open-consul/agent/consul/helper_test.go

558 lines
16 KiB
Go
Raw Normal View History

2017-05-05 10:29:49 +00:00
package consul
import (
"errors"
2017-05-05 10:29:49 +00:00
"fmt"
"net"
"net/rpc"
2017-05-05 10:29:49 +00:00
"testing"
"github.com/hashicorp/consul/agent/structs"
"github.com/hashicorp/consul/api"
"github.com/hashicorp/consul/sdk/testutil/retry"
"github.com/hashicorp/consul/types"
msgpackrpc "github.com/hashicorp/net-rpc-msgpackrpc"
"github.com/hashicorp/raft"
"github.com/hashicorp/serf/serf"
"github.com/stretchr/testify/require"
2017-05-05 10:29:49 +00:00
)
func waitForLeader(servers ...*Server) error {
if len(servers) == 0 {
return errors.New("no servers")
}
dc := servers[0].config.Datacenter
for _, s := range servers {
if s.config.Datacenter != dc {
return fmt.Errorf("servers are in different datacenters %s and %s", s.config.Datacenter, dc)
}
}
for _, s := range servers {
if s.IsLeader() {
return nil
}
}
return errors.New("no leader")
}
2017-05-05 10:29:49 +00:00
// wantPeers determines whether the server has the given
// number of voting raft peers.
2017-05-05 10:29:49 +00:00
func wantPeers(s *Server, peers int) error {
n, err := s.numPeers()
if err != nil {
return err
}
if got, want := n, peers; got != want {
return fmt.Errorf("got %d peers want %d", got, want)
}
return nil
}
// wantRaft determines if the servers have all of each other in their
// Raft configurations,
func wantRaft(servers []*Server) error {
// Make sure all the servers are represented in the Raft config,
// and that there are no extras.
verifyRaft := func(c raft.Configuration) error {
want := make(map[raft.ServerID]bool)
for _, s := range servers {
want[s.config.RaftConfig.LocalID] = true
}
for _, s := range c.Servers {
if !want[s.ID] {
return fmt.Errorf("don't want %q", s.ID)
}
delete(want, s.ID)
}
if len(want) > 0 {
return fmt.Errorf("didn't find %v", want)
}
return nil
}
for _, s := range servers {
future := s.raft.GetConfiguration()
if err := future.Error(); err != nil {
return err
}
if err := verifyRaft(future.Configuration()); err != nil {
return err
}
}
return nil
}
2017-05-05 10:29:49 +00:00
// joinAddrLAN returns the address other servers can
// use to join the cluster on the LAN interface.
func joinAddrLAN(s *Server) string {
if s == nil {
panic("no server")
}
port := s.config.SerfLANConfig.MemberlistConfig.BindPort
return fmt.Sprintf("127.0.0.1:%d", port)
}
// joinAddrWAN returns the address other servers can
// use to join the cluster on the WAN interface.
