open-consul/agent/consul/catalog_endpoint_test.go

2745 lines
75 KiB
Go
Raw Normal View History

2013-12-11 22:57:40 +00:00
package consul
import (
2013-12-12 00:33:19 +00:00
"fmt"
2013-12-19 20:03:57 +00:00
"net/rpc"
2013-12-11 22:57:40 +00:00
"os"
"strings"
2013-12-11 22:57:40 +00:00
"testing"
"time"
"github.com/hashicorp/consul/acl"
"github.com/hashicorp/consul/agent/structs"
"github.com/hashicorp/consul/api"
"github.com/hashicorp/consul/lib"
2020-05-27 16:47:32 +00:00
"github.com/hashicorp/consul/lib/stringslice"
"github.com/hashicorp/consul/sdk/testutil/retry"
"github.com/hashicorp/consul/testrpc"
2017-01-18 22:26:42 +00:00
"github.com/hashicorp/consul/types"
msgpackrpc "github.com/hashicorp/net-rpc-msgpackrpc"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
2013-12-11 22:57:40 +00:00
)
func TestCatalog_Register(t *testing.T) {
t.Parallel()
2013-12-11 22:57:40 +00:00
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
2013-12-11 22:57:40 +00:00
2013-12-19 20:03:57 +00:00
arg := structs.RegisterRequest{
Datacenter: "dc1",
Node: "foo",
Address: "127.0.0.1",
Service: &structs.NodeService{
Service: "db",
2014-04-03 19:03:10 +00:00
Tags: []string{"master"},
Port: 8000,
},
Check: &structs.HealthCheck{
CheckID: types.CheckID("db-check"),
ServiceID: "db",
},
2013-12-11 22:57:40 +00:00
}
var out struct{}
err := msgpackrpc.CallWithCodec(codec, "Catalog.Register", &arg, &out)
if err != nil {
2013-12-11 22:57:40 +00:00
t.Fatalf("err: %v", err)
}
2017-01-18 22:26:42 +00:00
}
func TestCatalog_RegisterService_InvalidAddress(t *testing.T) {
t.Parallel()
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
for _, addr := range []string{"0.0.0.0", "::", "[::]"} {
t.Run("addr "+addr, func(t *testing.T) {
arg := structs.RegisterRequest{
Datacenter: "dc1",
Node: "foo",
Address: "127.0.0.1",
Service: &structs.NodeService{
Service: "db",
Address: addr,
Port: 8000,
},
}
var out struct{}
err := msgpackrpc.CallWithCodec(codec, "Catalog.Register", &arg, &out)
if err == nil || err.Error() != "Invalid service address" {
t.Fatalf("got error %v want 'Invalid service address'", err)
}
})
}
}
2017-11-01 21:25:46 +00:00
func TestCatalog_RegisterService_SkipNodeUpdate(t *testing.T) {
t.Parallel()
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
// Register a node
arg := structs.RegisterRequest{
Datacenter: "dc1",
Node: "foo",
Address: "127.0.0.1",
}
var out struct{}
err := msgpackrpc.CallWithCodec(codec, "Catalog.Register", &arg, &out)
if err != nil {
t.Fatal(err)
}
// Update it with a blank address, should fail.
arg.Address = ""
arg.Service = &structs.NodeService{
Service: "db",
Port: 8000,
}
err = msgpackrpc.CallWithCodec(codec, "Catalog.Register", &arg, &out)
if err == nil || err.Error() != "Must provide address if SkipNodeUpdate is not set" {
t.Fatalf("got error %v want 'Must provide address...'", err)
}
// Set SkipNodeUpdate, should succeed
arg.SkipNodeUpdate = true
err = msgpackrpc.CallWithCodec(codec, "Catalog.Register", &arg, &out)
if err != nil {
t.Fatal(err)
}
}
2017-01-18 22:26:42 +00:00
func TestCatalog_Register_NodeID(t *testing.T) {
t.Parallel()
2017-01-18 22:26:42 +00:00
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
arg := structs.RegisterRequest{
Datacenter: "dc1",
ID: "nope",
Node: "foo",
Address: "127.0.0.1",
Service: &structs.NodeService{
Service: "db",
Tags: []string{"master"},
Port: 8000,
},
Check: &structs.HealthCheck{
CheckID: types.CheckID("db-check"),
2017-01-18 22:26:42 +00:00
ServiceID: "db",
},
}
var out struct{}
2013-12-11 22:57:40 +00:00
2017-01-18 22:26:42 +00:00
err := msgpackrpc.CallWithCodec(codec, "Catalog.Register", &arg, &out)
if err == nil || !strings.Contains(err.Error(), "Bad node ID") {
2013-12-11 22:57:40 +00:00
t.Fatalf("err: %v", err)
2017-01-18 22:26:42 +00:00
}
arg.ID = types.NodeID("adf4238a-882b-9ddc-4a9d-5b6758e4159e")
if err := msgpackrpc.CallWithCodec(codec, "Catalog.Register", &arg, &out); err != nil {
t.Fatalf("err: %v", err)
}
2013-12-11 22:57:40 +00:00
}
2013-12-11 23:34:10 +00:00
func TestCatalog_Register_ACLDeny(t *testing.T) {
t.Parallel()
dir1, s1 := testServerWithConfig(t, func(c *Config) {
c.ACLDatacenter = "dc1"
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
c.ACLsEnabled = true
c.ACLMasterToken = "root"
c.ACLDefaultPolicy = "deny"
})
defer os.RemoveAll(dir1)
defer s1.Shutdown()
testrpc.WaitForTestAgent(t, s1.RPC, "dc1", testrpc.WithToken("root"))
codec := rpcClient(t, s1)
defer codec.Close()
// Create the ACL.
arg := structs.ACLRequest{
Datacenter: "dc1",
Op: structs.ACLSet,
ACL: structs.ACL{
Name: "User token",
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
Type: structs.ACLTokenTypeClient,
Rules: `
service "foo" {
policy = "write"
}
node "foo" {
policy = "write"
}
`,
},
WriteRequest: structs.WriteRequest{Token: "root"},
}
var id string
if err := msgpackrpc.CallWithCodec(codec, "ACL.Apply", &arg, &id); err != nil {
t.Fatalf("err: %v", err)
}
argR := structs.RegisterRequest{
Datacenter: "dc1",
Node: "foo",
Address: "127.0.0.1",
Service: &structs.NodeService{
Service: "db",
Tags: []string{"master"},
Port: 8000,
},
WriteRequest: structs.WriteRequest{Token: id},
}
var outR struct{}
// This should fail since we are writing to the "db" service, which isn't
// allowed.
