open-nomad/nomad/worker_test.go
Michael Schurter ace5faf948
core: backoff considerably when worker is behind raft (#15523)
Upon dequeuing an evaluation workers snapshot their state store at the
eval's wait index or later. This ensures we process an eval at a point
in time after it was created or updated. Processing an eval on an old
snapshot could cause any number of problems such as:

1. Since job registration atomically updates an eval and job in a single
   raft entry, scheduling against indexes before that may not have the
   eval's job or may have an older version.
2. The older the scheduler's snapshot, the higher the likelihood
   something has changed in the cluster state which will cause the plan
   applier to reject the scheduler's plan. This could waste work or
   even cause eval's to be failed needlessly.

However, the workers run in parallel with a new server pulling the
cluster state from a peer. During this time, which may be many minutes
long, the state store is likely far behind the minimum index required
to process evaluations.

This PR addresses this by adding an additional long backoff period after
an eval is nacked. If the scheduler's indexes catches up within the
additional backoff, it will unblock early to dequeue the next eval.

When the server shuts down we'll get a `context.Canceled` error from the state
store method. We need to bubble this error up so that other callers can detect
it. Handle this case separately when waiting after dequeue so that we can warn
on shutdown instead of throwing an ambiguous error message with just the text
"canceled."

While there may be more precise ways to block scheduling until the
server catches up, this approach adds little risk and covers additional
cases where a server may be temporarily behind due to a spike in load or
a saturated network.

For testing, we make the `raftSyncLimit` into a parameter on the worker's `run` method 
so that we can run backoff tests without waiting 30+ seconds. We haven't followed thru
and made all the worker globals into worker parameters, because there isn't much
use outside of testing, but we can consider that in the future.

Co-authored-by: Tim Gross <tgross@hashicorp.com>
2023-01-24 08:56:35 -05:00

967 lines
24 KiB
Go

package nomad
import (
"context"
"errors"
"fmt"
"reflect"
"sync"
"testing"
"time"
log "github.com/hashicorp/go-hclog"
"github.com/hashicorp/go-memdb"
"github.com/hashicorp/nomad/ci"
"github.com/hashicorp/nomad/helper"
"github.com/shoenig/test/must"
"github.com/stretchr/testify/require"
"github.com/hashicorp/nomad/helper/testlog"
"github.com/hashicorp/nomad/helper/uuid"
"github.com/hashicorp/nomad/nomad/mock"
"github.com/hashicorp/nomad/nomad/structs"
"github.com/hashicorp/nomad/scheduler"
"github.com/hashicorp/nomad/testutil"
"github.com/stretchr/testify/assert"
)
type NoopScheduler struct {
state scheduler.State
planner scheduler.Planner
eval *structs.Evaluation
eventsCh chan<- interface{}
err error
}
func (n *NoopScheduler) Process(eval *structs.Evaluation) error {
if n.state == nil {
panic("missing state")
}
if n.planner == nil {
panic("missing planner")
}
n.eval = eval
return n.err
}
func init() {
scheduler.BuiltinSchedulers["noop"] = func(logger log.Logger, eventsCh chan<- interface{}, s scheduler.State, p scheduler.Planner) scheduler.Scheduler {
n := &NoopScheduler{
state: s,
planner: p,
}
return n
}
}
// NewTestWorker returns the worker without calling it's run method.
