open-consul/sdk/testutil/retry/retry.go

273 lines
5.8 KiB
Go
Raw Normal View History

// Package retry provides support for repeating operations in tests.
//
// A sample retry operation looks like this:
//
// func TestX(t *testing.T) {
// retry.Run(t, func(r *retry.R) {
// if err := foo(); err != nil {
// r.Errorf("foo: %s", err)
// return
// }
// })
// }
//
// Run uses the DefaultFailer, which is a Timer with a Timeout of 7s,
// and a Wait of 25ms. To customize, use RunWith.
//
// WARNING: unlike *testing.T, *retry.R#Fatal and FailNow *do not*
// fail the test function entirely, only the current run the retry func
package retry
import (
"bytes"
"fmt"
"runtime"
"strings"
"time"
)
// Failer is an interface compatible with testing.T.
type Failer interface {
2020-08-14 20:54:44 +00:00
Helper()
// Log is called for the final test output
Log(args ...interface{})
// FailNow is called when the retrying is abandoned.
FailNow()
}
// R provides context for the retryer.
//
// Logs from Logf, (Error|Fatal)(f) are gathered in an internal buffer
// and printed only if the retryer fails. Printed logs are deduped and
// prefixed with source code line numbers
type R struct {
// fail is set by FailNow and (Fatal|Error)(f). It indicates the pass
// did not succeed, and should be retried
fail bool
// done is set by Stop. It indicates the entire run was a failure,
// and triggers t.FailNow()
done bool
output []string
}
func (r *R) Logf(format string, args ...interface{}) {
r.log(fmt.Sprintf(format, args...))
}
func (r *R) Helper() {}
// runFailed is a sentinel value to indicate that the func itself
// didn't panic, rather that `FailNow` was called.
type runFailed struct{}
// FailNow stops run execution. It is roughly equivalent to:
//
// r.Error("")
// return
//
// inside the function being run.
func (r *R) FailNow() {
r.fail = true
panic(runFailed{})
}
// Fatal is equivalent to r.Logf(args) followed by r.FailNow(), i.e. the run
// function should be exited. Retries on the next run are allowed. Fatal is
// equivalent to
//
// r.Error(args)
// return
//
// inside the function being run.
func (r *R) Fatal(args ...interface{}) {
r.log(fmt.Sprint(args...))
r.FailNow()
}
// Fatalf is like Fatal but allows a format string
func (r *R) Fatalf(format string, args ...interface{}) {
r.log(fmt.Sprintf(format, args...))
r.FailNow()
}
// Error indicates the current run encountered an error and should be retried.
// It *does not* stop execution of the rest of the run function.
func (r *R) Error(args ...interface{}) {
r.log(fmt.Sprint(args...))
r.fail = true
}
// Errorf is like Error but allows a format string
New ACLs (#4791) This PR is almost a complete rewrite of the ACL system within Consul. It brings the features more in line with other HashiCorp products. Obviously there is quite a bit left to do here but most of it is related docs, testing and finishing the last few commands in the CLI. I will update the PR description and check off the todos as I finish them over the next few days/week. Description At a high level this PR is mainly to split ACL tokens from Policies and to split the concepts of Authorization from Identities. A lot of this PR is mostly just to support CRUD operations on ACLTokens and ACLPolicies. These in and of themselves are not particularly interesting. The bigger conceptual changes are in how tokens get resolved, how backwards compatibility is handled and the separation of policy from identity which could lead the way to allowing for alternative identity providers. On the surface and with a new cluster the ACL system will look very similar to that of Nomads. Both have tokens and policies. Both have local tokens. The ACL management APIs for both are very similar. I even ripped off Nomad's ACL bootstrap resetting procedure. There are a few key differences though. Nomad requires token and policy replication where Consul only requires policy replication with token replication being opt-in. In Consul local tokens only work with token replication being enabled though. All policies in Nomad are globally applicable. In Consul all policies are stored and replicated globally but can be scoped to a subset of the datacenters. This allows for more granular access management. Unlike Nomad, Consul has legacy baggage in the form of the original ACL system. The ramifications of this are: A server running the new system must still support other clients using the legacy system. A client running the new system must be able to use the legacy RPCs when the servers in its datacenter are running the legacy system. The primary ACL DC's servers running in legacy mode needs to be a gate that keeps everything else in the entire multi-DC cluster running in legacy mode. So not only does this PR implement the new ACL system but has a legacy mode built in for when the cluster isn't ready for new ACLs. Also detecting that new ACLs can be used is automatic and requires no configuration on the part of administrators. This process is detailed more in the "Transitioning from Legacy to New ACL Mode" section below.