err := msgpackrpc.CallWithCodec(codec, "Catalog.Register", &argR, &outR)
if !acl.IsErrPermissionDenied(err) {
t.Fatalf("err: %v", err)
}
// The "foo" service should work, though.
argR.Service.Service = "foo"
err = msgpackrpc.CallWithCodec(codec, "Catalog.Register", &argR, &outR)
if err != nil {
t.Fatalf("err: %v", err)
}
// Try the former special case for the "consul" service.
argR.Service.Service = "consul"
err = msgpackrpc.CallWithCodec(codec, "Catalog.Register", &argR, &outR)
if !acl.IsErrPermissionDenied(err) {
t.Fatalf("err: %v", err)
}
// Register a db service using the root token.
argR.Service.Service = "db"
argR.Service.ID = "my-id"
argR.Token = "root"
err = msgpackrpc.CallWithCodec(codec, "Catalog.Register", &argR, &outR)
if err != nil {
t.Fatalf("err: %v", err)
}
// Prove that we are properly looking up the node services and passing
// that to the ACL helper. We can vet the helper independently in its
// own unit test after this. This is trying to register over the db
// service we created above, which is a check that depends on looking
// at the existing registration data with that service ID. This is a new
// check for version 8.
argR.Service.Service = "foo"
argR.Service.ID = "my-id"
argR.Token = id
err = msgpackrpc.CallWithCodec(codec, "Catalog.Register", &argR, &outR)
if !acl.IsErrPermissionDenied(err) {
t.Fatalf("err: %v", err)
}
}
func TestCatalog_Register_ForwardLeader(t *testing.T) {
t.Parallel()
2013-12-12 00:42:19 +00:00
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec1 := rpcClient(t, s1)
defer codec1.Close()
2013-12-12 00:42:19 +00:00
dir2, s2 := testServer(t)
defer os.RemoveAll(dir2)
defer s2.Shutdown()
codec2 := rpcClient(t, s2)
defer codec2.Close()
2013-12-12 00:42:19 +00:00
// Try to join
2017-05-05 10:29:49 +00:00
joinLAN(t, s2, s1)
2013-12-12 00:42:19 +00:00
testrpc.WaitForLeader(t, s1.RPC, "dc1")
testrpc.WaitForLeader(t, s2.RPC, "dc1")
2013-12-12 00:42:19 +00:00
// Use the follower as the client
var codec rpc.ClientCodec
2013-12-12 00:42:19 +00:00
if !s1.IsLeader() {
codec = codec1
2013-12-12 00:42:19 +00:00
} else {
codec = codec2
2013-12-12 00:42:19 +00:00
}
2013-12-19 20:03:57 +00:00
arg := structs.RegisterRequest{
Datacenter: "dc1",
Node: "foo",
Address: "127.0.0.1",
Service: &structs.NodeService{
Service: "db",
2014-04-03 19:03:10 +00:00
Tags: []string{"master"},
Port: 8000,
},
2013-12-12 00:42:19 +00:00
}
var out struct{}
if err := msgpackrpc.CallWithCodec(codec, "Catalog.Register", &arg, &out); err != nil {
2013-12-12 00:42:19 +00:00
t.Fatalf("err: %v", err)
}
}
func TestCatalog_Register_ForwardDC(t *testing.T) {
t.Parallel()
2013-12-12 00:33:19 +00:00
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
2013-12-12 00:33:19 +00:00
dir2, s2 := testServerDC(t, "dc2")
defer os.RemoveAll(dir2)
defer s2.Shutdown()
// Try to join
2017-05-05 10:29:49 +00:00
joinWAN(t, s2, s1)
2013-12-12 00:33:19 +00:00
testrpc.WaitForLeader(t, s1.RPC, "dc2")
2013-12-12 00:33:19 +00:00
2013-12-19 20:03:57 +00:00
arg := structs.RegisterRequest{
2015-09-15 12:22:08 +00:00
Datacenter: "dc2", // Should forward through s1
Node: "foo",
Address: "127.0.0.1",
Service: &structs.NodeService{
Service: "db",
2014-04-03 19:03:10 +00:00
Tags: []string{"master"},
Port: 8000,
},
2013-12-12 00:33:19 +00:00
}
var out struct{}
if err := msgpackrpc.CallWithCodec(codec, "Catalog.Register", &arg, &out); err != nil {
2013-12-12 00:33:19 +00:00
t.Fatalf("err: %v", err)
}
}
func TestCatalog_Register_ConnectProxy(t *testing.T) {
t.Parallel()
assert := assert.New(t)
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
args := structs.TestRegisterRequestProxy(t)
// Register
var out struct{}
assert.Nil(msgpackrpc.CallWithCodec(codec, "Catalog.Register", &args, &out))
// List
req := structs.ServiceSpecificRequest{
Datacenter: "dc1",
ServiceName: args.Service.Service,
}
var resp structs.IndexedServiceNodes
assert.Nil(msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &req, &resp))
assert.Len(resp.ServiceNodes, 1)
v := resp.ServiceNodes[0]
assert.Equal(structs.ServiceKindConnectProxy, v.ServiceKind)
Add Proxy Upstreams to Service Definition (#4639) * Refactor Service Definition ProxyDestination. This includes: - Refactoring all internal structs used - Updated tests for both deprecated and new input for: - Agent Services endpoint response - Agent Service endpoint response - Agent Register endpoint - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Register - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Services endpoint response - Catalog Node endpoint response - Catalog Service endpoint response - Updated API tests for all of the above too (both deprecated and new forms of register) TODO: - config package changes for on-disk service definitions - proxy config endpoint - built-in proxy support for new fields * Agent proxy config endpoint updated with upstreams * Config file changes for upstreams. * Add upstream opaque config and update all tests to ensure it works everywhere. * Built in proxy working with new Upstreams config * Command fixes and deprecations * Fix key translation, upstream type defaults and a spate of other subtele bugs found with ned to end test scripts... TODO: tests still failing on one case that needs a fix. I think it's key translation for upstreams nested in Managed proxy struct. * Fix translated keys in API registration. ≈ * Fixes from docs - omit some empty undocumented fields in API - Bring back ServiceProxyDestination in Catalog responses to not break backwards compat - this was removed assuming it was only used internally. * Documentation updates for Upstreams in service definition * Fixes for tests broken by many refactors. * Enable travis on f-connect branch in this branch too. * Add consistent Deprecation comments to ProxyDestination uses * Update version number on deprecation notices, and correct upstream datacenter field with explanation in docs
2018-09-12 16:07:47 +00:00
assert.Equal(args.Service.Proxy.DestinationServiceName, v.ServiceProxy.DestinationServiceName)
}
// Test an invalid ConnectProxy. We don't need to exhaustively test because
// this is all tested in structs on the Validate method.