func NewTestWorker(shutdownCtx context.Context, srv *Server) *Worker {
w := &Worker{
srv: srv,
start: time.Now(),
id: uuid.Generate(),
enabledSchedulers: srv.config.EnabledSchedulers,
}
w.logger = srv.logger.ResetNamed("worker").With("worker_id", w.id)
w.pauseCond = sync.NewCond(&w.pauseLock)
w.ctx, w.cancelFn = context.WithCancel(shutdownCtx)
return w
}
func TestWorker_dequeueEvaluation(t *testing.T) {
ci.Parallel(t)
s1, cleanupS1 := TestServer(t, func(c *Config) {
c.NumSchedulers = 0
c.EnabledSchedulers = []string{structs.JobTypeService}
})
defer cleanupS1()
testutil.WaitForLeader(t, s1.RPC)
// Create the evaluation
eval1 := mock.Eval()
s1.evalBroker.Enqueue(eval1)
// Create a worker
poolArgs := getSchedulerWorkerPoolArgsFromConfigLocked(s1.config).Copy()
w, _ := NewWorker(s1.shutdownCtx, s1, poolArgs)
// Attempt dequeue
eval, token, waitIndex, shutdown := w.dequeueEvaluation(10 * time.Millisecond)
if shutdown {
t.Fatalf("should not shutdown")
}
if token == "" {
t.Fatalf("should get token")
}
if waitIndex != eval1.ModifyIndex {
t.Fatalf("bad wait index; got %d; want %d", waitIndex, eval1.ModifyIndex)
}
// Ensure we get a sane eval
if !reflect.DeepEqual(eval, eval1) {
t.Fatalf("bad: %#v %#v", eval, eval1)
}
}
// Test that the worker picks up the correct wait index when there are multiple
// evals for the same job.
func TestWorker_dequeueEvaluation_SerialJobs(t *testing.T) {
ci.Parallel(t)
s1, cleanupS1 := TestServer(t, func(c *Config) {
c.NumSchedulers = 0
c.EnabledSchedulers = []string{structs.JobTypeService}
})
defer cleanupS1()
testutil.WaitForLeader(t, s1.RPC)
// Create the evaluation
eval1 := mock.Eval()
eval2 := mock.Eval()
eval2.JobID = eval1.JobID
// Insert the evals into the state store
must.NoError(t, s1.fsm.State().UpsertEvals(
structs.MsgTypeTestSetup, 1000, []*structs.Evaluation{eval1}))
must.NoError(t, s1.fsm.State().UpsertEvals(
structs.MsgTypeTestSetup, 2000, []*structs.Evaluation{eval2}))
s1.evalBroker.Enqueue(eval1)
s1.evalBroker.Enqueue(eval2)
// Create a worker
poolArgs := getSchedulerWorkerPoolArgsFromConfigLocked(s1.config).Copy()
w := newWorker(s1.shutdownCtx, s1, poolArgs)
// Attempt dequeue
eval, token, waitIndex, shutdown := w.dequeueEvaluation(10 * time.Millisecond)
must.False(t, shutdown, must.Sprint("should not be shutdown"))
must.NotEq(t, token, "", must.Sprint("should get a token"))
must.NotEq(t, eval1.ModifyIndex, waitIndex, must.Sprintf("bad wait index"))
must.Eq(t, eval, eval1)
// Update the modify index of the first eval
must.NoError(t, s1.fsm.State().UpsertEvals(
structs.MsgTypeTestSetup, 1500, []*structs.Evaluation{eval1}))
// Send the Ack
w.sendAck(eval1, token)
// Attempt second dequeue; it should succeed because the 2nd eval has a
// lower modify index than the snapshot used to schedule the 1st
// eval. Normally this can only happen if the worker is on a follower that's
// trailing behind in raft logs
eval, token, waitIndex, shutdown = w.dequeueEvaluation(10 * time.Millisecond)
must.False(t, shutdown, must.Sprint("should not be shutdown"))
must.NotEq(t, token, "", must.Sprint("should get a token"))
