt.Sprintf("Unable to contact gitea: %v", err.Error())
}
defer resp.Body.Close()
if resp.StatusCode != http.StatusOK {
return resp.StatusCode, decodeJSONError(resp).Err
}
return http.StatusOK, "Flushed"
}

63
modules/queue/helper.go Normal file
View File

@ -0,0 +1,63 @@
// Copyright 2020 The Gitea Authors. All rights reserved.
// Use of this source code is governed by a MIT-style
// license that can be found in the LICENSE file.
package queue
import (
"encoding/json"
"reflect"
)
// toConfig will attempt to convert a given configuration cfg into the provided exemplar type.
//
// It will tolerate the cfg being passed as a []byte or string of a json representation of the
// exemplar or the correct type of the exemplar itself
func toConfig(exemplar, cfg interface{}) (interface{}, error) {
if reflect.TypeOf(cfg).AssignableTo(reflect.TypeOf(exemplar)) {
return cfg, nil
}
configBytes, ok := cfg.([]byte)
if !ok {
configStr, ok := cfg.(string)
if !ok {
return nil, ErrInvalidConfiguration{cfg: cfg}
}
configBytes = []byte(configStr)
}
newVal := reflect.New(reflect.TypeOf(exemplar))
if err := json.Unmarshal(configBytes, newVal.Interface()); err != nil {
return nil, ErrInvalidConfiguration{cfg: cfg, err: err}
}
return newVal.Elem().Interface(), nil
}
// unmarshalAs will attempt to unmarshal provided bytes as the provided exemplar
func unmarshalAs(bs []byte, exemplar interface{}) (data Data, err error) {
if exemplar != nil {
t := reflect.TypeOf(exemplar)
n := reflect.New(t)
ne := n.Elem()
err = json.Unmarshal(bs, ne.Addr().Interface())
data = ne.Interface().(Data)
} else {
err = json.Unmarshal(bs, &data)
}
return
}
// assignableTo will check if provided data is assignable to the same type as the exemplar
// if the provided exemplar is nil then it will always return true
func assignableTo(data Data, exemplar interface{}) bool {
if exemplar == nil {
return true
}
// Assert data is of same type as exemplar
t := reflect.TypeOf(data)
exemplarType := reflect.TypeOf(exemplar)
return t.AssignableTo(exemplarType) && data != nil
}

View File

@ -26,36 +26,57 @@ type Manager struct {
Queues map[int64]*ManagedQueue
}
// ManagedQueue represents a working queue inheriting from Gitea.
// ManagedQueue represents a working queue with a Pool of workers.
//
// Although a ManagedQueue should really represent a Queue this does not
// necessarily have to be the case. This could be used to describe any queue.WorkerPool.
type ManagedQueue struct {
mutex sync.Mutex
QID int64
Queue Queue
Type Type
Name string
Configuration interface{}
ExemplarType string
Pool ManagedPool
Managed interface{}
counter int64
PoolWorkers map[int64]*PoolWorkers
}
// Flushable represents a pool or queue that is flushable
type Flushable interface {
// Flush will add a flush worker to the pool - the worker should be autoregistered with the manager
Flush(time.Duration) error
// FlushWithContext is very similar to Flush
// NB: The worker will not be registered with the manager.
FlushWithContext(ctx context.Context) error
// IsEmpty will return if the managed pool is empty and has no work
IsEmpty() bool
}
// ManagedPool is a simple interface to get certain details from a worker pool
type ManagedPool interface {
// AddWorkers adds a number of worker as group to the pool with the provided timeout. A CancelFunc is provided to cancel the group
AddWorkers(number int, timeout time.Duration) context.CancelFunc
// NumberOfWorkers returns the total number of workers in the pool
NumberOfWorkers() int
// MaxNumberOfWorkers returns the maximum number of workers the pool can dynamically grow to
MaxNumberOfWorkers() int
// SetMaxNumberOfWorkers sets the maximum number of workers the pool can dynamically grow to
SetMaxNumberOfWorkers(int)
// BoostTimeout returns the current timeout for worker groups created during a boost
BoostTimeout() time.Duration
// BlockTimeout returns the timeout the internal channel can block for before a boost would occur
BlockTimeout() time.Duration
// BoostWorkers sets the number of workers to be created during a boost
BoostWorkers() int
SetSettings(maxNumberOfWorkers, boostWorkers int, timeout time.Duration)
// SetPoolSettings sets the user updatable settings for the pool
SetPoolSettings(maxNumberOfWorkers, boostWorkers int, timeout time.Duration)
}
// ManagedQueueList implements the sort.Interface
type ManagedQueueList []*ManagedQueue
// PoolWorkers represents a working queue inheriting from Gitea.
// PoolWorkers represents a group of workers working on a queue
type PoolWorkers struct {
PID int64
Workers int
@ -63,9 +84,10 @@ type PoolWorkers struct {
Timeout time.Time
HasTimeout bool
Cancel context.CancelFunc
IsFlusher bool
}
// PoolWorkersList implements the sort.Interface
// PoolWorkersList implements the sort.Interface for PoolWorkers
type PoolWorkersList []*PoolWorkers
func init() {
@ -83,27 +105,28 @@ func GetManager() *Manager {
}
// Add adds a queue to this manager
func (m *Manager) Add(queue Queue,
func (m *Manager) Add(managed interface{},
t Type,
configuration,
exemplar interface{},
pool ManagedPool) int64 {
exemplar interface{}) int64 {
cfg, _ := json.Marshal(configuration)
mq := &ManagedQueue{
Queue: queue,
Type: t,
Configuration: string(cfg),
ExemplarType: reflect.TypeOf(exemplar).String(),
PoolWorkers: make(map[int64]*PoolWorkers),
Pool: pool,
Managed: managed,
}
m.mutex.Lock()
m.counter++
mq.QID = m.counter
mq.Name = fmt.Sprintf("queue-%d", mq.QID)
if named, ok := queue.(Named); ok {
mq.Name = named.Name()
if named, ok := managed.(Named); ok {
name := named.Name()
if len(name) > 0 {
mq.Name = name
}
}
m.Queues[mq.QID] = mq
m.mutex.Unlock()
@ -127,6 +150,64 @@ func (m *Manager) GetManagedQueue(qid int64) *ManagedQueue {
return m.Queues[qid]
}
// FlushAll flushes all the flushable queues attached to this manager
func (m *Manager) FlushAll(baseCtx context.Context, timeout time.Duration) error {
var ctx context.Context
var cancel context.CancelFunc
start := time.Now()
end := start
hasTimeout := false
if timeout > 0 {
ctx, cancel = context.WithTimeout(baseCtx, timeout)
end = start.Add(timeout)
hasTimeout = true
} else {
ctx, cancel = context.WithCancel(baseCtx)
}
defer cancel()
for {
select {
case <-ctx.Done():
return ctx.Err()
default:
}
mqs := m.ManagedQueues()
wg := sync.WaitGroup{}
wg.Add(len(mqs))
allEmpty := true
for _, mq := range mqs {
if mq.IsEmpty() {
wg.Done()
continue
}
allEmpty = false
if flushable, ok := mq.Managed.(Flushable); ok {
go func() {
localCtx, localCancel := context.WithCancel(ctx)
pid := mq.RegisterWorkers(1, start, hasTimeout, end, localCancel, true)
err := flushable.FlushWithContext(localCtx)
if err != nil && err != ctx.Err() {
cancel()
}
mq.CancelWorkers(pid)
localCancel()
wg.Done()
}()
} else {
wg.Done()
}
}
if allEmpty {
break
}
wg.Wait()
}
return nil
}
// ManagedQueues returns the managed queues
func (m *Manager) ManagedQueues() []*ManagedQueue {
m.mutex.Lock()
@ -152,7 +233,7 @@ func (q *ManagedQueue) Workers() []*PoolWorkers {
}
// RegisterWorkers registers workers to this queue
func (q *ManagedQueue) RegisterWorkers(number int, start time.Time, hasTimeout bool, timeout time.Time, cancel context.CancelFunc) int64 {
func (q *ManagedQueue) RegisterWorkers(number int, start time.Time, hasTimeout bool, timeout time.Time, cancel context.CancelFunc, isFlusher bool) int64 {
q.mutex.Lock()
defer q.mutex.Unlock()
q.counter++
@ -163,6 +244,7 @@ func (q *ManagedQueue) RegisterWorkers(number int, start time.Time, hasTimeout b
Timeout: timeout,
HasTimeout: hasTimeout,
Cancel: cancel,
IsFlusher: isFlusher,
}
return q.counter
}
@ -191,57 +273,74 @@ func (q *ManagedQueue) RemoveWorkers(pid int64) {
// AddWorkers adds workers to the queue if it has registered an add worker function
func (q *ManagedQueue) AddWorkers(number int, timeout time.Duration) context.CancelFunc {
if q.Pool != nil {
if pool, ok := q.Managed.(ManagedPool); ok {
// the cancel will be added to the pool workers description above
return q.Pool.AddWorkers(number, timeout)
return pool.AddWorkers(number, timeout)
}
return nil
}
// Flush flushes the queue with a timeout
func (q *ManagedQueue) Flush(timeout time.Duration) error {
if flushable, ok := q.Managed.(Flushable); ok {
// the cancel will be added to the pool workers description above
return flushable.Flush(timeout)
}
return nil
}
// IsEmpty returns if the queue is empty
func (q *ManagedQueue) IsEmpty() bool {
if flushable, ok := q.Managed.(Flushable); ok {
return flushable.IsEmpty()
}
return true
}
// NumberOfWorkers returns the number of workers in the queue
func (q *ManagedQueue) NumberOfWorkers() int {
if q.Pool != nil {
return q.Pool.NumberOfWorkers()
if pool, ok := q.Managed.(ManagedPool); ok {
return pool.NumberOfWorkers()
}
return -1
}
// MaxNumberOfWorkers returns the maximum number of workers for the pool
func (q *ManagedQueue) MaxNumberOfWorkers() int {
if q.Pool != nil {
return q.Pool.MaxNumberOfWorkers()
if pool, ok := q.Managed.(ManagedPool); ok {
return pool.MaxNumberOfWorkers()
}
return 0
}
// BoostWorkers returns the number of workers for a boost
func (q *ManagedQueue) BoostWorkers() int {
if q.Pool != nil {
return q.Pool.BoostWorkers()
if pool, ok := q.Managed.(ManagedPool); ok {
return pool.BoostWorkers()
}
return -1
}
// BoostTimeout returns the timeout of the next boost
func (q *ManagedQueue) BoostTimeout() time.Duration {
if q.Pool != nil {
return q.Pool.BoostTimeout()
if pool, ok := q.Managed.(ManagedPool); ok {
return pool.BoostTimeout()
}
return 0
}
// BlockTimeout returns the timeout til the next boost
func (q *ManagedQueue) BlockTimeout() time.Duration {
if q.Pool != nil {
return q.Pool.BlockTimeout()
if pool, ok := q.Managed.(ManagedPool); ok {
return pool.BlockTimeout()
}
return 0
}
// SetSettings sets the setable boost values
func (q *ManagedQueue) SetSettings(maxNumberOfWorkers, boostWorkers int, timeout time.Duration) {
if q.Pool != nil {
q.Pool.SetSettings(maxNumberOfWorkers, boostWorkers, timeout)
// SetPoolSettings sets the setable boost values
func (q *ManagedQueue) SetPoolSettings(maxNumberOfWorkers, boostWorkers int, timeout time.Duration) {
if pool, ok := q.Managed.(ManagedPool); ok {
pool.SetPoolSettings(maxNumberOfWorkers, boostWorkers, timeout)
}
}