func joinAddrWAN(s *Server) string {
if s == nil {
panic("no server")
}
wan federation via mesh gateways (#6884) This is like a Möbius strip of code due to the fact that low-level components (serf/memberlist) are connected to high-level components (the catalog and mesh-gateways) in a twisty maze of references which make it hard to dive into. With that in mind here's a high level summary of what you'll find in the patch: There are several distinct chunks of code that are affected: * new flags and config options for the server * retry join WAN is slightly different * retry join code is shared to discover primary mesh gateways from secondary datacenters * because retry join logic runs in the *agent* and the results of that operation for primary mesh gateways are needed in the *server* there are some methods like `RefreshPrimaryGatewayFallbackAddresses` that must occur at multiple layers of abstraction just to pass the data down to the right layer. * new cache type `FederationStateListMeshGatewaysName` for use in `proxycfg/xds` layers * the function signature for RPC dialing picked up a new required field (the node name of the destination) * several new RPCs for manipulating a FederationState object: `FederationState:{Apply,Get,List,ListMeshGateways}` * 3 read-only internal APIs for debugging use to invoke those RPCs from curl * raft and fsm changes to persist these FederationStates * replication for FederationStates as they are canonically stored in the Primary and replicated to the Secondaries. * a special derivative of anti-entropy that runs in secondaries to snapshot their local mesh gateway `CheckServiceNodes` and sync them into their upstream FederationState in the primary (this works in conjunction with the replication to distribute addresses for all mesh gateways in all DCs to all other DCs) * a "gateway locator" convenience object to make use of this data to choose the addresses of gateways to use for any given RPC or gossip operation to a remote DC. This gets data from the "retry join" logic in the agent and also directly calls into the FSM. * RPC (`:8300`) on the server sniffs the first byte of a new connection to determine if it's actually doing native TLS. If so it checks the ALPN header for protocol determination (just like how the existing system uses the type-byte marker). * 2 new kinds of protocols are exclusively decoded via this native TLS mechanism: one for ferrying "packet" operations (udp-like) from the gossip layer and one for "stream" operations (tcp-like). The packet operations re-use sockets (using length-prefixing) to cut down on TLS re-negotiation overhead. * the server instances specially wrap the `memberlist.NetTransport` when running with gateway federation enabled (in a `wanfed.Transport`). The general gist is that if it tries to dial a node in the SAME datacenter (deduced by looking at the suffix of the node name) there is no change. If dialing a DIFFERENT datacenter it is wrapped up in a TLS+ALPN blob and sent through some mesh gateways to eventually end up in a server's :8300 port. * a new flag when launching a mesh gateway via `consul connect envoy` to indicate that the servers are to be exposed. This sets a special service meta when registering the gateway into the catalog. * `proxycfg/xds` notice this metadata blob to activate additional watches for the FederationState objects as well as the location of all of the consul servers in that datacenter. * `xds:` if the extra metadata is in place additional clusters are defined in a DC to bulk sink all traffic to another DC's gateways. For the current datacenter we listen on a wildcard name (`server.<dc>.consul`) that load balances all servers as well as one mini-cluster per node (`<node>.server.<dc>.consul`) * the `consul tls cert create` command got a new flag (`-node`) to help create an additional SAN in certs that can be used with this flavor of federation.