must.Eq(t, waitIndex, 2000, must.Sprintf("bad wait index"))
must.Eq(t, eval, eval2)
}
func TestWorker_dequeueEvaluation_paused(t *testing.T) {
ci.Parallel(t)
s1, cleanupS1 := TestServer(t, func(c *Config) {
c.NumSchedulers = 0
c.EnabledSchedulers = []string{structs.JobTypeService}
})
defer cleanupS1()
testutil.WaitForLeader(t, s1.RPC)
// Create the evaluation
eval1 := mock.Eval()
s1.evalBroker.Enqueue(eval1)
// Create a worker
poolArgs := getSchedulerWorkerPoolArgsFromConfigLocked(s1.config).Copy()
w := newWorker(s1.shutdownCtx, s1, poolArgs)
w.pauseCond = sync.NewCond(&w.pauseLock)
// PAUSE the worker
w.Pause()
go func() {
time.Sleep(100 * time.Millisecond)
w.Resume()
}()
// Attempt dequeue
start := time.Now()
eval, token, waitIndex, shutdown := w.dequeueEvaluation(10 * time.Millisecond)
if diff := time.Since(start); diff < 100*time.Millisecond {
t.Fatalf("should have paused: %v", diff)
}
if shutdown {
t.Fatalf("should not shutdown")
}
if token == "" {
t.Fatalf("should get token")
}
if waitIndex != eval1.ModifyIndex {
t.Fatalf("bad wait index; got %d; want %d", waitIndex, eval1.ModifyIndex)
}
// Ensure we get a sane eval
if !reflect.DeepEqual(eval, eval1) {
t.Fatalf("bad: %#v %#v", eval, eval1)
}
}
func TestWorker_dequeueEvaluation_shutdown(t *testing.T) {
ci.Parallel(t)
s1, cleanupS1 := TestServer(t, func(c *Config) {
c.NumSchedulers = 0
c.EnabledSchedulers = []string{structs.JobTypeService}
})
defer cleanupS1()
testutil.WaitForLeader(t, s1.RPC)
// Create a worker
poolArgs := getSchedulerWorkerPoolArgsFromConfigLocked(s1.config).Copy()
w := newWorker(s1.shutdownCtx, s1, poolArgs)
go func() {
time.Sleep(10 * time.Millisecond)
s1.Shutdown()
}()
// Attempt dequeue
eval, _, _, shutdown := w.dequeueEvaluation(10 * time.Millisecond)
if !shutdown {
t.Fatalf("should not shutdown")
}
// Ensure we get a sane eval
if eval != nil {
t.Fatalf("bad: %#v", eval)
}
}
func TestWorker_Shutdown(t *testing.T) {
ci.Parallel(t)
s1, cleanupS1 := TestServer(t, func(c *Config) {
c.NumSchedulers = 0
c.EnabledSchedulers = []string{structs.JobTypeService}
})
defer cleanupS1()
testutil.WaitForLeader(t, s1.RPC)
poolArgs := getSchedulerWorkerPoolArgsFromConfigLocked(s1.config).Copy()
w := newWorker(s1.shutdownCtx, s1, poolArgs)
go func() {
time.Sleep(10 * time.Millisecond)
w.Stop()
}()
// Attempt dequeue
eval, _, _, shutdown := w.dequeueEvaluation(10 * time.Millisecond)
require.True(t, shutdown)
require.Nil(t, eval)
}
func TestWorker_Shutdown_paused(t *testing.T) {
ci.Parallel(t)
s1, cleanupS1 := TestServer(t, func(c *Config) {
c.NumSchedulers = 0
c.EnabledSchedulers = []string{structs.JobTypeService}
})
defer cleanupS1()
testutil.WaitForLeader(t, s1.RPC)
poolArgs := getSchedulerWorkerPoolArgsFromConfigLocked(s1.config).Copy()
w, _ := NewWorker(s1.shutdownCtx, s1, poolArgs)
w.Pause()
// pausing can take up to 500ms because of the blocking query timeout in dequeueEvaluation.
require.Eventually(t, w.IsPaused, 550*time.Millisecond, 10*time.Millisecond, "should pause")
go func() {
w.Stop()
}()
// transitioning to stopped from paused should be very quick,
// but might not be immediate.