func TestCatalog_Register_ConnectProxy_invalid(t *testing.T) {
t.Parallel()
assert := assert.New(t)
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
args := structs.TestRegisterRequestProxy(t)
Add Proxy Upstreams to Service Definition (#4639) * Refactor Service Definition ProxyDestination. This includes: - Refactoring all internal structs used - Updated tests for both deprecated and new input for: - Agent Services endpoint response - Agent Service endpoint response - Agent Register endpoint - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Register - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Services endpoint response - Catalog Node endpoint response - Catalog Service endpoint response - Updated API tests for all of the above too (both deprecated and new forms of register) TODO: - config package changes for on-disk service definitions - proxy config endpoint - built-in proxy support for new fields * Agent proxy config endpoint updated with upstreams * Config file changes for upstreams. * Add upstream opaque config and update all tests to ensure it works everywhere. * Built in proxy working with new Upstreams config * Command fixes and deprecations * Fix key translation, upstream type defaults and a spate of other subtele bugs found with ned to end test scripts... TODO: tests still failing on one case that needs a fix. I think it's key translation for upstreams nested in Managed proxy struct. * Fix translated keys in API registration. ≈ * Fixes from docs - omit some empty undocumented fields in API - Bring back ServiceProxyDestination in Catalog responses to not break backwards compat - this was removed assuming it was only used internally. * Documentation updates for Upstreams in service definition * Fixes for tests broken by many refactors. * Enable travis on f-connect branch in this branch too. * Add consistent Deprecation comments to ProxyDestination uses * Update version number on deprecation notices, and correct upstream datacenter field with explanation in docs
2018-09-12 16:07:47 +00:00
args.Service.Proxy.DestinationServiceName = ""
// Register
var out struct{}
err := msgpackrpc.CallWithCodec(codec, "Catalog.Register", &args, &out)
assert.NotNil(err)
Add Proxy Upstreams to Service Definition (#4639) * Refactor Service Definition ProxyDestination. This includes: - Refactoring all internal structs used - Updated tests for both deprecated and new input for: - Agent Services endpoint response - Agent Service endpoint response - Agent Register endpoint - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Register - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Services endpoint response - Catalog Node endpoint response - Catalog Service endpoint response - Updated API tests for all of the above too (both deprecated and new forms of register) TODO: - config package changes for on-disk service definitions - proxy config endpoint - built-in proxy support for new fields * Agent proxy config endpoint updated with upstreams * Config file changes for upstreams. * Add upstream opaque config and update all tests to ensure it works everywhere. * Built in proxy working with new Upstreams config * Command fixes and deprecations * Fix key translation, upstream type defaults and a spate of other subtele bugs found with ned to end test scripts... TODO: tests still failing on one case that needs a fix. I think it's key translation for upstreams nested in Managed proxy struct. * Fix translated keys in API registration. ≈ * Fixes from docs - omit some empty undocumented fields in API - Bring back ServiceProxyDestination in Catalog responses to not break backwards compat - this was removed assuming it was only used internally. * Documentation updates for Upstreams in service definition * Fixes for tests broken by many refactors. * Enable travis on f-connect branch in this branch too. * Add consistent Deprecation comments to ProxyDestination uses * Update version number on deprecation notices, and correct upstream datacenter field with explanation in docs
2018-09-12 16:07:47 +00:00
assert.Contains(err.Error(), "DestinationServiceName")
}
// Test that write is required for the proxy destination to register a proxy.
func TestCatalog_Register_ConnectProxy_ACLDestinationServiceName(t *testing.T) {
t.Parallel()
assert := assert.New(t)
dir1, s1 := testServerWithConfig(t, func(c *Config) {
c.ACLDatacenter = "dc1"
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
c.ACLsEnabled = true
c.ACLMasterToken = "root"
c.ACLDefaultPolicy = "deny"
})
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
testrpc.WaitForLeader(t, s1.RPC, "dc1")
// Create the ACL.
