Mirroristas/etcd
2
0
Fork 0
mirror of https://github.com/etcd-io/etcd.git synced 2024-09-27 06:25:44 +00:00
Code Issues Packages Projects Releases Wiki Activity

etcdserver: ensure initial-advertise-peer-urls match initial-cluster

Browse Source 
This adds a check to setupCluster to ensure that the list of URLs
specified in `initial-advertise-peer-urls` matches those configured in
`initial-cluster` for this node. Also updates the documentation to
clarify this and address some changes in wording.
This commit is contained in:
Jonathan Boulle 2014-11-10 12:55:35 -08:00
parent 0772987128
commit 3f358b6d5d
5 changed files with 119 additions and 21 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

30
Documentation/0.5/clustering.md
View File

@ -10,7 +10,7 @@ This guide will walk you through configuring a three machine etcd cluster with t
## Static
As we know the cluster members, their addresses and the size of the cluster before starting we can use an offline bootstrap configuration. Each machine will get either the following command line or environment variables:
As we know the cluster members, their addresses and the size of the cluster before starting, we can use an offline bootstrap configuration by setting the `initial-cluster` flag. Each machine will get either the following command line or environment variables:
```
ETCD_INITIAL_CLUSTER="infra0=http://10.0.1.10:2380,infra1=http://10.0.1.11:2380,infra2=http://10.0.1.12:2380"
@ -22,6 +22,8 @@ ETCD_INITIAL_CLUSTER_STATE=new
-initial-cluster-state new
```
Note that the URLs specified in `initial-cluster` are the _advertised peer URLs_, i.e. they should match the value of `initial-advertise-peer-urls` on the respective nodes.
If you are spinning up multiple clusters (or creating and destroying a single cluster) with same configuration for testing purpose, it is highly recommended that you specify a unique `initial-cluster-token` for the different clusters. By doing this, etcd can generate unique cluster IDs and member IDs for the clusters even if they otherwise have the exact same configuration. This can protect you from cross-cluster-interaction, which might corrupt your clusters.
On each machine you would start etcd with these flags:
@ -45,11 +47,11 @@ $ etcd -name infra2 -initial-advertise-peer-urls https://10.0.1.12:2380 \
-initial-cluster-state new
```
The command line parameters starting with `-initial-cluster` will be ignored on subsequent runs of etcd. You are free to remove the environment variables or command line flags after the initial bootstrap process. If you need to make changes to the configuration later see our guide on [runtime configuration](runtime-configuration.md).
The command line parameters starting with `-initial-cluster` will be ignored on subsequent runs of etcd. You are free to remove the environment variables or command line flags after the initial bootstrap process. If you need to make changes to the configuration later (for example, adding or removing members to/from the cluster), see the [runtime configuration](runtime-configuration.md) guide.
### Error Cases
In the following case we have not included our new host in the list of enumerated nodes. If this is a new cluster, the node must be added to the list of initial cluster members.
In the following example, we have not included our new host in the list of enumerated nodes. If this is a new cluster, the node _must_ be added to the list of initial cluster members.
```
$ etcd -name infra1 -initial-advertise-peer-urls http://10.0.1.11:2380 \
@ -59,13 +61,13 @@ etcd: infra1 not listed in the initial cluster config
exit 1
```
In this case we are attempting to map a node (infra0) on a different address (127.0.0.1:2380) than its enumerated address in the cluster list (10.0.1.10:2380). If this node is to listen on multiple addresses, all addresses must be reflected in the "initial-cluster" configuration directive.
In this example, we are attempting to map a node (infra0) on a different address (127.0.0.1:2380) than its enumerated address in the cluster list (10.0.1.10:2380). If this node is to listen on multiple addresses, all addresses _must_ be reflected in the "initial-cluster" configuration directive.
```
$ etcd -name infra0 -initial-advertise-peer-urls http://127.0.0.1:2380 \
-initial-cluster infra0=http://10.0.1.10:2380,infra1=http://10.0.1.11:2380,infra2=http://10.0.1.12:2380 \
-initial-cluster-state=new
etcd: infra0 has different advertised URLs in the cluster and advertised peer URLs list
etcd: error setting up initial cluster: infra0 has different advertised URLs in the cluster and advertised peer URLs list
exit 1
```
@ -81,7 +83,7 @@ exit 1
## Discovery
In a number of cases you might not know the IPs of your cluster peers ahead of time. This is common when utilizing cloud providers or when your network uses DHCP. In these cases you can use an existing etcd cluster to bootstrap a new one. We call this process "discovery".
In a number of cases, you might not know the IPs of your cluster peers ahead of time. This is common when utilizing cloud providers or when your network uses DHCP. In these cases, rather than specifying a static configuration, you can use an existing etcd cluster to bootstrap a new one. We call this process "discovery".