require.Eventually(t, w.IsStopped, 100*time.Millisecond, 10*time.Millisecond, "should stop when paused")
}
func TestWorker_sendAck(t *testing.T) {
ci.Parallel(t)
s1, cleanupS1 := TestServer(t, func(c *Config) {
c.NumSchedulers = 0
c.EnabledSchedulers = []string{structs.JobTypeService}
})
defer cleanupS1()
testutil.WaitForLeader(t, s1.RPC)
// Create the evaluation
eval1 := mock.Eval()
s1.evalBroker.Enqueue(eval1)
// Create a worker
poolArgs := getSchedulerWorkerPoolArgsFromConfigLocked(s1.config).Copy()
w := newWorker(s1.shutdownCtx, s1, poolArgs)
// Attempt dequeue
eval, token, _, _ := w.dequeueEvaluation(10 * time.Millisecond)
// Check the depth is 0, 1 unacked
stats := s1.evalBroker.Stats()
if stats.TotalReady != 0 && stats.TotalUnacked != 1 {
t.Fatalf("bad: %#v", stats)
}
// Send the Nack
w.sendNack(eval, token)
// Check the depth is 1, nothing unacked
stats = s1.evalBroker.Stats()
if stats.TotalReady != 1 && stats.TotalUnacked != 0 {
t.Fatalf("bad: %#v", stats)
}
// Attempt dequeue
eval, token, _, _ = w.dequeueEvaluation(10 * time.Millisecond)
// Send the Ack
w.sendAck(eval, token)
// Check the depth is 0
stats = s1.evalBroker.Stats()
if stats.TotalReady != 0 && stats.TotalUnacked != 0 {
t.Fatalf("bad: %#v", stats)
}
}
func TestWorker_runBackoff(t *testing.T) {
ci.Parallel(t)
srv, cleanupSrv := TestServer(t, func(c *Config) {
c.NumSchedulers = 0
c.EnabledSchedulers = []string{structs.JobTypeService}
})
defer cleanupSrv()
testutil.WaitForLeader(t, srv.RPC)
eval1 := mock.Eval()
eval1.ModifyIndex = 1000
srv.evalBroker.Enqueue(eval1)
must.Eq(t, 1, srv.evalBroker.Stats().TotalReady)
// make a new context here so we can still check the broker's state after
// we've shut down the worker
workerCtx, workerCancel := context.WithCancel(srv.shutdownCtx)
defer workerCancel()
w := NewTestWorker(workerCtx, srv)
doneCh := make(chan struct{})
go func() {
w.run(time.Millisecond)
doneCh <- struct{}{}
}()
// We expect to be paused for 10ms + 1ms but otherwise can't be all that
// precise here because of concurrency. But checking coverage for this test
// shows we've covered the logic
t1, cancelT1 := helper.NewSafeTimer(100 * time.Millisecond)
defer cancelT1()
select {
case <-doneCh:
t.Fatal("returned early")
case <-t1.C:
}
workerCancel()
<-doneCh
must.Eq(t, 1, srv.evalBroker.Stats().TotalWaiting)
must.Eq(t, 0, srv.evalBroker.Stats().TotalReady)
must.Eq(t, 0, srv.evalBroker.Stats().TotalPending)
must.Eq(t, 0, srv.evalBroker.Stats().TotalUnacked)
}
func TestWorker_waitForIndex(t *testing.T) {
ci.Parallel(t)
s1, cleanupS1 := TestServer(t, func(c *Config) {
c.NumSchedulers = 0
c.EnabledSchedulers = []string{structs.JobTypeService}
})
defer cleanupS1()
testutil.WaitForLeader(t, s1.RPC)
// Get the current index
index := s1.raft.AppliedIndex()
// Cause an increment
errCh := make(chan error, 1)
go func() {
time.Sleep(10 * time.Millisecond)
n := mock.Node()
errCh <- s1.fsm.state.UpsertNode(structs.MsgTypeTestSetup, index+1, n)
}()
// Wait for a future index
poolArgs := getSchedulerWorkerPoolArgsFromConfigLocked(s1.config).Copy()
w := newWorker(s1.shutdownCtx, s1, poolArgs)
snap, err := w.snapshotMinIndex(index+1, time.Second)
require.NoError(t, err)
require.NotNil(t, snap)
// No error from upserting