2018-10-19 16:04:07 +00:00
func (r *R) Errorf(format string, args ...interface{}) {
r.log(fmt.Sprintf(format, args...))
r.fail = true
}
// If err is non-nil, equivalent to r.Fatal(err.Error()) followed by
// r.FailNow(). Otherwise a no-op.
func (r *R) Check(err error) {
if err != nil {
r.log(err.Error())
r.FailNow()
}
}
func (r *R) log(s string) {
r.output = append(r.output, decorate(s))
}
// Stop retrying, and fail the test, logging the specified error.
// Does not stop execution, so return should be called after.
func (r *R) Stop(err error) {
r.log(err.Error())
r.done = true
}
func decorate(s string) string {
_, file, line, ok := runtime.Caller(3)
if ok {
n := strings.LastIndex(file, "/")
if n >= 0 {
file = file[n+1:]
}
} else {
file = "???"
line = 1
}
return fmt.Sprintf("%s:%d: %s", file, line, s)
}
func Run(t Failer, f func(r *R)) {
t.Helper()
2018-02-08 01:29:30 +00:00
run(DefaultFailer(), t, f)
}
func RunWith(r Retryer, t Failer, f func(r *R)) {
t.Helper()
run(r, t, f)
}
func dedup(a []string) string {
if len(a) == 0 {
return ""
}
seen := map[string]struct{}{}
var b bytes.Buffer
for _, s := range a {
if _, ok := seen[s]; ok {
continue
}
seen[s] = struct{}{}
b.WriteString(s)
b.WriteRune('\n')
}
return b.String()
}
func run(r Retryer, t Failer, f func(r *R)) {
t.Helper()
rr := &R{}
fail := func() {
t.Helper()
out := dedup(rr.output)
if out != "" {
t.Log(out)
}
t.FailNow()
}
for r.Continue() {
// run f(rr), but if recover yields a runFailed value, we know
// FailNow was called.
func() {
defer func() {
if p := recover(); p != nil && p != (runFailed{}) {
panic(p)
}
}()
f(rr)
}()
switch {
case rr.done:
fail()
return
case !rr.fail:
return
}
rr.fail = false
}
fail()
}
// DefaultFailer provides default retry.Run() behavior for unit tests, namely
// 7s timeout with a wait of 25ms
2018-02-08 01:29:30 +00:00
func DefaultFailer() *Timer {
return &Timer{Timeout: 7 * time.Second, Wait: 25 * time.Millisecond}
}
// TwoSeconds repeats an operation for two seconds and waits 25ms in between.
func TwoSeconds() *Timer {
return &Timer{Timeout: 2 * time.Second, Wait: 25 * time.Millisecond}
}
// ThreeTimes repeats an operation three times and waits 25ms in between.
func ThreeTimes() *Counter {
return &Counter{Count: 3, Wait: 25 * time.Millisecond}
}
// Retryer provides an interface for repeating operations
// until they succeed or an exit condition is met.
type Retryer interface {
// Continue returns true if the operation should be repeated, otherwise it
// returns false to indicate retrying should stop.
Continue() bool
}
// Counter repeats an operation a given number of
// times and waits between subsequent operations.
type Counter struct {
Count int
Wait time.Duration
count int
}
func (r *Counter) Continue() bool {
if r.count == r.Count {
return false
}
if r.count > 0 {
time.Sleep(r.Wait)
}
r.count++
return true
}
// Timer repeats an operation for a given amount
// of time and waits between subsequent operations.
type Timer struct {
Timeout time.Duration
Wait time.Duration
// stop is the timeout deadline.
// TODO: Next()?
// Set on the first invocation of Next().
stop time.Time
}
func (r *Timer) Continue() bool {
if r.stop.IsZero() {
r.stop = time.Now().Add(r.Timeout)
return true
}
if time.Now().After(r.stop) {
return false
}
time.Sleep(r.Wait)
return true
}