### Lifetime of a Discovery URL
@ -99,7 +101,7 @@ Discovery uses an existing cluster to bootstrap itself. If you are using your ow
$ curl -X PUT https://myetcd.local/v2/keys/discovery/6c007a14875d53d9bf0ef5a6fc0257c817f0fb83/_config/size -d value=3
```
By setting the size key to the URL, you create a discovery URL with expected-cluster-size of 3.
By setting the size key to the URL, you create a discovery URL with an expected cluster size of 3.
If you bootstrap an etcd cluster using discovery service with more than the expected number of etcd members, the extra etcd processes will [fall back][fall-back] to being [proxies][proxy] by default.
@ -124,14 +126,14 @@ This will cause each member to register itself with the custom etcd discovery se
### Public discovery service
If you do not have access to an existing cluster you can use the public discovery service hosted at discovery.etcd.io. You can create a private discovery URL using the "new" endpoint like so:
If you do not have access to an existing cluster, you can use the public discovery service hosted at `discovery.etcd.io`. You can create a private discovery URL using the "new" endpoint like so:
```
$ curl https://discovery.etcd.io/new?size=3
https://discovery.etcd.io/3e86b59982e49066c5d813af1c2e2579cbf573de
```
This will create the cluster with an initial expected size of 3 members. If you do not specify a size a default of 3 will be used.
This will create the cluster with an initial expected size of 3 members. If you do not specify a size, a default of 3 will be used.
If you bootstrap an etcd cluster using discovery service with more than the expected number of etcd members, the extra etcd processes will [fall back][fall-back] to being [proxies][proxy] by default.
@ -169,6 +171,7 @@ You can use the environment variable `ETCD_DISCOVERY_PROXY` to cause etcd to use
#### Discovery Server Errors
```
$ etcd -name infra0 -initial-advertise-peer-urls http://10.0.1.10:2380 \
-discovery https://discovery.etcd.io/3e86b59982e49066c5d813af1c2e2579cbf573de
@ -178,10 +181,13 @@ exit 1
#### User Errors
This error will occur if the discovery cluster already has the configured number of members, and `discovery-fallback` is explicitly disabled
```
$ etcd -name infra0 -initial-advertise-peer-urls http://10.0.1.10:2380 \
-discovery https://discovery.etcd.io/3e86b59982e49066c5d813af1c2e2579cbf573de
etcd: error: the cluster using discovery https://discovery.etcd.io/3e86b59982e49066c5d813af1c2e2579cbf573de has already started with all 5 members
-discovery https://discovery.etcd.io/3e86b59982e49066c5d813af1c2e2579cbf573de \
-discovery-fallback exit
etcd: discovery: cluster is full
exit 1
```
@ -193,7 +199,7 @@ ignored on this machine.
```
$ etcd -name infra0 -initial-advertise-peer-urls http://10.0.1.10:2380 \
-discovery https://discovery.etcd.io/3e86b59982e49066c5d813af1c2e2579cbf573de
etcd: warn: ignoring discovery URL: etcd has already been initialized and has a valid log in /var/lib/etcd
etcdserver: warn: ignoring discovery: etcd has already been initialized and has a valid log in /var/lib/etcd
```
# 0.4 to 0.5+ Migration Guide

29
etcdmain/etcd.go
View File

@ -22,7 +22,10 @@ import (
"log"
"net"
"net/http"
"net/url"
"os"
"reflect"
"sort"
"strings"
"github.com/coreos/etcd/discovery"
@ -363,7 +366,7 @@ func startProxy() error {
return nil
}
// setupCluster sets up the cluster definition for bootstrap or discovery.
// setupCluster sets up an initial cluster definition for bootstrap or discovery.
func setupCluster() (*etcdserver.Cluster, error) {
set := make(map[string]bool)
fs.Visit(func(f *flag.Flag) {
@ -380,18 +383,40 @@ func setupCluster() (*etcdserver.Cluster, error) {
var cls *etcdserver.Cluster
switch {
case set["discovery"]:
// If using discovery, generate a temporary cluster based on
// self's advertised peer URLs
clusterStr := genClusterString(*name, apurls)
cls, err = etcdserver.NewClusterFromString(*durl, clusterStr)
case set["initial-cluster"]:
fallthrough
default:
// We're statically configured, and cluster has appropriately been set.