arg := structs.ACLRequest{
Datacenter: "dc1",
Op: structs.ACLSet,
ACL: structs.ACL{
Name: "User token",
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
Type: structs.ACLTokenTypeClient,
Rules: `
service "foo" {
policy = "write"
}
node "foo" {
policy = "write"
}
`,
},
WriteRequest: structs.WriteRequest{Token: "root"},
}
var token string
assert.Nil(msgpackrpc.CallWithCodec(codec, "ACL.Apply", &arg, &token))
// Register should fail because we don't have permission on the destination
args := structs.TestRegisterRequestProxy(t)
args.Service.Service = "foo"
Add Proxy Upstreams to Service Definition (#4639) * Refactor Service Definition ProxyDestination. This includes: - Refactoring all internal structs used - Updated tests for both deprecated and new input for: - Agent Services endpoint response - Agent Service endpoint response - Agent Register endpoint - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Register - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Services endpoint response - Catalog Node endpoint response - Catalog Service endpoint response - Updated API tests for all of the above too (both deprecated and new forms of register) TODO: - config package changes for on-disk service definitions - proxy config endpoint - built-in proxy support for new fields * Agent proxy config endpoint updated with upstreams * Config file changes for upstreams. * Add upstream opaque config and update all tests to ensure it works everywhere. * Built in proxy working with new Upstreams config * Command fixes and deprecations * Fix key translation, upstream type defaults and a spate of other subtele bugs found with ned to end test scripts... TODO: tests still failing on one case that needs a fix. I think it's key translation for upstreams nested in Managed proxy struct. * Fix translated keys in API registration. ≈ * Fixes from docs - omit some empty undocumented fields in API - Bring back ServiceProxyDestination in Catalog responses to not break backwards compat - this was removed assuming it was only used internally. * Documentation updates for Upstreams in service definition * Fixes for tests broken by many refactors. * Enable travis on f-connect branch in this branch too. * Add consistent Deprecation comments to ProxyDestination uses * Update version number on deprecation notices, and correct upstream datacenter field with explanation in docs
2018-09-12 16:07:47 +00:00
args.Service.Proxy.DestinationServiceName = "bar"
args.WriteRequest.Token = token
var out struct{}
err := msgpackrpc.CallWithCodec(codec, "Catalog.Register", &args, &out)
assert.True(acl.IsErrPermissionDenied(err))
// Register should fail with the right destination but wrong name
args = structs.TestRegisterRequestProxy(t)
args.Service.Service = "bar"
Add Proxy Upstreams to Service Definition (#4639) * Refactor Service Definition ProxyDestination. This includes: - Refactoring all internal structs used - Updated tests for both deprecated and new input for: - Agent Services endpoint response - Agent Service endpoint response - Agent Register endpoint - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Register - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Services endpoint response - Catalog Node endpoint response - Catalog Service endpoint response - Updated API tests for all of the above too (both deprecated and new forms of register) TODO: - config package changes for on-disk service definitions - proxy config endpoint - built-in proxy support for new fields * Agent proxy config endpoint updated with upstreams * Config file changes for upstreams. * Add upstream opaque config and update all tests to ensure it works everywhere. * Built in proxy working with new Upstreams config * Command fixes and deprecations * Fix key translation, upstream type defaults and a spate of other subtele bugs found with ned to end test scripts... TODO: tests still failing on one case that needs a fix. I think it's key translation for upstreams nested in Managed proxy struct. * Fix translated keys in API registration. ≈ * Fixes from docs - omit some empty undocumented fields in API - Bring back ServiceProxyDestination in Catalog responses to not break backwards compat - this was removed assuming it was only used internally. * Documentation updates for Upstreams in service definition * Fixes for tests broken by many refactors. * Enable travis on f-connect branch in this branch too. * Add consistent Deprecation comments to ProxyDestination uses * Update version number on deprecation notices, and correct upstream datacenter field with explanation in docs
2018-09-12 16:07:47 +00:00
args.Service.Proxy.DestinationServiceName = "foo"
args.WriteRequest.Token = token
err = msgpackrpc.CallWithCodec(codec, "Catalog.Register", &args, &out)
assert.True(acl.IsErrPermissionDenied(err))
// Register should work with the right destination
args = structs.TestRegisterRequestProxy(t)
args.Service.Service = "foo"
Add Proxy Upstreams to Service Definition (#4639) * Refactor Service Definition ProxyDestination. This includes: - Refactoring all internal structs used - Updated tests for both deprecated and new input for: - Agent Services endpoint response - Agent Service endpoint response - Agent Register endpoint - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Register - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Services endpoint response - Catalog Node endpoint response - Catalog Service endpoint response - Updated API tests for all of the above too (both deprecated and new forms of register) TODO: - config package changes for on-disk service definitions - proxy config endpoint - built-in proxy support for new fields * Agent proxy config endpoint updated with upstreams * Config file changes for upstreams. * Add upstream opaque config and update all tests to ensure it works everywhere. * Built in proxy working with new Upstreams config * Command fixes and deprecations * Fix key translation, upstream type defaults and a spate of other subtele bugs found with ned to end test scripts... TODO: tests still failing on one case that needs a fix. I think it's key translation for upstreams nested in Managed proxy struct. * Fix translated keys in API registration. ≈ * Fixes from docs - omit some empty undocumented fields in API - Bring back ServiceProxyDestination in Catalog responses to not break backwards compat - this was removed assuming it was only used internally. * Documentation updates for Upstreams in service definition * Fixes for tests broken by many refactors. * Enable travis on f-connect branch in this branch too. * Add consistent Deprecation comments to ProxyDestination uses * Update version number on deprecation notices, and correct upstream datacenter field with explanation in docs
2018-09-12 16:07:47 +00:00
args.Service.Proxy.DestinationServiceName = "foo"
args.WriteRequest.Token = token
assert.Nil(msgpackrpc.CallWithCodec(codec, "Catalog.Register", &args, &out))
}
func TestCatalog_Register_ConnectNative(t *testing.T) {
t.Parallel()
assert := assert.New(t)
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
args := structs.TestRegisterRequest(t)
args.Service.Connect.Native = true
// Register
var out struct{}
assert.Nil(msgpackrpc.CallWithCodec(codec, "Catalog.Register", &args, &out))
// List
req := structs.ServiceSpecificRequest{
Datacenter: "dc1",
ServiceName: args.Service.Service,
}
var resp structs.IndexedServiceNodes
assert.Nil(msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &req, &resp))
assert.Len(resp.ServiceNodes, 1)
v := resp.ServiceNodes[0]
assert.Equal(structs.ServiceKindTypical, v.ServiceKind)
assert.True(v.ServiceConnect.Native)
}
func TestCatalog_Deregister(t *testing.T) {
t.Parallel()
2013-12-11 23:34:10 +00:00
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
2013-12-11 23:34:10 +00:00
2013-12-19 20:03:57 +00:00
arg := structs.DeregisterRequest{
2013-12-11 23:34:10 +00:00
Datacenter: "dc1",
Node: "foo",
}
var out struct{}
err := msgpackrpc.CallWithCodec(codec, "Catalog.Deregister", &arg, &out)
if err != nil {
2013-12-11 23:34:10 +00:00
t.Fatalf("err: %v", err)
}
testrpc.WaitForLeader(t, s1.RPC, "dc1")
2013-12-11 23:34:10 +00:00
if err := msgpackrpc.CallWithCodec(codec, "Catalog.Deregister", &arg, &out); err != nil {
2013-12-11 23:34:10 +00:00
t.Fatalf("err: %v", err)
}
}
2013-12-12 18:35:50 +00:00
func TestCatalog_Deregister_ACLDeny(t *testing.T) {
t.Parallel()
dir1, s1 := testServerWithConfig(t, func(c *Config) {
c.ACLDatacenter = "dc1"
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
c.ACLsEnabled = true
c.ACLMasterToken = "root"
c.ACLDefaultPolicy = "deny"
})
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
testrpc.WaitForLeader(t, s1.RPC, "dc1")
// Create the ACL.