View File

@ -6,9 +6,8 @@ package queue
import (
"context"
"encoding/json"
"fmt"
"reflect"
"time"
)
// ErrInvalidConfiguration is called when there is invalid configuration for a queue
@ -53,8 +52,11 @@ type Named interface {
Name() string
}
// Queue defines an interface to save an issue indexer queue
// Queue defines an interface of a queue-like item
//
// Queues will handle their own contents in the Run method
type Queue interface {
Flushable
Run(atShutdown, atTerminate func(context.Context, func()))
Push(Data) error
}
@ -71,32 +73,27 @@ func NewDummyQueue(handler HandlerFunc, opts, exemplar interface{}) (Queue, erro
type DummyQueue struct {
}
// Run starts to run the queue
// Run does nothing
func (b *DummyQueue) Run(_, _ func(context.Context, func())) {}
// Push pushes data to the queue
// Push fakes a push of data to the queue
func (b *DummyQueue) Push(Data) error {
return nil
}
func toConfig(exemplar, cfg interface{}) (interface{}, error) {
if reflect.TypeOf(cfg).AssignableTo(reflect.TypeOf(exemplar)) {
return cfg, nil
}
// Flush always returns nil
func (b *DummyQueue) Flush(time.Duration) error {
return nil
}
configBytes, ok := cfg.([]byte)
if !ok {
configStr, ok := cfg.(string)
if !ok {
return nil, ErrInvalidConfiguration{cfg: cfg}
}
configBytes = []byte(configStr)
}
newVal := reflect.New(reflect.TypeOf(exemplar))
if err := json.Unmarshal(configBytes, newVal.Interface()); err != nil {
return nil, ErrInvalidConfiguration{cfg: cfg, err: err}
}
return newVal.Elem().Interface(), nil
// FlushWithContext always returns nil
func (b *DummyQueue) FlushWithContext(context.Context) error {
return nil
}
// IsEmpty asserts that the queue is empty
func (b *DummyQueue) IsEmpty() bool {
return true
}
var queuesMap = map[Type]NewQueueFunc{DummyQueueType: NewDummyQueue}
@ -123,7 +120,7 @@ func RegisteredTypesAsString() []string {
return types
}
// NewQueue takes a queue Type and HandlerFunc some options and possibly an exemplar and returns a Queue or an error
// NewQueue takes a queue Type, HandlerFunc, some options and possibly an exemplar and returns a Queue or an error
func NewQueue(queueType Type, handlerFunc HandlerFunc, opts, exemplar interface{}) (Queue, error) {
newFn, ok := queuesMap[queueType]
if !ok {

View File

@ -7,8 +7,6 @@ package queue
import (
"context"
"fmt"
"reflect"
"time"
"code.gitea.io/gitea/modules/log"
)
@ -18,25 +16,23 @@ const ChannelQueueType Type = "channel"
// ChannelQueueConfiguration is the configuration for a ChannelQueue
type ChannelQueueConfiguration struct {
QueueLength int
BatchLength int
WorkerPoolConfiguration
Workers int
MaxWorkers int
BlockTimeout time.Duration
BoostTimeout time.Duration
BoostWorkers int
Name string
}
// ChannelQueue implements
// ChannelQueue implements Queue
//
// A channel queue is not persistable and does not shutdown or terminate cleanly
// It is basically a very thin wrapper around a WorkerPool
type ChannelQueue struct {
pool *WorkerPool
*WorkerPool
exemplar interface{}
workers int
name string
}
// NewChannelQueue create a memory channel queue
// NewChannelQueue creates a memory channel queue
func NewChannelQueue(handle HandlerFunc, cfg, exemplar interface{}) (Queue, error) {
configInterface, err := toConfig(ChannelQueueConfiguration{}, cfg)
if err != nil {
@ -46,26 +42,13 @@ func NewChannelQueue(handle HandlerFunc, cfg, exemplar interface{}) (Queue, erro
if config.BatchLength == 0 {
config.BatchLength = 1
}
dataChan := make(chan Data, config.QueueLength)
ctx, cancel := context.WithCancel(context.Background())
queue := &ChannelQueue{
pool: &WorkerPool{
baseCtx: ctx,
cancel: cancel,
batchLength: config.BatchLength,
handle: handle,
dataChan: dataChan,
blockTimeout: config.BlockTimeout,
boostTimeout: config.BoostTimeout,
boostWorkers: config.BoostWorkers,
maxNumberOfWorkers: config.MaxWorkers,
},
WorkerPool: NewWorkerPool(handle, config.WorkerPoolConfiguration),
exemplar: exemplar,
workers: config.Workers,
name: config.Name,
}
queue.pool.qid = GetManager().Add(queue, ChannelQueueType, config, exemplar, queue.pool)
queue.qid = GetManager().Add(queue, ChannelQueueType, config, exemplar)
return queue, nil
}
@ -77,22 +60,18 @@ func (c *ChannelQueue) Run(atShutdown, atTerminate func(context.Context, func())
atTerminate(context.Background(), func() {
log.Warn("ChannelQueue: %s is not terminatable!", c.name)
})
log.Debug("ChannelQueue: %s Starting", c.name)
go func() {
_ = c.pool.AddWorkers(c.workers, 0)
_ = c.AddWorkers(c.workers, 0)
}()
}
// Push will push data into the queue
func (c *ChannelQueue) Push(data Data) error {
if c.exemplar != nil {
// Assert data is of same type as r.exemplar
t := reflect.TypeOf(data)
exemplarType := reflect.TypeOf(c.exemplar)
if !t.AssignableTo(exemplarType) || data == nil {
if !assignableTo(data, c.exemplar) {
return fmt.Errorf("Unable to assign data: %v to same type as exemplar: %v in queue: %s", data, c.exemplar, c.name)
}
}
c.pool.Push(data)
c.WorkerPool.Push(data)
return nil
}