2020-03-09 20:59:02 +00:00
name := s.config.NodeName
dc := s.config.Datacenter
2017-05-05 10:29:49 +00:00
port := s.config.SerfWANConfig.MemberlistConfig.BindPort
wan federation via mesh gateways (#6884) This is like a Möbius strip of code due to the fact that low-level components (serf/memberlist) are connected to high-level components (the catalog and mesh-gateways) in a twisty maze of references which make it hard to dive into. With that in mind here's a high level summary of what you'll find in the patch: There are several distinct chunks of code that are affected: * new flags and config options for the server * retry join WAN is slightly different * retry join code is shared to discover primary mesh gateways from secondary datacenters * because retry join logic runs in the *agent* and the results of that operation for primary mesh gateways are needed in the *server* there are some methods like `RefreshPrimaryGatewayFallbackAddresses` that must occur at multiple layers of abstraction just to pass the data down to the right layer. * new cache type `FederationStateListMeshGatewaysName` for use in `proxycfg/xds` layers * the function signature for RPC dialing picked up a new required field (the node name of the destination) * several new RPCs for manipulating a FederationState object: `FederationState:{Apply,Get,List,ListMeshGateways}` * 3 read-only internal APIs for debugging use to invoke those RPCs from curl * raft and fsm changes to persist these FederationStates * replication for FederationStates as they are canonically stored in the Primary and replicated to the Secondaries. * a special derivative of anti-entropy that runs in secondaries to snapshot their local mesh gateway `CheckServiceNodes` and sync them into their upstream FederationState in the primary (this works in conjunction with the replication to distribute addresses for all mesh gateways in all DCs to all other DCs) * a "gateway locator" convenience object to make use of this data to choose the addresses of gateways to use for any given RPC or gossip operation to a remote DC. This gets data from the "retry join" logic in the agent and also directly calls into the FSM. * RPC (`:8300`) on the server sniffs the first byte of a new connection to determine if it's actually doing native TLS. If so it checks the ALPN header for protocol determination (just like how the existing system uses the type-byte marker). * 2 new kinds of protocols are exclusively decoded via this native TLS mechanism: one for ferrying "packet" operations (udp-like) from the gossip layer and one for "stream" operations (tcp-like). The packet operations re-use sockets (using length-prefixing) to cut down on TLS re-negotiation overhead. * the server instances specially wrap the `memberlist.NetTransport` when running with gateway federation enabled (in a `wanfed.Transport`). The general gist is that if it tries to dial a node in the SAME datacenter (deduced by looking at the suffix of the node name) there is no change. If dialing a DIFFERENT datacenter it is wrapped up in a TLS+ALPN blob and sent through some mesh gateways to eventually end up in a server's :8300 port. * a new flag when launching a mesh gateway via `consul connect envoy` to indicate that the servers are to be exposed. This sets a special service meta when registering the gateway into the catalog. * `proxycfg/xds` notice this metadata blob to activate additional watches for the FederationState objects as well as the location of all of the consul servers in that datacenter. * `xds:` if the extra metadata is in place additional clusters are defined in a DC to bulk sink all traffic to another DC's gateways. For the current datacenter we listen on a wildcard name (`server.<dc>.consul`) that load balances all servers as well as one mini-cluster per node (`<node>.server.<dc>.consul`) * the `consul tls cert create` command got a new flag (`-node`) to help create an additional SAN in certs that can be used with this flavor of federation.