arg := structs.ACLRequest{
Datacenter: "dc1",
Op: structs.ACLSet,
ACL: structs.ACL{
Name: "User token",
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
Type: structs.ACLTokenTypeClient,
Rules: `
node "node" {
policy = "write"
}
service "service" {
policy = "write"
}
`,
},
WriteRequest: structs.WriteRequest{Token: "root"},
}
var id string
if err := msgpackrpc.CallWithCodec(codec, "ACL.Apply", &arg, &id); err != nil {
t.Fatalf("err: %v", err)
}
// Register a node, node check, service, and service check.
argR := structs.RegisterRequest{
Datacenter: "dc1",
Node: "node",
Address: "127.0.0.1",
Service: &structs.NodeService{
Service: "service",
Port: 8000,
},
Checks: structs.HealthChecks{
&structs.HealthCheck{
Node: "node",
CheckID: "node-check",
},
&structs.HealthCheck{
Node: "node",
CheckID: "service-check",
ServiceID: "service",
},
},
WriteRequest: structs.WriteRequest{Token: id},
}
var outR struct{}
if err := msgpackrpc.CallWithCodec(codec, "Catalog.Register", &argR, &outR); err != nil {
t.Fatalf("err: %v", err)
}
// We should be not be able to deregister everything without a token.
var err error
var out struct{}
err = msgpackrpc.CallWithCodec(codec, "Catalog.Deregister",
&structs.DeregisterRequest{
Datacenter: "dc1",
Node: "node",
CheckID: "service-check"}, &out)
if !acl.IsErrPermissionDenied(err) {
t.Fatalf("err: %v", err)
}
err = msgpackrpc.CallWithCodec(codec, "Catalog.Deregister",
&structs.DeregisterRequest{
Datacenter: "dc1",
Node: "node",
CheckID: "node-check"}, &out)
if !acl.IsErrPermissionDenied(err) {
t.Fatalf("err: %v", err)
}
err = msgpackrpc.CallWithCodec(codec, "Catalog.Deregister",
&structs.DeregisterRequest{
Datacenter: "dc1",
Node: "node",
ServiceID: "service"}, &out)
if !acl.IsErrPermissionDenied(err) {
t.Fatalf("err: %v", err)
}
err = msgpackrpc.CallWithCodec(codec, "Catalog.Deregister",
&structs.DeregisterRequest{
Datacenter: "dc1",
Node: "node"}, &out)
if !acl.IsErrPermissionDenied(err) {
t.Fatalf("err: %v", err)
}
// Second pass these should all go through with the token set.
err = msgpackrpc.CallWithCodec(codec, "Catalog.Deregister",
&structs.DeregisterRequest{
Datacenter: "dc1",
Node: "node",
CheckID: "service-check",
WriteRequest: structs.WriteRequest{
Token: id,
}}, &out)
if err != nil {
t.Fatalf("err: %v", err)
}
err = msgpackrpc.CallWithCodec(codec, "Catalog.Deregister",
&structs.DeregisterRequest{
Datacenter: "dc1",
Node: "node",
CheckID: "node-check",
WriteRequest: structs.WriteRequest{
Token: id,
}}, &out)
if err != nil {
t.Fatalf("err: %v", err)
}
err = msgpackrpc.CallWithCodec(codec, "Catalog.Deregister",
&structs.DeregisterRequest{
Datacenter: "dc1",
Node: "node",
ServiceID: "service",
WriteRequest: structs.WriteRequest{
Token: id,
}}, &out)
if err != nil {
t.Fatalf("err: %v", err)
}
err = msgpackrpc.CallWithCodec(codec, "Catalog.Deregister",
&structs.DeregisterRequest{
Datacenter: "dc1",
Node: "node",
WriteRequest: structs.WriteRequest{
Token: id,
}}, &out)
if err != nil {
t.Fatalf("err: %v", err)
}
// Try a few error cases.
err = msgpackrpc.CallWithCodec(codec, "Catalog.Deregister",
&structs.DeregisterRequest{
Datacenter: "dc1",
Node: "nope",
ServiceID: "nope",
WriteRequest: structs.WriteRequest{
Token: id,
}}, &out)
if err == nil || !strings.Contains(err.Error(), "Unknown service") {
t.Fatalf("err: %v", err)
}
err = msgpackrpc.CallWithCodec(codec, "Catalog.Deregister",
&structs.DeregisterRequest{
Datacenter: "dc1",
Node: "nope",
CheckID: "nope",
WriteRequest: structs.WriteRequest{
Token: id,
}}, &out)
if err == nil || !strings.Contains(err.Error(), "Unknown check") {
t.Fatalf("err: %v", err)
}
}
func TestCatalog_ListDatacenters(t *testing.T) {
t.Parallel()
2013-12-12 18:35:50 +00:00
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
2013-12-12 18:35:50 +00:00
dir2, s2 := testServerDC(t, "dc2")
defer os.RemoveAll(dir2)
defer s2.Shutdown()
// Try to join
2017-05-05 10:29:49 +00:00
joinWAN(t, s2, s1)
testrpc.WaitForLeader(t, s1.RPC, "dc1")
2013-12-12 18:35:50 +00:00
var out []string
if err := msgpackrpc.CallWithCodec(codec, "Catalog.ListDatacenters", struct{}{}, &out); err != nil {
2013-12-12 18:35:50 +00:00
t.Fatalf("err: %v", err)
}
// The DCs should come out sorted by default.