View File

@ -25,12 +25,14 @@ func TestChannelQueue(t *testing.T) {
queue, err := NewChannelQueue(handle,
ChannelQueueConfiguration{
WorkerPoolConfiguration: WorkerPoolConfiguration{
QueueLength: 20,
Workers: 1,
MaxWorkers: 10,
BlockTimeout: 1 * time.Second,
BoostTimeout: 5 * time.Minute,
BoostWorkers: 5,
},
Workers: 1,
}, &testData{})
assert.NoError(t, err)
@ -60,13 +62,15 @@ func TestChannelQueue_Batch(t *testing.T) {
queue, err := NewChannelQueue(handle,
ChannelQueueConfiguration{
WorkerPoolConfiguration: WorkerPoolConfiguration{
QueueLength: 20,
BatchLength: 2,
Workers: 1,
MaxWorkers: 10,
BlockTimeout: 1 * time.Second,
BoostTimeout: 5 * time.Minute,
BoostWorkers: 5,
MaxWorkers: 10,
},
Workers: 1,
}, &testData{})
assert.NoError(t, err)

View File

@ -8,8 +8,8 @@ import (
"context"
"encoding/json"
"fmt"
"reflect"
"sync"
"sync/atomic"
"time"
"code.gitea.io/gitea/modules/log"
@ -22,20 +22,15 @@ const LevelQueueType Type = "level"
// LevelQueueConfiguration is the configuration for a LevelQueue
type LevelQueueConfiguration struct {
WorkerPoolConfiguration
DataDir string
QueueLength int
BatchLength int
Workers int
MaxWorkers int
BlockTimeout time.Duration
BoostTimeout time.Duration
BoostWorkers int
Name string
}
// LevelQueue implements a disk library queue
type LevelQueue struct {
pool *WorkerPool
*WorkerPool
queue *levelqueue.Queue
closed chan struct{}
terminated chan struct{}
@ -58,21 +53,8 @@ func NewLevelQueue(handle HandlerFunc, cfg, exemplar interface{}) (Queue, error)
return nil, err
}
dataChan := make(chan Data, config.QueueLength)
ctx, cancel := context.WithCancel(context.Background())
queue := &LevelQueue{
pool: &WorkerPool{
baseCtx: ctx,
cancel: cancel,
batchLength: config.BatchLength,
handle: handle,
dataChan: dataChan,
blockTimeout: config.BlockTimeout,
boostTimeout: config.BoostTimeout,
boostWorkers: config.BoostWorkers,
maxNumberOfWorkers: config.MaxWorkers,
},
WorkerPool: NewWorkerPool(handle, config.WorkerPoolConfiguration),
queue: internal,
exemplar: exemplar,
closed: make(chan struct{}),
@ -80,7 +62,7 @@ func NewLevelQueue(handle HandlerFunc, cfg, exemplar interface{}) (Queue, error)
workers: config.Workers,
name: config.Name,
}
queue.pool.qid = GetManager().Add(queue, LevelQueueType, config, exemplar, queue.pool)
queue.qid = GetManager().Add(queue, LevelQueueType, config, exemplar)
return queue, nil
}
@ -88,9 +70,10 @@ func NewLevelQueue(handle HandlerFunc, cfg, exemplar interface{}) (Queue, error)
func (l *LevelQueue) Run(atShutdown, atTerminate func(context.Context, func())) {
atShutdown(context.Background(), l.Shutdown)
atTerminate(context.Background(), l.Terminate)
log.Debug("LevelQueue: %s Starting", l.name)
go func() {
_ = l.pool.AddWorkers(l.workers, 0)
_ = l.AddWorkers(l.workers, 0)
}()
go l.readToChan()
@ -99,12 +82,12 @@ func (l *LevelQueue) Run(atShutdown, atTerminate func(context.Context, func()))
<-l.closed
log.Trace("LevelQueue: %s Waiting til done", l.name)
l.pool.Wait()
l.Wait()
log.Trace("LevelQueue: %s Waiting til cleaned", l.name)
ctx, cancel := context.WithCancel(context.Background())
atTerminate(ctx, cancel)
l.pool.CleanUp(ctx)
l.CleanUp(ctx)
cancel()
log.Trace("LevelQueue: %s Cleaned", l.name)
@ -115,57 +98,46 @@ func (l *LevelQueue) readToChan() {
select {
case <-l.closed:
// tell the pool to shutdown.
l.pool.cancel()
l.cancel()
return
default:
atomic.AddInt64(&l.numInQueue, 1)
bs, err := l.queue.RPop()
if err != nil {
if err != levelqueue.ErrNotFound {
log.Error("LevelQueue: %s Error on RPop: %v", l.name, err)
}
atomic.AddInt64(&l.numInQueue, -1)
time.Sleep(time.Millisecond * 100)
continue
}
if len(bs) == 0 {
atomic.AddInt64(&l.numInQueue, -1)
time.Sleep(time.Millisecond * 100)
continue
}
var data Data
if l.exemplar != nil {
t := reflect.TypeOf(l.exemplar)
n := reflect.New(t)
ne := n.Elem()
err = json.Unmarshal(bs, ne.Addr().Interface())
data = ne.Interface().(Data)
} else {
err = json.Unmarshal(bs, &data)
}
data, err := unmarshalAs(bs, l.exemplar)
if err != nil {
log.Error("LevelQueue: %s Failed to unmarshal with error: %v", l.name, err)
atomic.AddInt64(&l.numInQueue, -1)
time.Sleep(time.Millisecond * 100)
continue
}
log.Trace("LevelQueue %s: Task found: %#v", l.name, data)
l.pool.Push(data)
l.WorkerPool.Push(data)
atomic.AddInt64(&l.numInQueue, -1)
}
}
}
// Push will push the indexer data to queue
func (l *LevelQueue) Push(data Data) error {
if l.exemplar != nil {
// Assert data is of same type as r.exemplar
value := reflect.ValueOf(data)
t := value.Type()
exemplarType := reflect.ValueOf(l.exemplar).Type()
if !t.AssignableTo(exemplarType) || data == nil {
if !assignableTo(data, l.exemplar) {
return fmt.Errorf("Unable to assign data: %v to same type as exemplar: %v in %s", data, l.exemplar, l.name)
}
}
bs, err := json.Marshal(data)
if err != nil {
return err
@ -173,16 +145,25 @@ func (l *LevelQueue) Push(data Data) error {
return l.queue.LPush(bs)
}
// IsEmpty checks whether the queue is empty
func (l *LevelQueue) IsEmpty() bool {
if !l.WorkerPool.IsEmpty() {
return false
}
return l.queue.Len() == 0
}
// Shutdown this queue and stop processing
func (l *LevelQueue) Shutdown() {
l.lock.Lock()
defer l.lock.Unlock()
log.Trace("LevelQueue: %s Shutdown", l.name)
log.Trace("LevelQueue: %s Shutting down", l.name)
select {
case <-l.closed:
default:
close(l.closed)
}
log.Debug("LevelQueue: %s Shutdown", l.name)
}
// Terminate this queue and close the queue
@ -196,11 +177,15 @@ func (l *LevelQueue) Terminate() {
default:
close(l.terminated)
l.lock.Unlock()
if log.IsDebug() {
log.Debug("LevelQueue: %s Closing with %d tasks left in queue", l.name, l.queue.Len())
}
if err := l.queue.Close(); err != nil && err.Error() != "leveldb: closed" {
log.Error("Error whilst closing internal queue in %s: %v", l.name, err)
}
}
log.Debug("LevelQueue: %s Terminated", l.name)
}
// Name returns the name of this queue