require.NoError(t, <-errCh)
// Cause a timeout
waitIndex := index + 100
timeout := 10 * time.Millisecond
snap, err = w.snapshotMinIndex(index+100, timeout)
require.Nil(t, snap)
require.EqualError(t, err,
fmt.Sprintf("timed out after %s waiting for index=%d", timeout, waitIndex))
require.True(t, errors.Is(err, context.DeadlineExceeded), "expect error to wrap DeadlineExceeded")
}
func TestWorker_invokeScheduler(t *testing.T) {
ci.Parallel(t)
s1, cleanupS1 := TestServer(t, func(c *Config) {
c.NumSchedulers = 0
c.EnabledSchedulers = []string{structs.JobTypeService}
})
defer cleanupS1()
poolArgs := getSchedulerWorkerPoolArgsFromConfigLocked(s1.config).Copy()
w := newWorker(s1.shutdownCtx, s1, poolArgs)
eval := mock.Eval()
eval.Type = "noop"
snap, err := s1.fsm.state.Snapshot()
require.NoError(t, err)
err = w.invokeScheduler(snap, eval, uuid.Generate())
require.NoError(t, err)
}
func TestWorker_SubmitPlan(t *testing.T) {
ci.Parallel(t)
s1, cleanupS1 := TestServer(t, func(c *Config) {
c.NumSchedulers = 0
c.EnabledSchedulers = []string{structs.JobTypeService}
})
defer cleanupS1()
testutil.WaitForLeader(t, s1.RPC)
// Register node
node := mock.Node()
testRegisterNode(t, s1, node)
job := mock.Job()
eval1 := mock.Eval()
eval1.JobID = job.ID
s1.fsm.State().UpsertJob(structs.MsgTypeTestSetup, 1000, job)
s1.fsm.State().UpsertEvals(structs.MsgTypeTestSetup, 1000, []*structs.Evaluation{eval1})
// Create the register request
s1.evalBroker.Enqueue(eval1)
evalOut, token, err := s1.evalBroker.Dequeue([]string{eval1.Type}, time.Second)
if err != nil {
t.Fatalf("err: %v", err)
}
if evalOut != eval1 {
t.Fatalf("Bad eval")
}
// Create an allocation plan
alloc := mock.Alloc()
plan := &structs.Plan{
Job: job,
EvalID: eval1.ID,
NodeAllocation: map[string][]*structs.Allocation{
node.ID: {alloc},
},
}
// Attempt to submit a plan
poolArgs := getSchedulerWorkerPoolArgsFromConfigLocked(s1.config).Copy()
w := newWorker(s1.shutdownCtx, s1, poolArgs)
w.evalToken = token
result, state, err := w.SubmitPlan(plan)
if err != nil {
t.Fatalf("err: %v", err)
}
// Should have no update
if state != nil {
t.Fatalf("unexpected state update")
}
// Result should have allocated
if result == nil {
t.Fatalf("missing result")
}
if result.AllocIndex == 0 {
t.Fatalf("Bad: %#v", result)
}
if len(result.NodeAllocation) != 1 {
t.Fatalf("Bad: %#v", result)
}
}
func TestWorker_SubmitPlanNormalizedAllocations(t *testing.T) {
ci.Parallel(t)
s1, cleanupS1 := TestServer(t, func(c *Config) {
c.NumSchedulers = 0
c.EnabledSchedulers = []string{structs.JobTypeService}
c.Build = "1.4.0"
})
defer cleanupS1()
testutil.WaitForLeader(t, s1.RPC)
// Register node
node := mock.Node()
testRegisterNode(t, s1, node)
job := mock.Job()
eval1 := mock.Eval()
eval1.JobID = job.ID
s1.fsm.State().UpsertJob(structs.MsgTypeTestSetup, 0, job)
s1.fsm.State().UpsertEvals(structs.MsgTypeTestSetup, 0, []*structs.Evaluation{eval1})
stoppedAlloc := mock.Alloc()
preemptedAlloc := mock.Alloc()
s1.fsm.State().UpsertAllocs(structs.MsgTypeTestSetup, 5, []*structs.Allocation{stoppedAlloc, preemptedAlloc})
// Create an allocation plan
plan := &structs.Plan{
Job: job,
EvalID: eval1.ID,
NodeUpdate: make(map[string][]*structs.Allocation),
NodePreemptions: make(map[string][]*structs.Allocation),