// Try to configure by indexing the static cluster by name.
cls, err = etcdserver.NewClusterFromString(*initialClusterToken, *initialCluster)
// Ensure our own advertised peer URLs match those specified in cluster
if err == nil && !clusterPeerURLsMatch(*name, cls, apurls) {
cls = nil
err = fmt.Errorf("%s has different advertised URLs in the cluster and advertised peer URLs list", *name)
}
}
return cls, err
}
// clusterPeerURLsMatch checks whether the peer URLs of the member by the given
// name in the given cluster match the provided set of URLs
func clusterPeerURLsMatch(name string, cls *etcdserver.Cluster, urls []url.URL) bool {
m := cls.MemberByName(name)
if m == nil {
// should never happen
log.Panicf("could not find %q in cluster!", name)
}
purls := make([]string, len(urls))
for i, u := range urls {
purls[i] = u.String()
}
sort.Strings(purls)
return reflect.DeepEqual(purls, m.PeerURLs)
}
func genClusterString(name string, urls types.URLs) string {
addrs := make([]string, 0)
for _, u := range urls {

67
etcdmain/etcd_test.go
View File

@ -17,11 +17,73 @@
package etcdmain
import (
"net/url"
"testing"
"github.com/coreos/etcd/etcdserver"
"github.com/coreos/etcd/pkg/types"
)
func mustClsFromString(t *testing.T, s string) *etcdserver.Cluster {
cls, err := etcdserver.NewClusterFromString("", s)
if err != nil {
t.Fatalf("error creating cluster from %q: %v", s, err)
}
return cls
}
func mustNewURLs(t *testing.T, urls []string) []url.URL {
u, err := types.NewURLs(urls)
if err != nil {
t.Fatalf("unexpected new urls error: %v", err)
}
return u
}
func TestClusterPeerURLsMatch(t *testing.T) {
tests := []struct {
name string
cls *etcdserver.Cluster
urls []url.URL
w bool
}{
{
name: "default",
cls: mustClsFromString(t, "default=http://localhost:12345"),
urls: mustNewURLs(t, []string{"http://localhost:12345"}),
w: true,
},
{
name: "default",
cls: mustClsFromString(t, "default=http://localhost:7001,other=http://192.168.0.1:7002,default=http://localhost:12345"),
urls: mustNewURLs(t, []string{"http://localhost:7001", "http://localhost:12345"}),
w: true,
},
{
name: "infra1",
cls: mustClsFromString(t, "infra1=http://localhost:7001"),
urls: mustNewURLs(t, []string{"http://localhost:12345"}),
w: false,
},
{
name: "infra1",
cls: mustClsFromString(t, "infra1=http://localhost:7001,infra2=http://localhost:12345"),
urls: mustNewURLs(t, []string{"http://localhost:12345"}),
w: false,
},
}
for i, tt := range tests {
if g := clusterPeerURLsMatch(tt.name, tt.cls, tt.urls); g != tt.w {
t.Errorf("#%d: clusterPeerURLsMatch=%t, want %t", i, g, tt.w)
}
}
}
func TestGenClusterString(t *testing.T) {
tests := []struct {
token string
@ -38,10 +100,7 @@ func TestGenClusterString(t *testing.T) {
},
}
for i, tt := range tests {
urls, err := types.NewURLs(tt.urls)
if err != nil {
t.Fatalf("unexpected new urls error: %v", err)
}
urls := mustNewURLs(t, tt.urls)
str := genClusterString(tt.token, urls)
if str != tt.wstr {
t.Errorf("#%d: cluster = %s, want %s", i, str, tt.wstr)

12
etcdserver/cluster.go
View File

@ -42,11 +42,18 @@ const (
)
type ClusterInfo interface {
// ID returns the cluster ID
ID() types.ID
// ClientURLs returns an aggregate set of all URLs on which this
// cluster is listening for client requests
ClientURLs() []string
// Members returns a slice of members sorted by their ID
Members() []*Member
// Member retrieves a particular member based on ID, or nil if the
// member does not exist in the cluster
Member(id types.ID) *Member
// IsIDRemoved checks whether the given ID has been removed from this
// cluster at some point in the past
IsIDRemoved(id types.ID) bool
}
@ -62,8 +69,9 @@ type Cluster struct {
sync.Mutex
}
// NewClusterFromString returns Cluster through given cluster token and parsing
// members from a sets of names to IPs discovery formatted like:
// NewClusterFromString returns a Cluster instantiated from the given cluster token
// and cluster string, by parsing members from a set of discovery-formatted
// names-to-IPs, like:
// mach0=http://1.1.1.1,mach0=http://2.2.2.2,mach1=http://3.3.3.3,mach2=http://4.4.4.4
func NewClusterFromString(token string, cluster string) (*Cluster, error) {
c := newCluster(token)

2
etcdserver/config.go
View File

@ -45,7 +45,7 @@ func (c *ServerConfig) VerifyBootstrapConfig() error {
m := c.Cluster.MemberByName(c.Name)
// Make sure the cluster at least contains the local server.
if m == nil {
return fmt.Errorf("couldn't find local name %s in the initial cluster configuration", c.Name)
return fmt.Errorf("couldn't find local name %q in the initial cluster configuration", c.Name)
}
if uint64(m.ID) == raft.None {
return fmt.Errorf("cannot use %x as member id", raft.None)