View File

@ -6,7 +6,9 @@ package queue
import (
"context"
"fmt"
"sync"
"sync/atomic"
"time"
"code.gitea.io/gitea/modules/log"
@ -31,8 +33,10 @@ type PersistableChannelQueueConfiguration struct {
}
// PersistableChannelQueue wraps a channel queue and level queue together
// The disk level queue will be used to store data at shutdown and terminate - and will be restored
// on start up.
type PersistableChannelQueue struct {
*ChannelQueue
channelQueue *ChannelQueue
delayedStarter
lock sync.Mutex
closed chan struct{}
@ -48,13 +52,15 @@ func NewPersistableChannelQueue(handle HandlerFunc, cfg, exemplar interface{}) (
config := configInterface.(PersistableChannelQueueConfiguration)
channelQueue, err := NewChannelQueue(handle, ChannelQueueConfiguration{
WorkerPoolConfiguration: WorkerPoolConfiguration{
QueueLength: config.QueueLength,
BatchLength: config.BatchLength,
Workers: config.Workers,
MaxWorkers: config.MaxWorkers,
BlockTimeout: config.BlockTimeout,
BoostTimeout: config.BoostTimeout,
BoostWorkers: config.BoostWorkers,
MaxWorkers: config.MaxWorkers,
},
Workers: config.Workers,
Name: config.Name + "-channel",
}, exemplar)
if err != nil {
@ -63,28 +69,30 @@ func NewPersistableChannelQueue(handle HandlerFunc, cfg, exemplar interface{}) (
// the level backend only needs temporary workers to catch up with the previously dropped work
levelCfg := LevelQueueConfiguration{
DataDir: config.DataDir,
WorkerPoolConfiguration: WorkerPoolConfiguration{
QueueLength: config.QueueLength,
BatchLength: config.BatchLength,
Workers: 1,
MaxWorkers: 6,
BlockTimeout: 1 * time.Second,
BoostTimeout: 5 * time.Minute,
BoostWorkers: 5,
MaxWorkers: 6,
},
DataDir: config.DataDir,
Workers: 1,
Name: config.Name + "-level",
}
levelQueue, err := NewLevelQueue(handle, levelCfg, exemplar)
if err == nil {
queue := &PersistableChannelQueue{
ChannelQueue: channelQueue.(*ChannelQueue),
channelQueue: channelQueue.(*ChannelQueue),
delayedStarter: delayedStarter{
internal: levelQueue.(*LevelQueue),
name: config.Name,
},
closed: make(chan struct{}),
}
_ = GetManager().Add(queue, PersistableChannelQueueType, config, exemplar, nil)
_ = GetManager().Add(queue, PersistableChannelQueueType, config, exemplar)
return queue, nil
}
if IsErrInvalidConfiguration(err) {
@ -93,7 +101,7 @@ func NewPersistableChannelQueue(handle HandlerFunc, cfg, exemplar interface{}) (
}
queue := &PersistableChannelQueue{
ChannelQueue: channelQueue.(*ChannelQueue),
channelQueue: channelQueue.(*ChannelQueue),
delayedStarter: delayedStarter{
cfg: levelCfg,
underlying: LevelQueueType,
@ -103,7 +111,7 @@ func NewPersistableChannelQueue(handle HandlerFunc, cfg, exemplar interface{}) (
},
closed: make(chan struct{}),
}
_ = GetManager().Add(queue, PersistableChannelQueueType, config, exemplar, nil)
_ = GetManager().Add(queue, PersistableChannelQueueType, config, exemplar)
return queue, nil
}
@ -118,15 +126,17 @@ func (p *PersistableChannelQueue) Push(data Data) error {
case <-p.closed:
return p.internal.Push(data)
default:
return p.ChannelQueue.Push(data)
return p.channelQueue.Push(data)
}
}
// Run starts to run the queue
func (p *PersistableChannelQueue) Run(atShutdown, atTerminate func(context.Context, func())) {
log.Debug("PersistableChannelQueue: %s Starting", p.delayedStarter.name)
p.lock.Lock()
if p.internal == nil {
err := p.setInternal(atShutdown, p.ChannelQueue.pool.handle, p.exemplar)
err := p.setInternal(atShutdown, p.channelQueue.handle, p.channelQueue.exemplar)
p.lock.Unlock()
if err != nil {
log.Fatal("Unable to create internal queue for %s Error: %v", p.Name(), err)
@ -142,31 +152,83 @@ func (p *PersistableChannelQueue) Run(atShutdown, atTerminate func(context.Conte
go p.internal.Run(func(_ context.Context, _ func()) {}, func(_ context.Context, _ func()) {})
go func() {
_ = p.ChannelQueue.pool.AddWorkers(p.workers, 0)
_ = p.channelQueue.AddWorkers(p.channelQueue.workers, 0)
}()
log.Trace("PersistableChannelQueue: %s Waiting til closed", p.delayedStarter.name)
<-p.closed
log.Trace("PersistableChannelQueue: %s Cancelling pools", p.delayedStarter.name)
p.ChannelQueue.pool.cancel()
p.internal.(*LevelQueue).pool.cancel()
p.channelQueue.cancel()
p.internal.(*LevelQueue).cancel()
log.Trace("PersistableChannelQueue: %s Waiting til done", p.delayedStarter.name)
p.ChannelQueue.pool.Wait()
p.internal.(*LevelQueue).pool.Wait()
p.channelQueue.Wait()
p.internal.(*LevelQueue).Wait()
// Redirect all remaining data in the chan to the internal channel
go func() {
log.Trace("PersistableChannelQueue: %s Redirecting remaining data", p.delayedStarter.name)
for data := range p.ChannelQueue.pool.dataChan {
for data := range p.channelQueue.dataChan {
_ = p.internal.Push(data)
atomic.AddInt64(&p.channelQueue.numInQueue, -1)
}
log.Trace("PersistableChannelQueue: %s Done Redirecting remaining data", p.delayedStarter.name)
}()
log.Trace("PersistableChannelQueue: %s Done main loop", p.delayedStarter.name)
}
// Flush flushes the queue and blocks till the queue is empty
func (p *PersistableChannelQueue) Flush(timeout time.Duration) error {
var ctx context.Context
var cancel context.CancelFunc
if timeout > 0 {
ctx, cancel = context.WithTimeout(context.Background(), timeout)
} else {
ctx, cancel = context.WithCancel(context.Background())
}
defer cancel()
return p.FlushWithContext(ctx)
}
// FlushWithContext flushes the queue and blocks till the queue is empty
func (p *PersistableChannelQueue) FlushWithContext(ctx context.Context) error {
errChan := make(chan error, 1)
go func() {
errChan <- p.channelQueue.FlushWithContext(ctx)
}()
go func() {
p.lock.Lock()
if p.internal == nil {
p.lock.Unlock()
errChan <- fmt.Errorf("not ready to flush internal queue %s yet", p.Name())
return
}
p.lock.Unlock()
errChan <- p.internal.FlushWithContext(ctx)
}()
err1 := <-errChan
err2 := <-errChan
if err1 != nil {
return err1
}
return err2
}
// IsEmpty checks if a queue is empty
func (p *PersistableChannelQueue) IsEmpty() bool {
if !p.channelQueue.IsEmpty() {
return false
}
p.lock.Lock()
defer p.lock.Unlock()
if p.internal == nil {
return false
}
return p.internal.IsEmpty()
}
// Shutdown processing this queue
func (p *PersistableChannelQueue) Shutdown() {
log.Trace("PersistableChannelQueue: %s Shutdown", p.delayedStarter.name)
log.Trace("PersistableChannelQueue: %s Shutting down", p.delayedStarter.name)
select {
case <-p.closed:
default:
@ -177,6 +239,7 @@ func (p *PersistableChannelQueue) Shutdown() {
}
close(p.closed)
}
log.Debug("PersistableChannelQueue: %s Shutdown", p.delayedStarter.name)
}
// Terminate this queue and close the queue
@ -188,6 +251,7 @@ func (p *PersistableChannelQueue) Terminate() {
if p.internal != nil {
p.internal.(*LevelQueue).Terminate()
}
log.Debug("PersistableChannelQueue: %s Terminated", p.delayedStarter.name)
}
func init() {

View File

@ -32,14 +32,16 @@ func TestLevelQueue(t *testing.T) {
defer os.RemoveAll(tmpDir)
queue, err := NewLevelQueue(handle, LevelQueueConfiguration{
DataDir: tmpDir,
BatchLength: 2,
Workers: 1,
MaxWorkers: 10,
WorkerPoolConfiguration: WorkerPoolConfiguration{
QueueLength: 20,
BatchLength: 2,
BlockTimeout: 1 * time.Second,
BoostTimeout: 5 * time.Minute,
BoostWorkers: 5,
MaxWorkers: 10,
},
DataDir: tmpDir,
Workers: 1,
}, &testData{})
assert.NoError(t, err)
@ -92,14 +94,16 @@ func TestLevelQueue(t *testing.T) {
WrappedQueueConfiguration{
Underlying: LevelQueueType,
Config: LevelQueueConfiguration{
DataDir: tmpDir,
BatchLength: 2,
Workers: 1,
MaxWorkers: 10,
WorkerPoolConfiguration: WorkerPoolConfiguration{
QueueLength: 20,
BatchLength: 2,
BlockTimeout: 1 * time.Second,
BoostTimeout: 5 * time.Minute,
BoostWorkers: 5,
MaxWorkers: 10,
},
DataDir: tmpDir,
Workers: 1,
},
}, &testData{})
assert.NoError(t, err)