2020-03-09 20:59:02 +00:00
return fmt.Sprintf("%s.%s/127.0.0.1:%d", name, dc, port)
2017-05-05 10:29:49 +00:00
}
type clientOrServer interface {
JoinLAN(addrs []string) (int, error)
LANMembers() []serf.Member
2017-05-05 10:29:49 +00:00
}
// joinLAN is a convenience function for
//
// member.JoinLAN("127.0.0.1:"+leader.config.SerfLANConfig.MemberlistConfig.BindPort)
func joinLAN(t *testing.T, member clientOrServer, leader *Server) {
if member == nil || leader == nil {
panic("no server")
}
var memberAddr string
switch x := member.(type) {
case *Server:
memberAddr = joinAddrLAN(x)
case *Client:
memberAddr = fmt.Sprintf("127.0.0.1:%d", x.config.SerfLANConfig.MemberlistConfig.BindPort)
}
leaderAddr := joinAddrLAN(leader)
if _, err := member.JoinLAN([]string{leaderAddr}); err != nil {
2017-05-05 10:29:49 +00:00
t.Fatal(err)
}
retry.Run(t, func(r *retry.R) {
if !seeEachOther(leader.LANMembers(), member.LANMembers(), leaderAddr, memberAddr) {
r.Fatalf("leader and member cannot see each other on LAN")
}
})
if !seeEachOther(leader.LANMembers(), member.LANMembers(), leaderAddr, memberAddr) {
t.Fatalf("leader and member cannot see each other on LAN")
}
2017-05-05 10:29:49 +00:00
}
// joinWAN is a convenience function for
//
// member.JoinWAN("127.0.0.1:"+leader.config.SerfWANConfig.MemberlistConfig.BindPort)
func joinWAN(t *testing.T, member, leader *Server) {
if member == nil || leader == nil {
panic("no server")
}
leaderAddr, memberAddr := joinAddrWAN(leader), joinAddrWAN(member)
if _, err := member.JoinWAN([]string{leaderAddr}); err != nil {
2017-05-05 10:29:49 +00:00
t.Fatal(err)
}
retry.Run(t, func(r *retry.R) {
if !seeEachOther(leader.WANMembers(), member.WANMembers(), leaderAddr, memberAddr) {
r.Fatalf("leader and member cannot see each other on WAN")
}
})
if !seeEachOther(leader.WANMembers(), member.WANMembers(), leaderAddr, memberAddr) {
t.Fatalf("leader and member cannot see each other on WAN")
}
}
func waitForNewACLs(t *testing.T, server *Server) {
t.Helper()
retry.Run(t, func(r *retry.R) {
require.False(r, server.UseLegacyACLs(), "Server cannot use new ACLs")
})
require.False(t, server.UseLegacyACLs(), "Server cannot use new ACLs")
}
func waitForNewACLReplication(t *testing.T, server *Server, expectedReplicationType structs.ACLReplicationType, minPolicyIndex, minTokenIndex, minRoleIndex uint64) {
t.Helper()
retry.Run(t, func(r *retry.R) {
status := server.getACLReplicationStatus()
require.Equal(r, expectedReplicationType, status.ReplicationType, "Server not running new replicator yet")
require.True(r, status.Running, "Server not running new replicator yet")
require.True(r, status.ReplicatedIndex >= minPolicyIndex, "Server hasn't replicated enough policies")
require.True(r, status.ReplicatedTokenIndex >= minTokenIndex, "Server hasn't replicated enough tokens")
require.True(r, status.ReplicatedRoleIndex >= minRoleIndex, "Server hasn't replicated enough roles")
})
}
func seeEachOther(a, b []serf.Member, addra, addrb string) bool {
return serfMembersContains(a, addrb) && serfMembersContains(b, addra)
}
func serfMembersContains(members []serf.Member, addr string) bool {
// There are tests that manipulate the advertise address, so we just
// compare port numbers here, since that uniquely identifies a member
// as we use the loopback interface for everything.
_, want, err := net.SplitHostPort(addr)
if err != nil {
panic(err)
}
for _, m := range members {
if got := fmt.Sprintf("%d", m.Port); got == want {
return true
}
}
return false
2017-05-05 10:29:49 +00:00
}
func registerTestCatalogEntries(t *testing.T, codec rpc.ClientCodec) {
t.Helper()
// prep the cluster with some data we can use in our filters
registrations := map[string]*structs.RegisterRequest{
"Node foo": &structs.RegisterRequest{