2013-12-12 18:35:50 +00:00
if len(out) != 2 {
t.Fatalf("bad: %v", out)
}
if out[0] != "dc1" {
t.Fatalf("bad: %v", out)
}
if out[1] != "dc2" {
t.Fatalf("bad: %v", out)
}
}
2013-12-12 18:48:36 +00:00
func TestCatalog_ListDatacenters_DistanceSort(t *testing.T) {
t.Parallel()
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
2015-10-15 23:07:16 +00:00
codec := rpcClient(t, s1)
defer codec.Close()
dir2, s2 := testServerDC(t, "dc2")
defer os.RemoveAll(dir2)
defer s2.Shutdown()
dir3, s3 := testServerDC(t, "acdc")
defer os.RemoveAll(dir3)
defer s3.Shutdown()
// Try to join
2017-05-05 10:29:49 +00:00
joinWAN(t, s2, s1)
joinWAN(t, s3, s1)
testrpc.WaitForLeader(t, s1.RPC, "dc1")
var out []string
2015-10-15 23:07:16 +00:00
if err := msgpackrpc.CallWithCodec(codec, "Catalog.ListDatacenters", struct{}{}, &out); err != nil {
t.Fatalf("err: %v", err)
}
// It's super hard to force the Serfs into a known configuration of
// coordinates, so the best we can do is make sure that the sorting
// function is getting called (it's tested extensively in rtt_test.go).
// Since this is relative to dc1, it will be listed first (proving we
// went into the sort fn).
if len(out) != 3 {
t.Fatalf("bad: %v", out)
}
if out[0] != "dc1" {
t.Fatalf("bad: %v", out)
}
}
func TestCatalog_ListNodes(t *testing.T) {
t.Parallel()
2013-12-12 18:48:36 +00:00
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
2013-12-12 18:48:36 +00:00
args := structs.DCSpecificRequest{
Datacenter: "dc1",
}
var out structs.IndexedNodes
err := msgpackrpc.CallWithCodec(codec, "Catalog.ListNodes", &args, &out)
if err != nil {
2013-12-12 18:48:36 +00:00
t.Fatalf("err: %v", err)
}
testrpc.WaitForLeader(t, s1.RPC, "dc1")
2013-12-12 18:48:36 +00:00
// Just add a node
if err := s1.fsm.State().EnsureNode(1, &structs.Node{Node: "foo", Address: "127.0.0.1"}); err != nil {
t.Fatalf("err: %v", err)
}
retry.Run(t, func(r *retry.R) {
msgpackrpc.CallWithCodec(codec, "Catalog.ListNodes", &args, &out)
if got, want := len(out.Nodes), 2; got != want {
r.Fatalf("got %d nodes want %d", got, want)
}
})
2014-01-10 01:22:01 +00:00
// Server node is auto added from Serf
2014-07-23 08:34:03 +00:00
if out.Nodes[1].Node != s1.config.NodeName {
2013-12-12 18:48:36 +00:00
t.Fatalf("bad: %v", out)
}
2014-07-23 08:34:03 +00:00
if out.Nodes[0].Node != "foo" {
2013-12-12 18:48:36 +00:00
t.Fatalf("bad: %v", out)
}
2014-07-23 08:34:03 +00:00
if out.Nodes[0].Address != "127.0.0.1" {
2013-12-12 18:48:36 +00:00
t.Fatalf("bad: %v", out)
}
}
2013-12-12 19:07:14 +00:00
func TestCatalog_ListNodes_NodeMetaFilter(t *testing.T) {
t.Parallel()
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
testrpc.WaitForLeader(t, s1.RPC, "dc1")
// Add a new node with the right meta k/v pair
node := &structs.Node{Node: "foo", Address: "127.0.0.1", Meta: map[string]string{"somekey": "somevalue"}}
if err := s1.fsm.State().EnsureNode(1, node); err != nil {
t.Fatalf("err: %v", err)
}
// Filter by a specific meta k/v pair
args := structs.DCSpecificRequest{
Datacenter: "dc1",
NodeMetaFilters: map[string]string{
"somekey": "somevalue",
},
}
var out structs.IndexedNodes
retry.Run(t, func(r *retry.R) {
msgpackrpc.CallWithCodec(codec, "Catalog.ListNodes", &args, &out)
if got, want := len(out.Nodes), 1; got != want {
r.Fatalf("got %d nodes want %d", got, want)
}
})
// Verify that only the correct node was returned
if out.Nodes[0].Node != "foo" {
t.Fatalf("bad: %v", out)
}
if out.Nodes[0].Address != "127.0.0.1" {
t.Fatalf("bad: %v", out)
}
if v, ok := out.Nodes[0].Meta["somekey"]; !ok || v != "somevalue" {
t.Fatalf("bad: %v", out)
}
// Now filter on a nonexistent meta k/v pair
args = structs.DCSpecificRequest{
Datacenter: "dc1",
NodeMetaFilters: map[string]string{
"somekey": "invalid",
},
}
out = structs.IndexedNodes{}
err := msgpackrpc.CallWithCodec(codec, "Catalog.ListNodes", &args, &out)
if err != nil {
t.Fatalf("err: %v", err)
}
// Should get an empty list of nodes back
retry.Run(t, func(r *retry.R) {
msgpackrpc.CallWithCodec(codec, "Catalog.ListNodes", &args, &out)
if len(out.Nodes) != 0 {
r.Fatal(nil)
}
})
}
func TestCatalog_RPC_Filter(t *testing.T) {
t.Parallel()
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
testrpc.WaitForLeader(t, s1.RPC, "dc1")
// prep the cluster with some data we can use in our filters
registerTestCatalogEntries(t, codec)
// Run the tests against the test server
t.Run("ListNodes", func(t *testing.T) {
args := structs.DCSpecificRequest{
Datacenter: "dc1",
QueryOptions: structs.QueryOptions{Filter: "Meta.os == linux"},
}
out := new(structs.IndexedNodes)