View File

@ -9,9 +9,9 @@ import (
"encoding/json"
"errors"
"fmt"
"reflect"
"strings"
"sync"
"sync/atomic"
"time"
"code.gitea.io/gitea/modules/log"
@ -25,13 +25,14 @@ const RedisQueueType Type = "redis"
type redisClient interface {
RPush(key string, args ...interface{}) *redis.IntCmd
LPop(key string) *redis.StringCmd
LLen(key string) *redis.IntCmd
Ping() *redis.StatusCmd
Close() error
}
// RedisQueue redis queue
type RedisQueue struct {
pool *WorkerPool
*WorkerPool
client redisClient
queueName string
closed chan struct{}
@ -44,18 +45,13 @@ type RedisQueue struct {
// RedisQueueConfiguration is the configuration for the redis queue
type RedisQueueConfiguration struct {
WorkerPoolConfiguration
Network string
Addresses string
Password string
DBIndex int
BatchLength int
QueueLength int
QueueName string
Workers int
MaxWorkers int
BlockTimeout time.Duration
BoostTimeout time.Duration
BoostWorkers int
Name string
}
@ -69,21 +65,8 @@ func NewRedisQueue(handle HandlerFunc, cfg, exemplar interface{}) (Queue, error)
dbs := strings.Split(config.Addresses, ",")
dataChan := make(chan Data, config.QueueLength)
ctx, cancel := context.WithCancel(context.Background())
var queue = &RedisQueue{
pool: &WorkerPool{
baseCtx: ctx,
cancel: cancel,
batchLength: config.BatchLength,
handle: handle,
dataChan: dataChan,
blockTimeout: config.BlockTimeout,
boostTimeout: config.BoostTimeout,
boostWorkers: config.BoostWorkers,
maxNumberOfWorkers: config.MaxWorkers,
},
WorkerPool: NewWorkerPool(handle, config.WorkerPoolConfiguration),
queueName: config.QueueName,
exemplar: exemplar,
closed: make(chan struct{}),
@ -108,7 +91,7 @@ func NewRedisQueue(handle HandlerFunc, cfg, exemplar interface{}) (Queue, error)
if err := queue.client.Ping().Err(); err != nil {
return nil, err
}
queue.pool.qid = GetManager().Add(queue, RedisQueueType, config, exemplar, queue.pool)
queue.qid = GetManager().Add(queue, RedisQueueType, config, exemplar)
return queue, nil
}
@ -117,9 +100,10 @@ func NewRedisQueue(handle HandlerFunc, cfg, exemplar interface{}) (Queue, error)
func (r *RedisQueue) Run(atShutdown, atTerminate func(context.Context, func())) {
atShutdown(context.Background(), r.Shutdown)
atTerminate(context.Background(), r.Terminate)
log.Debug("RedisQueue: %s Starting", r.name)
go func() {
_ = r.pool.AddWorkers(r.workers, 0)
_ = r.AddWorkers(r.workers, 0)
}()
go r.readToChan()
@ -127,12 +111,12 @@ func (r *RedisQueue) Run(atShutdown, atTerminate func(context.Context, func()))
log.Trace("RedisQueue: %s Waiting til closed", r.name)
<-r.closed
log.Trace("RedisQueue: %s Waiting til done", r.name)
r.pool.Wait()
r.Wait()
log.Trace("RedisQueue: %s Waiting til cleaned", r.name)
ctx, cancel := context.WithCancel(context.Background())
atTerminate(ctx, cancel)
r.pool.CleanUp(ctx)
r.CleanUp(ctx)
cancel()
}
@ -141,54 +125,44 @@ func (r *RedisQueue) readToChan() {
select {
case <-r.closed:
// tell the pool to shutdown
r.pool.cancel()
r.cancel()
return
default:
atomic.AddInt64(&r.numInQueue, 1)
bs, err := r.client.LPop(r.queueName).Bytes()
if err != nil && err != redis.Nil {
log.Error("RedisQueue: %s Error on LPop: %v", r.name, err)
atomic.AddInt64(&r.numInQueue, -1)
time.Sleep(time.Millisecond * 100)
continue
}
if len(bs) == 0 {
atomic.AddInt64(&r.numInQueue, -1)
time.Sleep(time.Millisecond * 100)
continue
}
var data Data
if r.exemplar != nil {
t := reflect.TypeOf(r.exemplar)
n := reflect.New(t)
ne := n.Elem()
err = json.Unmarshal(bs, ne.Addr().Interface())
data = ne.Interface().(Data)
} else {
err = json.Unmarshal(bs, &data)
}
data, err := unmarshalAs(bs, r.exemplar)
if err != nil {
log.Error("RedisQueue: %s Error on Unmarshal: %v", r.name, err)
atomic.AddInt64(&r.numInQueue, -1)
time.Sleep(time.Millisecond * 100)
continue
}
log.Trace("RedisQueue: %s Task found: %#v", r.name, data)
r.pool.Push(data)
r.WorkerPool.Push(data)
atomic.AddInt64(&r.numInQueue, -1)
}
}
}
// Push implements Queue
func (r *RedisQueue) Push(data Data) error {
if r.exemplar != nil {
// Assert data is of same type as r.exemplar
value := reflect.ValueOf(data)
t := value.Type()
exemplarType := reflect.ValueOf(r.exemplar).Type()
if !t.AssignableTo(exemplarType) || data == nil {
if !assignableTo(data, r.exemplar) {
return fmt.Errorf("Unable to assign data: %v to same type as exemplar: %v in %s", data, r.exemplar, r.name)
}
}
bs, err := json.Marshal(data)
if err != nil {
return err
@ -196,9 +170,22 @@ func (r *RedisQueue) Push(data Data) error {
return r.client.RPush(r.queueName, bs).Err()
}
// IsEmpty checks if the queue is empty
func (r *RedisQueue) IsEmpty() bool {
if !r.WorkerPool.IsEmpty() {
return false
}
length, err := r.client.LLen(r.queueName).Result()
if err != nil {
log.Error("Error whilst getting queue length for %s: Error: %v", r.name, err)
return false
}
return length == 0
}
// Shutdown processing from this queue
func (r *RedisQueue) Shutdown() {
log.Trace("Shutdown: %s", r.name)
log.Trace("RedisQueue: %s Shutting down", r.name)
r.lock.Lock()
select {
case <-r.closed:
@ -206,11 +193,12 @@ func (r *RedisQueue) Shutdown() {
close(r.closed)
}
r.lock.Unlock()
log.Debug("RedisQueue: %s Shutdown", r.name)
}
// Terminate this queue and close the queue
func (r *RedisQueue) Terminate() {
log.Trace("Terminating: %s", r.name)
log.Trace("RedisQueue: %s Terminating", r.name)
r.Shutdown()
r.lock.Lock()
select {
@ -219,10 +207,14 @@ func (r *RedisQueue) Terminate() {
default:
close(r.terminated)
r.lock.Unlock()
if log.IsDebug() {
log.Debug("RedisQueue: %s Closing with %d tasks left in queue", r.name, r.client.LLen(r.queueName))
}
if err := r.client.Close(); err != nil {
log.Error("Error whilst closing internal redis client in %s: %v", r.name, err)
}
}
log.Debug("RedisQueue: %s Terminated", r.name)
}
// Name returns the name of this queue