}
desiredDescription := "desired desc"
plan.AppendStoppedAlloc(stoppedAlloc, desiredDescription, structs.AllocClientStatusLost, "")
preemptingAllocID := uuid.Generate()
plan.AppendPreemptedAlloc(preemptedAlloc, preemptingAllocID)
// Attempt to submit a plan
poolArgs := getSchedulerWorkerPoolArgsFromConfigLocked(s1.config).Copy()
w := newWorker(s1.shutdownCtx, s1, poolArgs)
w.SubmitPlan(plan)
assert.Equal(t, &structs.Allocation{
ID: preemptedAlloc.ID,
PreemptedByAllocation: preemptingAllocID,
}, plan.NodePreemptions[preemptedAlloc.NodeID][0])
assert.Equal(t, &structs.Allocation{
ID: stoppedAlloc.ID,
DesiredDescription: desiredDescription,
ClientStatus: structs.AllocClientStatusLost,
}, plan.NodeUpdate[stoppedAlloc.NodeID][0])
}
func TestWorker_SubmitPlan_MissingNodeRefresh(t *testing.T) {
ci.Parallel(t)
s1, cleanupS1 := TestServer(t, func(c *Config) {
c.NumSchedulers = 0
c.EnabledSchedulers = []string{structs.JobTypeService}
})
defer cleanupS1()
testutil.WaitForLeader(t, s1.RPC)
// Register node
node := mock.Node()
testRegisterNode(t, s1, node)
// Create the job
job := mock.Job()
s1.fsm.State().UpsertJob(structs.MsgTypeTestSetup, 1000, job)
// Create the register request
eval1 := mock.Eval()
eval1.JobID = job.ID
s1.evalBroker.Enqueue(eval1)
evalOut, token, err := s1.evalBroker.Dequeue([]string{eval1.Type}, time.Second)
if err != nil {
t.Fatalf("err: %v", err)
}
if evalOut != eval1 {
t.Fatalf("Bad eval")
}
// Create an allocation plan, with unregistered node
node2 := mock.Node()
alloc := mock.Alloc()
plan := &structs.Plan{
Job: job,
EvalID: eval1.ID,
NodeAllocation: map[string][]*structs.Allocation{
node2.ID: {alloc},
},
}
// Attempt to submit a plan
poolArgs := getSchedulerWorkerPoolArgsFromConfigLocked(s1.config).Copy()
w := newWorker(s1.shutdownCtx, s1, poolArgs)
w.evalToken = token
result, state, err := w.SubmitPlan(plan)
if err != nil {
t.Fatalf("err: %v", err)
}
// Result should have allocated
if result == nil {
t.Fatalf("missing result")
}
// Expect no allocation and forced refresh
if result.AllocIndex != 0 {
t.Fatalf("Bad: %#v", result)
}
if result.RefreshIndex == 0 {
t.Fatalf("Bad: %#v", result)
}
if len(result.NodeAllocation) != 0 {
t.Fatalf("Bad: %#v", result)
}
// Should have an update
if state == nil {
t.Fatalf("expected state update")
}
}
func TestWorker_UpdateEval(t *testing.T) {
ci.Parallel(t)
s1, cleanupS1 := TestServer(t, func(c *Config) {
c.NumSchedulers = 0
c.EnabledSchedulers = []string{structs.JobTypeService}
})
defer cleanupS1()
testutil.WaitForLeader(t, s1.RPC)
// Register node
node := mock.Node()
testRegisterNode(t, s1, node)
// Create the register request
eval1 := mock.Eval()
s1.evalBroker.Enqueue(eval1)
evalOut, token, err := s1.evalBroker.Dequeue([]string{eval1.Type}, time.Second)
if err != nil {
t.Fatalf("err: %v", err)
}
if evalOut != eval1 {
t.Fatalf("Bad eval")
}
eval2 := evalOut.Copy()
eval2.Status = structs.EvalStatusComplete
// Attempt to update eval
poolArgs := getSchedulerWorkerPoolArgsFromConfigLocked(s1.config).Copy()
w := newWorker(s1.shutdownCtx, s1, poolArgs)
w.evalToken = token
err = w.UpdateEval(eval2)
if err != nil {
t.Fatalf("err: %v", err)
}
ws := memdb.NewWatchSet()
out, err := s1.fsm.State().EvalByID(ws, eval2.ID)