require.NoError(t, msgpackrpc.CallWithCodec(codec, "Catalog.ListNodes", &args, out))
require.Len(t, out.Nodes, 2)
require.Condition(t, func() bool {
return (out.Nodes[0].Node == "foo" && out.Nodes[1].Node == "baz") ||
(out.Nodes[0].Node == "baz" && out.Nodes[1].Node == "foo")
})
args.Filter = "Meta.os == linux and Meta.env == qa"
out = new(structs.IndexedNodes)
require.NoError(t, msgpackrpc.CallWithCodec(codec, "Catalog.ListNodes", &args, out))
require.Len(t, out.Nodes, 1)
require.Equal(t, "baz", out.Nodes[0].Node)
})
t.Run("ServiceNodes", func(t *testing.T) {
args := structs.ServiceSpecificRequest{
Datacenter: "dc1",
ServiceName: "redis",
QueryOptions: structs.QueryOptions{Filter: "ServiceMeta.version == 1"},
}
out := new(structs.IndexedServiceNodes)
require.NoError(t, msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &args, &out))
require.Len(t, out.ServiceNodes, 2)
require.Condition(t, func() bool {
return (out.ServiceNodes[0].Node == "foo" && out.ServiceNodes[1].Node == "bar") ||
(out.ServiceNodes[0].Node == "bar" && out.ServiceNodes[1].Node == "foo")
})
args.Filter = "ServiceMeta.version == 2"
out = new(structs.IndexedServiceNodes)
require.NoError(t, msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &args, &out))
require.Len(t, out.ServiceNodes, 1)
require.Equal(t, "foo", out.ServiceNodes[0].Node)
})
t.Run("NodeServices", func(t *testing.T) {
args := structs.NodeSpecificRequest{
Datacenter: "dc1",
Node: "baz",
QueryOptions: structs.QueryOptions{Filter: "Service == web"},
}
out := new(structs.IndexedNodeServices)
require.NoError(t, msgpackrpc.CallWithCodec(codec, "Catalog.NodeServices", &args, &out))
require.Len(t, out.NodeServices.Services, 2)
args.Filter = "Service == web and Meta.version == 2"
out = new(structs.IndexedNodeServices)
require.NoError(t, msgpackrpc.CallWithCodec(codec, "Catalog.NodeServices", &args, &out))
require.Len(t, out.NodeServices.Services, 1)
})
}
2017-06-26 12:22:09 +00:00
func TestCatalog_ListNodes_StaleRead(t *testing.T) {
t.Parallel()
2014-04-19 00:26:59 +00:00
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec1 := rpcClient(t, s1)
defer codec1.Close()
testrpc.WaitForTestAgent(t, s1.RPC, "dc1")
2014-04-19 00:26:59 +00:00
2014-04-19 00:48:50 +00:00
dir2, s2 := testServerDCBootstrap(t, "dc1", false)
2014-04-19 00:26:59 +00:00
defer os.RemoveAll(dir2)
defer s2.Shutdown()
codec2 := rpcClient(t, s2)
defer codec2.Close()
2014-04-19 00:26:59 +00:00
// Try to join
2017-05-05 10:29:49 +00:00
joinLAN(t, s2, s1)
2014-04-19 00:26:59 +00:00
testrpc.WaitForTestAgent(t, s1.RPC, "dc1")
testrpc.WaitForTestAgent(t, s2.RPC, "dc1")
2014-04-19 00:26:59 +00:00
// Use the follower as the client
var codec rpc.ClientCodec
2014-04-19 00:26:59 +00:00
if !s1.IsLeader() {
codec = codec1
2014-04-19 00:26:59 +00:00
// Inject fake data on the follower!
if err := s1.fsm.State().EnsureNode(1, &structs.Node{Node: "foo", Address: "127.0.0.1"}); err != nil {
t.Fatalf("err: %v", err)
}
2014-04-19 00:26:59 +00:00
} else {
codec = codec2
2014-04-19 00:26:59 +00:00
// Inject fake data on the follower!
if err := s2.fsm.State().EnsureNode(1, &structs.Node{Node: "foo", Address: "127.0.0.1"}); err != nil {
t.Fatalf("err: %v", err)
}
2014-04-19 00:26:59 +00:00
}
args := structs.DCSpecificRequest{
Datacenter: "dc1",
QueryOptions: structs.QueryOptions{AllowStale: true},
}
2014-04-19 00:26:59 +00:00
var out structs.IndexedNodes
2019-07-12 15:52:26 +00:00
retry.Run(t, func(r *retry.R) {
if err := msgpackrpc.CallWithCodec(codec, "Catalog.ListNodes", &args, &out); err != nil {
r.Fatalf("err: %v", err)
2014-04-19 00:48:50 +00:00
}
2019-07-12 15:52:26 +00:00
found := false
for _, n := range out.Nodes {
if n.Node == "foo" {
found = true
}
}
if !found {
r.Fatalf("failed to find foo in %#v", out.Nodes)
}
if out.QueryMeta.LastContact == 0 {
r.Fatalf("should have a last contact time")
}
if !out.QueryMeta.KnownLeader {
r.Fatalf("should have known leader")
}
})
}
func TestCatalog_ListNodes_ConsistentRead_Fail(t *testing.T) {
t.Parallel()
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
dir2, s2 := testServerDCBootstrap(t, "dc1", false)
defer os.RemoveAll(dir2)
defer s2.Shutdown()
dir3, s3 := testServerDCBootstrap(t, "dc1", false)
defer os.RemoveAll(dir3)
defer s3.Shutdown()
// Try to join and wait for all servers to get promoted to voters.
joinLAN(t, s2, s1)
joinLAN(t, s3, s2)
servers := []*Server{s1, s2, s3}
retry.Run(t, func(r *retry.R) {
r.Check(wantRaft(servers))
for _, s := range servers {
r.Check(wantPeers(s, 3))
}
})
// Use the leader as the client, kill the followers.