View File

@ -7,8 +7,8 @@ package queue
import (
"context"
"fmt"
"reflect"
"sync"
"sync/atomic"
"time"
"code.gitea.io/gitea/modules/log"
@ -56,7 +56,7 @@ func (q *delayedStarter) setInternal(atShutdown func(context.Context, func()), h
for q.internal == nil {
select {
case <-ctx.Done():
return fmt.Errorf("Timedout creating queue %v with cfg %v in %s", q.underlying, q.cfg, q.name)
return fmt.Errorf("Timedout creating queue %v with cfg %s in %s", q.underlying, q.cfg, q.name)
default:
queue, err := NewQueue(q.underlying, handle, q.cfg, exemplar)
if err == nil {
@ -64,11 +64,11 @@ func (q *delayedStarter) setInternal(atShutdown func(context.Context, func()), h
break
}
if err.Error() != "resource temporarily unavailable" {
log.Warn("[Attempt: %d] Failed to create queue: %v for %s cfg: %v error: %v", i, q.underlying, q.name, q.cfg, err)
log.Warn("[Attempt: %d] Failed to create queue: %v for %s cfg: %s error: %v", i, q.underlying, q.name, q.cfg, err)
}
i++
if q.maxAttempts > 0 && i > q.maxAttempts {
return fmt.Errorf("Unable to create queue %v for %s with cfg %v by max attempts: error: %v", q.underlying, q.name, q.cfg, err)
return fmt.Errorf("Unable to create queue %v for %s with cfg %s by max attempts: error: %v", q.underlying, q.name, q.cfg, err)
}
sleepTime := 100 * time.Millisecond
if q.timeout > 0 && q.maxAttempts > 0 {
@ -92,6 +92,7 @@ type WrappedQueue struct {
handle HandlerFunc
exemplar interface{}
channel chan Data
numInQueue int64
}
// NewWrappedQueue will attempt to create a queue of the provided type,
@ -127,7 +128,7 @@ func NewWrappedQueue(handle HandlerFunc, cfg, exemplar interface{}) (Queue, erro
name: config.Name,
},
}
_ = GetManager().Add(queue, WrappedQueueType, config, exemplar, nil)
_ = GetManager().Add(queue, WrappedQueueType, config, exemplar)
return queue, nil
}
@ -138,21 +139,78 @@ func (q *WrappedQueue) Name() string {
// Push will push the data to the internal channel checking it against the exemplar
func (q *WrappedQueue) Push(data Data) error {
if q.exemplar != nil {
// Assert data is of same type as r.exemplar
value := reflect.ValueOf(data)
t := value.Type()
exemplarType := reflect.ValueOf(q.exemplar).Type()
if !t.AssignableTo(exemplarType) || data == nil {
return fmt.Errorf("Unable to assign data: %v to same type as exemplar: %v in %s", data, q.exemplar, q.name)
}
if !assignableTo(data, q.exemplar) {
return fmt.Errorf("unable to assign data: %v to same type as exemplar: %v in %s", data, q.exemplar, q.name)
}
atomic.AddInt64(&q.numInQueue, 1)
q.channel <- data
return nil
}
func (q *WrappedQueue) flushInternalWithContext(ctx context.Context) error {
q.lock.Lock()
if q.internal == nil {
q.lock.Unlock()
return fmt.Errorf("not ready to flush wrapped queue %s yet", q.Name())
}
q.lock.Unlock()
select {
case <-ctx.Done():
return ctx.Err()
default:
}
return q.internal.FlushWithContext(ctx)
}
// Flush flushes the queue and blocks till the queue is empty
func (q *WrappedQueue) Flush(timeout time.Duration) error {
var ctx context.Context
var cancel context.CancelFunc
if timeout > 0 {
ctx, cancel = context.WithTimeout(context.Background(), timeout)
} else {
ctx, cancel = context.WithCancel(context.Background())
}
defer cancel()
return q.FlushWithContext(ctx)
}
// FlushWithContext implements the final part of Flushable
func (q *WrappedQueue) FlushWithContext(ctx context.Context) error {
log.Trace("WrappedQueue: %s FlushWithContext", q.Name())
errChan := make(chan error, 1)
go func() {
errChan <- q.flushInternalWithContext(ctx)
close(errChan)
}()
select {
case err := <-errChan:
return err
case <-ctx.Done():
go func() {
<-errChan
}()
return ctx.Err()
}
}
// IsEmpty checks whether the queue is empty
func (q *WrappedQueue) IsEmpty() bool {
if atomic.LoadInt64(&q.numInQueue) != 0 {
return false
}
q.lock.Lock()
defer q.lock.Unlock()
if q.internal == nil {
return false
}
return q.internal.IsEmpty()
}
// Run starts to run the queue and attempts to create the internal queue
func (q *WrappedQueue) Run(atShutdown, atTerminate func(context.Context, func())) {
log.Debug("WrappedQueue: %s Starting", q.name)
q.lock.Lock()
if q.internal == nil {
err := q.setInternal(atShutdown, q.handle, q.exemplar)
@ -164,6 +222,7 @@ func (q *WrappedQueue) Run(atShutdown, atTerminate func(context.Context, func())
go func() {
for data := range q.channel {
_ = q.internal.Push(data)
atomic.AddInt64(&q.numInQueue, -1)
}
}()
} else {
@ -176,7 +235,7 @@ func (q *WrappedQueue) Run(atShutdown, atTerminate func(context.Context, func())
// Shutdown this queue and stop processing
func (q *WrappedQueue) Shutdown() {
log.Trace("WrappedQueue: %s Shutdown", q.name)
log.Trace("WrappedQueue: %s Shutting down", q.name)
q.lock.Lock()
defer q.lock.Unlock()
if q.internal == nil {
@ -185,6 +244,7 @@ func (q *WrappedQueue) Shutdown() {
if shutdownable, ok := q.internal.(Shutdownable); ok {
shutdownable.Shutdown()
}
log.Debug("WrappedQueue: %s Shutdown", q.name)
}
// Terminate this queue and close the queue
@ -198,6 +258,7 @@ func (q *WrappedQueue) Terminate() {
if shutdownable, ok := q.internal.(Shutdownable); ok {
shutdownable.Terminate()
}
log.Debug("WrappedQueue: %s Terminated", q.name)
}
func init() {

View File

@ -24,8 +24,7 @@ func validType(t string) (Type, error) {
return PersistableChannelQueueType, fmt.Errorf("Unknown queue type: %s defaulting to %s", t, string(PersistableChannelQueueType))
}
// CreateQueue for name with provided handler and exemplar
func CreateQueue(name string, handle HandlerFunc, exemplar interface{}) Queue {
func getQueueSettings(name string) (setting.QueueSettings, []byte) {
q := setting.GetQueueSettings(name)
opts := make(map[string]interface{})
opts["Name"] = name
@ -43,24 +42,33 @@ func CreateQueue(name string, handle HandlerFunc, exemplar interface{}) Queue {
opts["BoostTimeout"] = q.BoostTimeout
opts["BoostWorkers"] = q.BoostWorkers
typ, err := validType(q.Type)
if err != nil {
log.Error("Invalid type %s provided for queue named %s defaulting to %s", q.Type, name, string(typ))
}
cfg, err := json.Marshal(opts)
if err != nil {
log.Error("Unable to marshall generic options: %v Error: %v", opts, err)
log.Error("Unable to create queue for %s", name, err)
return q, []byte{}
}
return q, cfg
}
// CreateQueue for name with provided handler and exemplar
func CreateQueue(name string, handle HandlerFunc, exemplar interface{}) Queue {
q, cfg := getQueueSettings(name)
if len(cfg) == 0 {
return nil
}
typ, err := validType(q.Type)
if err != nil {
log.Error("Invalid type %s provided for queue named %s defaulting to %s", q.Type, name, string(typ))
}
returnable, err := NewQueue(typ, handle, cfg, exemplar)
if q.WrapIfNecessary && err != nil {
log.Warn("Unable to create queue for %s: %v", name, err)
log.Warn("Attempting to create wrapped queue")
returnable, err = NewQueue(WrappedQueueType, handle, WrappedQueueConfiguration{
Underlying: Type(q.Type),
Underlying: typ,
Timeout: q.Timeout,
MaxAttempts: q.MaxAttempts,
Config: cfg,