if err != nil {
t.Fatalf("err: %v", err)
}
if out.Status != structs.EvalStatusComplete {
t.Fatalf("bad: %v", out)
}
if out.SnapshotIndex != w.snapshotIndex {
t.Fatalf("bad: %v", out)
}
}
func TestWorker_CreateEval(t *testing.T) {
ci.Parallel(t)
s1, cleanupS1 := TestServer(t, func(c *Config) {
c.NumSchedulers = 0
c.EnabledSchedulers = []string{structs.JobTypeService}
})
defer cleanupS1()
testutil.WaitForLeader(t, s1.RPC)
// Register node
node := mock.Node()
testRegisterNode(t, s1, node)
// Create the register request
eval1 := mock.Eval()
s1.evalBroker.Enqueue(eval1)
evalOut, token, err := s1.evalBroker.Dequeue([]string{eval1.Type}, time.Second)
if err != nil {
t.Fatalf("err: %v", err)
}
if evalOut != eval1 {
t.Fatalf("Bad eval")
}
eval2 := mock.Eval()
eval2.PreviousEval = eval1.ID
// Attempt to create eval
poolArgs := getSchedulerWorkerPoolArgsFromConfigLocked(s1.config).Copy()
w := newWorker(s1.shutdownCtx, s1, poolArgs)
w.evalToken = token
err = w.CreateEval(eval2)
if err != nil {
t.Fatalf("err: %v", err)
}
ws := memdb.NewWatchSet()
out, err := s1.fsm.State().EvalByID(ws, eval2.ID)
if err != nil {
t.Fatalf("err: %v", err)
}
if out.PreviousEval != eval1.ID {
t.Fatalf("bad: %v", out)
}
if out.SnapshotIndex != w.snapshotIndex {
t.Fatalf("bad: %v", out)
}
}
func TestWorker_ReblockEval(t *testing.T) {
ci.Parallel(t)
s1, cleanupS1 := TestServer(t, func(c *Config) {
c.NumSchedulers = 0
c.EnabledSchedulers = []string{structs.JobTypeService}
})
defer cleanupS1()
testutil.WaitForLeader(t, s1.RPC)
// Create the blocked eval
eval1 := mock.Eval()
eval1.Status = structs.EvalStatusBlocked
eval1.QueuedAllocations = map[string]int{"cache": 100}
// Insert it into the state store
if err := s1.fsm.State().UpsertEvals(structs.MsgTypeTestSetup, 1000, []*structs.Evaluation{eval1}); err != nil {
t.Fatal(err)
}
// Create the job summary
js := mock.JobSummary(eval1.JobID)
tg := js.Summary["web"]
tg.Queued = 100
js.Summary["web"] = tg
if err := s1.fsm.State().UpsertJobSummary(1001, js); err != nil {
t.Fatal(err)
}
// Enqueue the eval and then dequeue
s1.evalBroker.Enqueue(eval1)
evalOut, token, err := s1.evalBroker.Dequeue([]string{eval1.Type}, time.Second)
if err != nil {
t.Fatalf("err: %v", err)
}
if evalOut != eval1 {
t.Fatalf("Bad eval")
}
eval2 := evalOut.Copy()
eval2.QueuedAllocations = map[string]int{"web": 50}
// Attempt to reblock eval
poolArgs := getSchedulerWorkerPoolArgsFromConfigLocked(s1.config).Copy()
w := newWorker(s1.shutdownCtx, s1, poolArgs)
w.evalToken = token
err = w.ReblockEval(eval2)
if err != nil {
t.Fatalf("err: %v", err)
}
// Ack the eval
w.sendAck(evalOut, token)
// Check that it is blocked
bStats := s1.blockedEvals.Stats()
if bStats.TotalBlocked+bStats.TotalEscaped != 1 {
t.Fatalf("ReblockEval didn't insert eval into the blocked eval tracker: %#v", bStats)
}
// Check that the eval was updated
ws := memdb.NewWatchSet()
eval, err := s1.fsm.State().EvalByID(ws, eval2.ID)
if err != nil {
t.Fatal(err)
}
if !reflect.DeepEqual(eval.QueuedAllocations, eval2.QueuedAllocations) {
t.Fatalf("expected: %#v, actual: %#v", eval2.QueuedAllocations, eval.QueuedAllocations)
}
// Check that the snapshot index was set properly by unblocking the eval and
// then dequeuing.