Datacenter: "dc1",
Node: "foo",
ID: types.NodeID("e0155642-135d-4739-9853-a1ee6c9f945b"),
Address: "127.0.0.2",
TaggedAddresses: map[string]string{
"lan": "127.0.0.2",
"wan": "198.18.0.2",
},
NodeMeta: map[string]string{
"env": "production",
"os": "linux",
},
Checks: structs.HealthChecks{
&structs.HealthCheck{
Node: "foo",
CheckID: "foo:alive",
Name: "foo-liveness",
Status: api.HealthPassing,
Notes: "foo is alive and well",
},
&structs.HealthCheck{
Node: "foo",
CheckID: "foo:ssh",
Name: "foo-remote-ssh",
Status: api.HealthPassing,
Notes: "foo has ssh access",
},
},
},
"Service redis v1 on foo": &structs.RegisterRequest{
Datacenter: "dc1",
Node: "foo",
SkipNodeUpdate: true,
Service: &structs.NodeService{
Kind: structs.ServiceKindTypical,
ID: "redisV1",
Service: "redis",
Tags: []string{"v1"},
Meta: map[string]string{"version": "1"},
Port: 1234,
Address: "198.18.1.2",
},
Checks: structs.HealthChecks{
&structs.HealthCheck{
Node: "foo",
CheckID: "foo:redisV1",
Name: "redis-liveness",
Status: api.HealthPassing,
Notes: "redis v1 is alive and well",
ServiceID: "redisV1",
ServiceName: "redis",
},
},
},
"Service redis v2 on foo": &structs.RegisterRequest{
Datacenter: "dc1",
Node: "foo",
SkipNodeUpdate: true,
Service: &structs.NodeService{
Kind: structs.ServiceKindTypical,
ID: "redisV2",
Service: "redis",
Tags: []string{"v2"},
Meta: map[string]string{"version": "2"},
Port: 1235,
Address: "198.18.1.2",
},
Checks: structs.HealthChecks{
&structs.HealthCheck{
Node: "foo",
CheckID: "foo:redisV2",
Name: "redis-v2-liveness",
Status: api.HealthPassing,
Notes: "redis v2 is alive and well",
ServiceID: "redisV2",
ServiceName: "redis",
},
},
},
"Node bar": &structs.RegisterRequest{
Datacenter: "dc1",
Node: "bar",
ID: types.NodeID("c6e7a976-8f4f-44b5-bdd3-631be7e8ecac"),
Address: "127.0.0.3",
TaggedAddresses: map[string]string{
"lan": "127.0.0.3",
"wan": "198.18.0.3",
},
NodeMeta: map[string]string{
"env": "production",
"os": "windows",
},
Checks: structs.HealthChecks{
&structs.HealthCheck{
Node: "bar",
CheckID: "bar:alive",
Name: "bar-liveness",
Status: api.HealthPassing,
Notes: "bar is alive and well",
},
},
},
"Service redis v1 on bar": &structs.RegisterRequest{
Datacenter: "dc1",
Node: "bar",
SkipNodeUpdate: true,
Service: &structs.NodeService{
Kind: structs.ServiceKindTypical,
ID: "redisV1",
Service: "redis",
Tags: []string{"v1"},
Meta: map[string]string{"version": "1"},
Port: 1234,
Address: "198.18.1.3",
},
Checks: structs.HealthChecks{
&structs.HealthCheck{
Node: "bar",
CheckID: "bar:redisV1",
Name: "redis-liveness",
Status: api.HealthPassing,
Notes: "redis v1 is alive and well",
ServiceID: "redisV1",
ServiceName: "redis",
},
},
},
"Service web v1 on bar": &structs.RegisterRequest{
Datacenter: "dc1",
Node: "bar",
SkipNodeUpdate: true,
Service: &structs.NodeService{
Kind: structs.ServiceKindTypical,
ID: "webV1",
Service: "web",
Tags: []string{"v1", "connect"},
Meta: map[string]string{"version": "1", "connect": "enabled"},
Port: 443,
Address: "198.18.1.4",
Connect: structs.ServiceConnect{Native: true},
},
Checks: structs.HealthChecks{
&structs.HealthCheck{
Node: "bar",
CheckID: "bar:web:v1",
Name: "web-v1-liveness",
Status: api.HealthPassing,
Notes: "web connect v1 is alive and well",
ServiceID: "webV1",
ServiceName: "web",
},
},
},
"Node baz": &structs.RegisterRequest{
Datacenter: "dc1",
Node: "baz",
ID: types.NodeID("12f96b27-a7b0-47bd-add7-044a2bfc7bfb"),
Address: "127.0.0.4",
TaggedAddresses: map[string]string{
"lan": "127.0.0.4",
},
NodeMeta: map[string]string{
"env": "qa",
"os": "linux",
},
Checks: structs.HealthChecks{
&structs.HealthCheck{
Node: "baz",
CheckID: "baz:alive",
Name: "baz-liveness",
Status: api.HealthPassing,
Notes: "baz is alive and well",
},
&structs.HealthCheck{
Node: "baz",
CheckID: "baz:ssh",
Name: "baz-remote-ssh",
Status: api.HealthPassing,
Notes: "baz has ssh access",
},
},
},
"Service web v1 on baz": &structs.RegisterRequest{