View File

@ -7,12 +7,16 @@ package queue
import (
"context"
"sync"
"sync/atomic"
"time"
"code.gitea.io/gitea/modules/log"
)
// WorkerPool takes
// WorkerPool represent a dynamically growable worker pool for a
// provided handler function. They have an internal channel which
// they use to detect if there is a block and will grow and shrink in
// response to demand as per configuration.
type WorkerPool struct {
lock sync.Mutex
baseCtx context.Context
@ -27,10 +31,42 @@ type WorkerPool struct {
blockTimeout time.Duration
boostTimeout time.Duration
boostWorkers int
numInQueue int64
}
// WorkerPoolConfiguration is the basic configuration for a WorkerPool
type WorkerPoolConfiguration struct {
QueueLength int
BatchLength int
BlockTimeout time.Duration
BoostTimeout time.Duration
BoostWorkers int
MaxWorkers int
}
// NewWorkerPool creates a new worker pool
func NewWorkerPool(handle HandlerFunc, config WorkerPoolConfiguration) *WorkerPool {
ctx, cancel := context.WithCancel(context.Background())
dataChan := make(chan Data, config.QueueLength)
pool := &WorkerPool{
baseCtx: ctx,
cancel: cancel,
batchLength: config.BatchLength,
dataChan: dataChan,
handle: handle,
blockTimeout: config.BlockTimeout,
boostTimeout: config.BoostTimeout,
boostWorkers: config.BoostWorkers,
maxNumberOfWorkers: config.MaxWorkers,
}
return pool
}
// Push pushes the data to the internal channel
func (p *WorkerPool) Push(data Data) {
atomic.AddInt64(&p.numInQueue, 1)
p.lock.Lock()
if p.blockTimeout > 0 && p.boostTimeout > 0 && (p.numberOfWorkers <= p.maxNumberOfWorkers || p.maxNumberOfWorkers < 0) {
p.lock.Unlock()
@ -80,7 +116,7 @@ func (p *WorkerPool) pushBoost(data Data) {
log.Warn("WorkerPool: %d (for %s) Channel blocked for %v - adding %d temporary workers for %s, block timeout now %v", p.qid, mq.Name, ourTimeout, boost, p.boostTimeout, p.blockTimeout)
start := time.Now()
pid := mq.RegisterWorkers(boost, start, false, start, cancel)
pid := mq.RegisterWorkers(boost, start, false, start, cancel, false)
go func() {
<-ctx.Done()
mq.RemoveWorkers(pid)
@ -138,8 +174,8 @@ func (p *WorkerPool) BlockTimeout() time.Duration {
return p.blockTimeout
}
// SetSettings sets the setable boost values
func (p *WorkerPool) SetSettings(maxNumberOfWorkers, boostWorkers int, timeout time.Duration) {
// SetPoolSettings sets the setable boost values
func (p *WorkerPool) SetPoolSettings(maxNumberOfWorkers, boostWorkers int, timeout time.Duration) {
p.lock.Lock()
defer p.lock.Unlock()
p.maxNumberOfWorkers = maxNumberOfWorkers
@ -156,8 +192,7 @@ func (p *WorkerPool) SetMaxNumberOfWorkers(newMax int) {
p.maxNumberOfWorkers = newMax
}
// AddWorkers adds workers to the pool - this allows the number of workers to go above the limit
func (p *WorkerPool) AddWorkers(number int, timeout time.Duration) context.CancelFunc {
func (p *WorkerPool) commonRegisterWorkers(number int, timeout time.Duration, isFlusher bool) (context.Context, context.CancelFunc) {
var ctx context.Context
var cancel context.CancelFunc
start := time.Now()
@ -173,7 +208,7 @@ func (p *WorkerPool) AddWorkers(number int, timeout time.Duration) context.Cance
mq := GetManager().GetManagedQueue(p.qid)
if mq != nil {
pid := mq.RegisterWorkers(number, start, hasTimeout, end, cancel)
pid := mq.RegisterWorkers(number, start, hasTimeout, end, cancel, isFlusher)
go func() {
<-ctx.Done()
mq.RemoveWorkers(pid)
@ -184,6 +219,12 @@ func (p *WorkerPool) AddWorkers(number int, timeout time.Duration) context.Cance
log.Trace("WorkerPool: %d adding %d workers (no group id)", p.qid, number)
}
return ctx, cancel
}
// AddWorkers adds workers to the pool - this allows the number of workers to go above the limit
func (p *WorkerPool) AddWorkers(number int, timeout time.Duration) context.CancelFunc {
ctx, cancel := p.commonRegisterWorkers(number, timeout, false)
p.addWorkers(ctx, number)
return cancel
}
@ -235,6 +276,7 @@ func (p *WorkerPool) CleanUp(ctx context.Context) {
close(p.dataChan)
for data := range p.dataChan {
p.handle(data)
atomic.AddInt64(&p.numInQueue, -1)
select {
case <-ctx.Done():
log.Warn("WorkerPool: %d Cleanup context closed before finishing clean-up", p.qid)
@ -245,6 +287,37 @@ func (p *WorkerPool) CleanUp(ctx context.Context) {
log.Trace("WorkerPool: %d CleanUp Done", p.qid)
}
// Flush flushes the channel with a timeout - the Flush worker will be registered as a flush worker with the manager
func (p *WorkerPool) Flush(timeout time.Duration) error {
ctx, cancel := p.commonRegisterWorkers(1, timeout, true)
defer cancel()
return p.FlushWithContext(ctx)
}
// IsEmpty returns if true if the worker queue is empty
func (p *WorkerPool) IsEmpty() bool {
return atomic.LoadInt64(&p.numInQueue) == 0
}
// FlushWithContext is very similar to CleanUp but it will return as soon as the dataChan is empty
// NB: The worker will not be registered with the manager.
func (p *WorkerPool) FlushWithContext(ctx context.Context) error {
log.Trace("WorkerPool: %d Flush", p.qid)
for {
select {
case data := <-p.dataChan:
p.handle(data)
atomic.AddInt64(&p.numInQueue, -1)
case <-p.baseCtx.Done():
return p.baseCtx.Err()
case <-ctx.Done():
return ctx.Err()
default:
return nil
}
}
}
func (p *WorkerPool) doWork(ctx context.Context) {
delay := time.Millisecond * 300
var data = make([]Data, 0, p.batchLength)
@ -254,6 +327,7 @@ func (p *WorkerPool) doWork(ctx context.Context) {
if len(data) > 0 {
log.Trace("Handling: %d data, %v", len(data), data)
p.handle(data...)
atomic.AddInt64(&p.numInQueue, -1*int64(len(data)))
}
log.Trace("Worker shutting down")
return
@ -263,6 +337,7 @@ func (p *WorkerPool) doWork(ctx context.Context) {
if len(data) > 0 {
log.Trace("Handling: %d data, %v", len(data), data)
p.handle(data...)
atomic.AddInt64(&p.numInQueue, -1*int64(len(data)))
}
log.Trace("Worker shutting down")
return
@ -271,6 +346,7 @@ func (p *WorkerPool) doWork(ctx context.Context) {
if len(data) >= p.batchLength {
log.Trace("Handling: %d data, %v", len(data), data)
p.handle(data...)
atomic.AddInt64(&p.numInQueue, -1*int64(len(data)))
data = make([]Data, 0, p.batchLength)
}
default:
@ -286,6 +362,7 @@ func (p *WorkerPool) doWork(ctx context.Context) {
if len(data) > 0 {
log.Trace("Handling: %d data, %v", len(data), data)
p.handle(data...)
atomic.AddInt64(&p.numInQueue, -1*int64(len(data)))
}
log.Trace("Worker shutting down")
return
@ -301,6 +378,7 @@ func (p *WorkerPool) doWork(ctx context.Context) {
if len(data) > 0 {
log.Trace("Handling: %d data, %v", len(data), data)
p.handle(data...)
atomic.AddInt64(&p.numInQueue, -1*int64(len(data)))
}
log.Trace("Worker shutting down")
return
@ -309,6 +387,7 @@ func (p *WorkerPool) doWork(ctx context.Context) {
if len(data) >= p.batchLength {
log.Trace("Handling: %d data, %v", len(data), data)
p.handle(data...)
atomic.AddInt64(&p.numInQueue, -1*int64(len(data)))
data = make([]Data, 0, p.batchLength)
}
case <-timer.C:
@ -316,6 +395,7 @@ func (p *WorkerPool) doWork(ctx context.Context) {
if len(data) > 0 {
log.Trace("Handling: %d data, %v", len(data), data)
p.handle(data...)
atomic.AddInt64(&p.numInQueue, -1*int64(len(data)))
data = make([]Data, 0, p.batchLength)
}

View File

@ -59,7 +59,7 @@ func GetQueueSettings(name string) QueueSettings {
if !filepath.IsAbs(q.DataDir) {
q.DataDir = filepath.Join(AppDataPath, q.DataDir)
}
sec.Key("DATADIR").SetValue(q.DataDir)
_, _ = sec.NewKey("DATADIR", q.DataDir)
// The rest are...
q.Length = sec.Key("LENGTH").MustInt(Queue.Length)
q.BatchLength = sec.Key("BATCH_LENGTH").MustInt(Queue.BatchLength)
@ -89,7 +89,7 @@ func NewQueueService() {
Queue.Length = sec.Key("LENGTH").MustInt(20)
Queue.BatchLength = sec.Key("BATCH_LENGTH").MustInt(20)
Queue.ConnectionString = sec.Key("CONN_STR").MustString(path.Join(AppDataPath, ""))
Queue.Type = sec.Key("TYPE").MustString("")
Queue.Type = sec.Key("TYPE").MustString("persistable-channel")
Queue.Network, Queue.Addresses, Queue.Password, Queue.DBIndex, _ = ParseQueueConnStr(Queue.ConnectionString)
Queue.WrapIfNecessary = sec.Key("WRAP_IF_NECESSARY").MustBool(true)
Queue.MaxAttempts = sec.Key("MAX_ATTEMPTS").MustInt(10)
@ -110,27 +110,27 @@ func NewQueueService() {
if _, ok := sectionMap["TYPE"]; !ok {
switch Indexer.IssueQueueType {
case LevelQueueType:
section.Key("TYPE").SetValue("level")
_, _ = section.NewKey("TYPE", "level")
case ChannelQueueType:
section.Key("TYPE").SetValue("persistable-channel")
_, _ = section.NewKey("TYPE", "persistable-channel")
case RedisQueueType:
section.Key("TYPE").SetValue("redis")
_, _ = section.NewKey("TYPE", "redis")
default:
log.Fatal("Unsupported indexer queue type: %v",
Indexer.IssueQueueType)
}
}
if _, ok := sectionMap["LENGTH"]; !ok {
section.Key("LENGTH").SetValue(fmt.Sprintf("%d", Indexer.UpdateQueueLength))
_, _ = section.NewKey("LENGTH", fmt.Sprintf("%d", Indexer.UpdateQueueLength))
}
if _, ok := sectionMap["BATCH_LENGTH"]; !ok {
section.Key("BATCH_LENGTH").SetValue(fmt.Sprintf("%d", Indexer.IssueQueueBatchNumber))
_, _ = section.NewKey("BATCH_LENGTH", fmt.Sprintf("%d", Indexer.IssueQueueBatchNumber))
}
if _, ok := sectionMap["DATADIR"]; !ok {
section.Key("DATADIR").SetValue(Indexer.IssueQueueDir)
_, _ = section.NewKey("DATADIR", Indexer.IssueQueueDir)
}
if _, ok := sectionMap["CONN_STR"]; !ok {
section.Key("CONN_STR").SetValue(Indexer.IssueQueueConnStr)
_, _ = section.NewKey("CONN_STR", Indexer.IssueQueueConnStr)
}
// Handle the old mailer configuration
@ -140,7 +140,7 @@ func NewQueueService() {
sectionMap[key.Name()] = true
}
if _, ok := sectionMap["LENGTH"]; !ok {
section.Key("LENGTH").SetValue(fmt.Sprintf("%d", Cfg.Section("mailer").Key("SEND_BUFFER_LEN").MustInt(100)))
_, _ = section.NewKey("LENGTH", fmt.Sprintf("%d", Cfg.Section("mailer").Key("SEND_BUFFER_LEN").MustInt(100)))
}
}

View File

@ -2083,6 +2083,10 @@ monitor.queue.pool.addworkers.numberworkers.placeholder = Number of Workers
monitor.queue.pool.addworkers.timeout.placeholder = Set to 0 for no timeout
monitor.queue.pool.addworkers.mustnumbergreaterzero = Number of Workers to add must be greater than zero
monitor.queue.pool.addworkers.musttimeoutduration = Timeout must be a golang duration eg. 5m or be 0
monitor.queue.pool.flush.title = Flush Queue
monitor.queue.pool.flush.desc = Flush will add a worker that will terminate once the queue is empty, or it times out.
monitor.queue.pool.flush.submit = Add Flush Worker
monitor.queue.pool.flush.added = Flush Worker added for %[1]s
monitor.queue.settings.title = Pool Settings
monitor.queue.settings.desc = Pools dynamically grow with a boost in response to their worker queue blocking. These changes will not affect current worker groups.