s1.blockedEvals.Unblock("foobar", 1000)
reblockedEval, _, err := s1.evalBroker.Dequeue([]string{eval1.Type}, 1*time.Second)
if err != nil {
t.Fatalf("err: %v", err)
}
if reblockedEval == nil {
t.Fatalf("Nil eval")
}
if reblockedEval.ID != eval1.ID {
t.Fatalf("Bad eval")
}
// Check that the SnapshotIndex is set
if reblockedEval.SnapshotIndex != w.snapshotIndex {
t.Fatalf("incorrect snapshot index; got %d; want %d",
reblockedEval.SnapshotIndex, w.snapshotIndex)
}
}
func TestWorker_Info(t *testing.T) {
ci.Parallel(t)
s1, cleanupS1 := TestServer(t, func(c *Config) {
c.NumSchedulers = 0
c.EnabledSchedulers = []string{structs.JobTypeService}
})
defer cleanupS1()
testutil.WaitForLeader(t, s1.RPC)
poolArgs := getSchedulerWorkerPoolArgsFromConfigLocked(s1.config).Copy()
// Create a worker
w := newWorker(s1.shutdownCtx, s1, poolArgs)
require.Equal(t, WorkerStarting, w.GetStatus())
workerInfo := w.Info()
require.Equal(t, WorkerStarting.String(), workerInfo.Status)
}
const (
longWait = 100 * time.Millisecond
tinyWait = 10 * time.Millisecond
)
func TestWorker_SetPause(t *testing.T) {
ci.Parallel(t)
logger := testlog.HCLogger(t)
srv := &Server{
logger: logger,
shutdownCtx: context.Background(),
}
args := SchedulerWorkerPoolArgs{
EnabledSchedulers: []string{structs.JobTypeCore, structs.JobTypeBatch, structs.JobTypeSystem},
}
w := newWorker(context.Background(), srv, args)
w._start(testWorkload)
require.Eventually(t, w.IsStarted, longWait, tinyWait, "should have started")
go func() {
time.Sleep(tinyWait)
w.Pause()
}()
require.Eventually(t, w.IsPaused, longWait, tinyWait, "should have paused")
go func() {
time.Sleep(tinyWait)
w.Pause()
}()
require.Eventually(t, w.IsPaused, longWait, tinyWait, "pausing a paused should be okay")
go func() {
time.Sleep(tinyWait)
w.Resume()
}()
require.Eventually(t, w.IsStarted, longWait, tinyWait, "should have restarted from pause")
go func() {
time.Sleep(tinyWait)
w.Stop()
}()
require.Eventually(t, w.IsStopped, longWait, tinyWait, "should have shutdown")
}
func TestWorker_SetPause_OutOfOrderEvents(t *testing.T) {
ci.Parallel(t)
logger := testlog.HCLogger(t)
srv := &Server{
logger: logger,
shutdownCtx: context.Background(),
}
args := SchedulerWorkerPoolArgs{
EnabledSchedulers: []string{structs.JobTypeCore, structs.JobTypeBatch, structs.JobTypeSystem},
}
w := newWorker(context.Background(), srv, args)
w._start(testWorkload)
require.Eventually(t, w.IsStarted, longWait, tinyWait, "should have started")
go func() {
time.Sleep(tinyWait)
w.Pause()
}()
require.Eventually(t, w.IsPaused, longWait, tinyWait, "should have paused")
go func() {
time.Sleep(tinyWait)
w.Stop()
}()
require.Eventually(t, w.IsStopped, longWait, tinyWait, "stop from pause should have shutdown")
go func() {
time.Sleep(tinyWait)
w.Pause()
}()
require.Eventually(t, w.IsStopped, longWait, tinyWait, "pausing a stopped should stay stopped")
}
// _start is a test helper function used to start a worker with an alternate workload
func (w *Worker) _start(inFunc func(w *Worker)) {
w.setStatus(WorkerStarting)
go inFunc(w)
}
// testWorkload is a very simple function that performs the same status updating behaviors that the
// real workload does.
func testWorkload(w *Worker) {
defer w.markStopped()
w.setStatuses(WorkerStarted, WorkloadRunning)
w.logger.Debug("testWorkload running")
for {
// ensure state variables are happy after resuming.
w.maybeWait()
if w.workerShuttingDown() {
w.logger.Debug("testWorkload stopped")
return
}
// do some fake work
time.Sleep(10 * time.Millisecond)
}
}