Merge pull request #15280 from meteor/glasser/unique-id-timestamp-again

Fix resource.UniqueId to be properly ordered over multiple runs
This commit is contained in:
James Bardin 2017-06-15 11:13:34 -04:00 committed by GitHub
commit 956ab165bd
2 changed files with 46 additions and 20 deletions

View File

@ -1,21 +1,17 @@
package resource
import (
"crypto/rand"
"fmt"
"math/big"
"strings"
"sync"
"time"
)
const UniqueIdPrefix = `terraform-`
// idCounter is a randomly seeded monotonic counter for generating ordered
// unique ids. It uses a big.Int so we can easily increment a long numeric
// string. The max possible hex value here with 12 random bytes is
// "01000000000000000000000000", so there's no chance of rollover during
// operation.
// idCounter is a monotonic counter for generating ordered unique ids.
var idMutex sync.Mutex
var idCounter = big.NewInt(0).SetBytes(randomBytes(12))
var idCounter uint32
// Helper for a resource to generate a unique identifier w/ default prefix
func UniqueId() string {
@ -25,15 +21,20 @@ func UniqueId() string {
// Helper for a resource to generate a unique identifier w/ given prefix
//
// After the prefix, the ID consists of an incrementing 26 digit value (to match
// previous timestamp output).
// previous timestamp output). After the prefix, the ID consists of a timestamp
// and an incrementing 8 hex digit value The timestamp means that multiple IDs
// created with the same prefix will sort in the order of their creation, even
// across multiple terraform executions, as long as the clock is not turned back
// between calls, and as long as any given terraform execution generates fewer
// than 4 billion IDs.
func PrefixedUniqueId(prefix string) string {
// Be precise to 4 digits of fractional seconds, but remove the dot before the
// fractional seconds.
timestamp := strings.Replace(
time.Now().UTC().Format("20060102150405.0000"), ".", "", 1)
idMutex.Lock()
defer idMutex.Unlock()
return fmt.Sprintf("%s%026x", prefix, idCounter.Add(idCounter, big.NewInt(1)))
}
func randomBytes(n int) []byte {
b := make([]byte, n)
rand.Read(b)
return b
idCounter++
return fmt.Sprintf("%s%s%08x", prefix, timestamp, idCounter)
}

View File

@ -4,11 +4,19 @@ import (
"regexp"
"strings"
"testing"
"time"
)
var allDigits = regexp.MustCompile(`^\d+$`)
var allHex = regexp.MustCompile(`^[a-f0-9]+$`)
func TestUniqueId(t *testing.T) {
split := func(rest string) (timestamp, increment string) {
return rest[:18], rest[18:]
}
const prefix = "terraform-"
iterations := 10000
ids := make(map[string]struct{})
var id, lastId string
@ -19,18 +27,24 @@ func TestUniqueId(t *testing.T) {
t.Fatalf("Got duplicated id! %s", id)
}
if !strings.HasPrefix(id, "terraform-") {
if !strings.HasPrefix(id, prefix) {
t.Fatalf("Unique ID didn't have terraform- prefix! %s", id)
}
rest := strings.TrimPrefix(id, "terraform-")
rest := strings.TrimPrefix(id, prefix)
if len(rest) != 26 {
t.Fatalf("Post-prefix part has wrong length! %s", rest)
}
if !allHex.MatchString(rest) {
t.Fatalf("Random part not all hex! %s", rest)
timestamp, increment := split(rest)
if !allDigits.MatchString(timestamp) {
t.Fatalf("Timestamp not all digits! %s", timestamp)
}
if !allHex.MatchString(increment) {
t.Fatalf("Increment part not all hex! %s", increment)
}
if lastId != "" && lastId >= id {
@ -40,4 +54,15 @@ func TestUniqueId(t *testing.T) {
ids[id] = struct{}{}
lastId = id
}
id1 := UniqueId()
time.Sleep(time.Millisecond)
id2 := UniqueId()
timestamp1, _ := split(strings.TrimPrefix(id1, prefix))
timestamp2, _ := split(strings.TrimPrefix(id2, prefix))
if timestamp1 == timestamp2 {
t.Fatalf("Timestamp part should update at least once a millisecond %s %s",
id1, id2)
}
}