View File

@ -404,6 +404,28 @@ func WorkerCancel(ctx *context.Context) {
})
}
// Flush flushes a queue
func Flush(ctx *context.Context) {
qid := ctx.ParamsInt64("qid")
mq := queue.GetManager().GetManagedQueue(qid)
if mq == nil {
ctx.Status(404)
return
}
timeout, err := time.ParseDuration(ctx.Query("timeout"))
if err != nil {
timeout = -1
}
ctx.Flash.Info(ctx.Tr("admin.monitor.queue.pool.flush.added", mq.Name))
go func() {
err := mq.Flush(timeout)
if err != nil {
log.Error("Flushing failure for %s: Error %v", mq.Name, err)
}
}()
ctx.Redirect(setting.AppSubURL + fmt.Sprintf("/admin/monitor/queue/%d", qid))
}
// AddWorkers adds workers to a worker group
func AddWorkers(ctx *context.Context) {
qid := ctx.ParamsInt64("qid")
@ -424,7 +446,7 @@ func AddWorkers(ctx *context.Context) {
ctx.Redirect(setting.AppSubURL + fmt.Sprintf("/admin/monitor/queue/%d", qid))
return
}
if mq.Pool == nil {
if _, ok := mq.Managed.(queue.ManagedPool); !ok {
ctx.Flash.Error(ctx.Tr("admin.monitor.queue.pool.none"))
ctx.Redirect(setting.AppSubURL + fmt.Sprintf("/admin/monitor/queue/%d", qid))
return
@ -442,7 +464,7 @@ func SetQueueSettings(ctx *context.Context) {
ctx.Status(404)
return
}
if mq.Pool == nil {
if _, ok := mq.Managed.(queue.ManagedPool); !ok {
ctx.Flash.Error(ctx.Tr("admin.monitor.queue.pool.none"))
ctx.Redirect(setting.AppSubURL + fmt.Sprintf("/admin/monitor/queue/%d", qid))
return
@ -488,10 +510,10 @@ func SetQueueSettings(ctx *context.Context) {
return
}
} else {
timeout = mq.Pool.BoostTimeout()
timeout = mq.BoostTimeout()
}
mq.SetSettings(maxNumber, number, timeout)
mq.SetPoolSettings(maxNumber, number, timeout)
ctx.Flash.Success(ctx.Tr("admin.monitor.queue.settings.changed"))
ctx.Redirect(setting.AppSubURL + fmt.Sprintf("/admin/monitor/queue/%d", qid))
}

View File

@ -89,5 +89,9 @@ func RegisterRoutes(m *macaron.Macaron) {
m.Post("/hook/set-default-branch/:owner/:repo/:branch", SetDefaultBranch)
m.Get("/serv/none/:keyid", ServNoCommand)
m.Get("/serv/command/:keyid/:owner/:repo", ServCommand)
m.Post("/manager/shutdown", Shutdown)
m.Post("/manager/restart", Restart)
m.Post("/manager/flush-queues", bind(private.FlushOptions{}), FlushQueues)
}, CheckInternalToken)
}

View File

@ -0,0 +1,41 @@
// Copyright 2020 The Gitea Authors. All rights reserved.
// Use of this source code is governed by a MIT-style
// license that can be found in the LICENSE file.
package private
import (
"net/http"
"code.gitea.io/gitea/modules/graceful"
"code.gitea.io/gitea/modules/log"
"code.gitea.io/gitea/modules/private"
"code.gitea.io/gitea/modules/queue"
"gitea.com/macaron/macaron"
)
// FlushQueues flushes all the Queues
func FlushQueues(ctx *macaron.Context, opts private.FlushOptions) {
if opts.NonBlocking {
// Save the hammer ctx here - as a new one is created each time you call this.
baseCtx := graceful.GetManager().HammerContext()
go func() {
err := queue.GetManager().FlushAll(baseCtx, opts.Timeout)
if err != nil {
log.Error("Flushing request timed-out with error: %v", err)
}
}()
ctx.JSON(http.StatusAccepted, map[string]interface{}{
"err": "Flushing",
})
return
}
err := queue.GetManager().FlushAll(ctx.Req.Request.Context(), opts.Timeout)
if err != nil {
ctx.JSON(http.StatusRequestTimeout, map[string]interface{}{
"err": err,
})
}
ctx.PlainText(http.StatusOK, []byte("success"))
}

View File

@ -0,0 +1,28 @@
// +build !windows
// Copyright 2020 The Gitea Authors. All rights reserved.
// Use of this source code is governed by a MIT-style
// license that can be found in the LICENSE file.
package private
import (
"net/http"
"code.gitea.io/gitea/modules/graceful"
"gitea.com/macaron/macaron"
)
// Restart causes the server to perform a graceful restart
func Restart(ctx *macaron.Context) {
graceful.GetManager().DoGracefulRestart()
ctx.PlainText(http.StatusOK, []byte("success"))
}
// Shutdown causes the server to perform a graceful shutdown
func Shutdown(ctx *macaron.Context) {
graceful.GetManager().DoGracefulShutdown()
ctx.PlainText(http.StatusOK, []byte("success"))
}

View File

@ -0,0 +1,28 @@
// +build windows
// Copyright 2020 The Gitea Authors. All rights reserved.
// Use of this source code is governed by a MIT-style
// license that can be found in the LICENSE file.
package private
import (
"net/http"
"code.gitea.io/gitea/modules/graceful"
"gitea.com/macaron/macaron"
)
// Restart is not implemented for Windows based servers as they can't fork
func Restart(ctx *macaron.Context) {
ctx.JSON(http.StatusNotImplemented, map[string]interface{}{
"err": "windows servers cannot be gracefully restarted - shutdown and restart manually",
})
}
// Shutdown causes the server to perform a graceful shutdown
func Shutdown(ctx *macaron.Context) {
graceful.GetManager().DoGracefulShutdown()
ctx.PlainText(http.StatusOK, []byte("success"))
}

View File

@ -423,6 +423,7 @@ func RegisterRoutes(m *macaron.Macaron) {
m.Post("/set", admin.SetQueueSettings)
m.Post("/add", admin.AddWorkers)
m.Post("/cancel/:pid", admin.WorkerCancel)
m.Post("/flush", admin.Flush)
})
})

View File

@ -92,6 +92,24 @@
</div>
</form>
</div>
<h4 class="ui top attached header">
{{.i18n.Tr "admin.monitor.queue.pool.flush.title"}}
</h4>
<div class="ui attached segment">
<p>{{.i18n.Tr "admin.monitor.queue.pool.flush.desc"}}</p>
<form method="POST" action="{{.Link}}/flush">
{{$.CsrfTokenHtml}}
<div class="ui form">
<div class="fields">
<div class="field">
<label>{{.i18n.Tr "admin.monitor.queue.pool.timeout"}}</label>
<input name="timeout" type="text" placeholder="{{.i18n.Tr "admin.monitor.queue.pool.addworkers.timeout.placeholder"}}">
</div>
</div>
<button class="ui submit button">{{.i18n.Tr "admin.monitor.queue.pool.flush.submit"}}</button>
</div>
</form>
</div>
<h4 class="ui top attached header">
{{.i18n.Tr "admin.monitor.queue.pool.workers.title"}}
</h4>
@ -108,7 +126,7 @@
<tbody>
{{range .Queue.Workers}}
<tr>
<td>{{.Workers}}</td>
<td>{{.Workers}}{{if .IsFlusher}}<i class="icon sync-alternate" title="{{.i18n.Tr "admin.monitor.queue.flush"}}"></i>{{end}}</td>
<td>{{DateFmtLong .Start}}</td>
<td>{{if .HasTimeout}}{{DateFmtLong .Timeout}}{{else}}-{{end}}</td>
<td>