Datacenter: "dc1",
Node: "baz",
SkipNodeUpdate: true,
Service: &structs.NodeService{
Kind: structs.ServiceKindTypical,
ID: "webV1",
Service: "web",
Tags: []string{"v1", "connect"},
Meta: map[string]string{"version": "1", "connect": "enabled"},
Port: 443,
Address: "198.18.1.4",
Connect: structs.ServiceConnect{Native: true},
},
Checks: structs.HealthChecks{
&structs.HealthCheck{
Node: "baz",
CheckID: "baz:web:v1",
Name: "web-v1-liveness",
Status: api.HealthPassing,
Notes: "web connect v1 is alive and well",
ServiceID: "webV1",
ServiceName: "web",
},
},
},
"Service web v2 on baz": &structs.RegisterRequest{
Datacenter: "dc1",
Node: "baz",
SkipNodeUpdate: true,
Service: &structs.NodeService{
Kind: structs.ServiceKindTypical,
ID: "webV2",
Service: "web",
Tags: []string{"v2", "connect"},
Meta: map[string]string{"version": "2", "connect": "enabled"},
Port: 8443,
Address: "198.18.1.4",
Connect: structs.ServiceConnect{Native: true},
},
Checks: structs.HealthChecks{
&structs.HealthCheck{
Node: "baz",
CheckID: "baz:web:v2",
Name: "web-v2-liveness",
Status: api.HealthPassing,
Notes: "web connect v2 is alive and well",
ServiceID: "webV2",
ServiceName: "web",
},
},
},
"Service critical on baz": &structs.RegisterRequest{
Datacenter: "dc1",
Node: "baz",
SkipNodeUpdate: true,
Service: &structs.NodeService{
Kind: structs.ServiceKindTypical,
ID: "criticalV2",
Service: "critical",
Tags: []string{"v2"},
Meta: map[string]string{"version": "2"},
Port: 8080,
Address: "198.18.1.4",
},
Checks: structs.HealthChecks{
&structs.HealthCheck{
Node: "baz",
CheckID: "baz:critical:v2",
Name: "critical-v2-liveness",
Status: api.HealthCritical,
Notes: "critical v2 is in the critical state",
ServiceID: "criticalV2",
ServiceName: "critical",
},
},
},
"Service warning on baz": &structs.RegisterRequest{
Datacenter: "dc1",
Node: "baz",
SkipNodeUpdate: true,
Service: &structs.NodeService{
Kind: structs.ServiceKindTypical,
ID: "warningV2",
Service: "warning",
Tags: []string{"v2"},
Meta: map[string]string{"version": "2"},
Port: 8081,
Address: "198.18.1.4",
},
Checks: structs.HealthChecks{
&structs.HealthCheck{
Node: "baz",
CheckID: "baz:warning:v2",
Name: "warning-v2-liveness",
Status: api.HealthWarning,
Notes: "warning v2 is in the warning state",
ServiceID: "warningV2",
ServiceName: "warning",
},
},
},
}
registerTestCatalogEntriesMap(t, codec, registrations)
}
func registerTestCatalogProxyEntries(t *testing.T, codec rpc.ClientCodec) {
t.Helper()
registrations := map[string]*structs.RegisterRequest{
"Service tg-gw": &structs.RegisterRequest{
Datacenter: "dc1",
Node: "terminating-gateway",
ID: types.NodeID("3a9d7530-20d4-443a-98d3-c10fe78f09f4"),
Address: "10.1.2.2",
Service: &structs.NodeService{
Kind: structs.ServiceKindTerminatingGateway,
ID: "tg-gw-01",
Service: "tg-gw",
Port: 8443,
Address: "198.18.1.3",
},
},
"Service mg-gw": &structs.RegisterRequest{
Datacenter: "dc1",
Node: "gateway",
ID: types.NodeID("72e18a4c-85ec-4520-978f-2fc0378b06aa"),
Address: "10.1.2.3",
Service: &structs.NodeService{
Kind: structs.ServiceKindMeshGateway,
ID: "mg-gw-01",
Service: "mg-gw",
Port: 8443,
Address: "198.18.1.4",
},
},
"Service web-proxy": &structs.RegisterRequest{
Datacenter: "dc1",
Node: "proxy",
ID: types.NodeID("2d31602c-3291-4f94-842d-446bc2f945ce"),
Address: "10.1.2.4",
Service: &structs.NodeService{
Kind: structs.ServiceKindConnectProxy,
ID: "web-proxy",
Service: "web-proxy",
Port: 8443,
Address: "198.18.1.5",
Proxy: structs.ConnectProxyConfig{
DestinationServiceName: "web",
},
},
},
}
registerTestCatalogEntriesMap(t, codec, registrations)
}
func registerTestCatalogEntriesMap(t *testing.T, codec rpc.ClientCodec, registrations map[string]*structs.RegisterRequest) {
t.Helper()
for name, reg := range registrations {
err := msgpackrpc.CallWithCodec(codec, "Catalog.Register", reg, nil)
require.NoError(t, err, "Failed catalog registration %q: %v", name, err)
}
}