var codec rpc.ClientCodec
for _, s := range servers {
if s.IsLeader() {
codec = rpcClient(t, s)
defer codec.Close()
} else {
s.Shutdown()
}
}
if codec == nil {
t.Fatalf("no leader")
}
args := structs.DCSpecificRequest{
Datacenter: "dc1",
QueryOptions: structs.QueryOptions{RequireConsistent: true},
}
var out structs.IndexedNodes
err := msgpackrpc.CallWithCodec(codec, "Catalog.ListNodes", &args, &out)
if err == nil || !strings.HasPrefix(err.Error(), "leadership lost") {
t.Fatalf("err: %v", err)
}
if out.QueryMeta.LastContact != 0 {
t.Fatalf("should not have a last contact time")
}
if out.QueryMeta.KnownLeader {
t.Fatalf("should have no known leader")
}
}
func TestCatalog_ListNodes_ConsistentRead(t *testing.T) {
t.Parallel()
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec1 := rpcClient(t, s1)
defer codec1.Close()
dir2, s2 := testServerDCBootstrap(t, "dc1", false)
defer os.RemoveAll(dir2)
defer s2.Shutdown()
codec2 := rpcClient(t, s2)
defer codec2.Close()
// Try to join
2017-05-05 10:29:49 +00:00
joinLAN(t, s2, s1)
testrpc.WaitForLeader(t, s1.RPC, "dc1")
testrpc.WaitForLeader(t, s2.RPC, "dc1")
// Use the leader as the client, kill the follower
var codec rpc.ClientCodec
if s1.IsLeader() {
codec = codec1
} else {
codec = codec2
}
args := structs.DCSpecificRequest{
Datacenter: "dc1",
QueryOptions: structs.QueryOptions{RequireConsistent: true},
}
var out structs.IndexedNodes
if err := msgpackrpc.CallWithCodec(codec, "Catalog.ListNodes", &args, &out); err != nil {
t.Fatalf("err: %v", err)
}
if out.QueryMeta.LastContact != 0 {
t.Fatalf("should not have a last contact time")
}
if !out.QueryMeta.KnownLeader {
t.Fatalf("should have known leader")
2014-04-19 00:26:59 +00:00
}
}
func TestCatalog_ListNodes_DistanceSort(t *testing.T) {
t.Parallel()
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
2015-10-15 23:07:16 +00:00
codec := rpcClient(t, s1)
defer codec.Close()
testrpc.WaitForLeader(t, s1.RPC, "dc1")
if err := s1.fsm.State().EnsureNode(1, &structs.Node{Node: "aaa", Address: "127.0.0.1"}); err != nil {
t.Fatalf("err: %v", err)
}
if err := s1.fsm.State().EnsureNode(2, &structs.Node{Node: "foo", Address: "127.0.0.2"}); err != nil {
t.Fatalf("err: %v", err)
}
if err := s1.fsm.State().EnsureNode(3, &structs.Node{Node: "bar", Address: "127.0.0.3"}); err != nil {
t.Fatalf("err: %v", err)
}
if err := s1.fsm.State().EnsureNode(4, &structs.Node{Node: "baz", Address: "127.0.0.4"}); err != nil {
t.Fatalf("err: %v", err)
}
// Set all but one of the nodes to known coordinates.
updates := structs.Coordinates{
{Node: "foo", Coord: lib.GenerateCoordinate(2 * time.Millisecond)},
{Node: "bar", Coord: lib.GenerateCoordinate(5 * time.Millisecond)},
{Node: "baz", Coord: lib.GenerateCoordinate(1 * time.Millisecond)},
}
if err := s1.fsm.State().CoordinateBatchUpdate(5, updates); err != nil {
t.Fatalf("err: %v", err)
}
// Query with no given source node, should get the natural order from
// the index.
args := structs.DCSpecificRequest{
Datacenter: "dc1",
}
var out structs.IndexedNodes
retry.Run(t, func(r *retry.R) {
2015-10-15 23:07:16 +00:00
msgpackrpc.CallWithCodec(codec, "Catalog.ListNodes", &args, &out)
if got, want := len(out.Nodes), 5; got != want {
r.Fatalf("got %d nodes want %d", got, want)
}
})
if out.Nodes[0].Node != "aaa" {
t.Fatalf("bad: %v", out)
}
if out.Nodes[1].Node != "bar" {
t.Fatalf("bad: %v", out)
}
if out.Nodes[2].Node != "baz" {
t.Fatalf("bad: %v", out)
}
if out.Nodes[3].Node != "foo" {
t.Fatalf("bad: %v", out)
}
if out.Nodes[4].Node != s1.config.NodeName {
t.Fatalf("bad: %v", out)
}
// Query relative to foo, note that there's no known coordinate for the
// default-added Serf node nor "aaa" so they will go at the end.
args = structs.DCSpecificRequest{
Datacenter: "dc1",
Source: structs.QuerySource{Datacenter: "dc1", Node: "foo"},
}
retry.Run(t, func(r *retry.R) {
2015-10-15 23:07:16 +00:00
msgpackrpc.CallWithCodec(codec, "Catalog.ListNodes", &args, &out)
if got, want := len(out.Nodes), 5; got != want {
r.Fatalf("got %d nodes want %d", got, want)
}
})
if out.Nodes[0].Node != "foo" {
t.Fatalf("bad: %v", out)
}
if out.Nodes[1].Node != "baz" {
t.Fatalf("bad: %v", out)
}
if out.Nodes[2].Node != "bar" {
t.Fatalf("bad: %v", out)
}
if out.Nodes[3].Node != "aaa" {
t.Fatalf("bad: %v", out)
}
if out.Nodes[4].Node != s1.config.NodeName {
t.Fatalf("bad: %v", out)
}
}
func TestCatalog_ListNodes_ACLFilter(t *testing.T) {
t.Parallel()
dir1, s1 := testServerWithConfig(t, func(c *Config) {
c.ACLDatacenter = "dc1"
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
c.ACLsEnabled = true
c.ACLMasterToken = "root"
c.ACLDefaultPolicy = "deny"
})
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
testrpc.WaitForLeader(t, s1.RPC, "dc1")
// We scope the reply in each of these since msgpack won't clear out an
// existing slice if the incoming one is nil, so it's best to start
// clean each time.
// The node policy should not be ignored.
args := structs.DCSpecificRequest{
Datacenter: "dc1",
}
{
reply := structs.IndexedNodes{}
if err := msgpackrpc.CallWithCodec(codec, "Catalog.ListNodes", &args, &reply); err != nil {
t.Fatalf("err: %v", err)
}
if len(reply.Nodes) != 0 {
t.Fatalf("bad: %v", reply.Nodes)
}
}
// Create an ACL that can read the node.
arg := structs.ACLRequest{
Datacenter: "dc1",
Op: structs.ACLSet,
ACL: structs.ACL{
Name: "User token",
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
Type: structs.ACLTokenTypeClient,
Rules: fmt.Sprintf(`
node "%s" {
policy = "read"
}
`, s1.config.NodeName),
},
WriteRequest: structs.WriteRequest{Token: "root"},
}
var id string
if err := msgpackrpc.CallWithCodec(codec, "ACL.Apply", &arg, &id); err != nil {
t.Fatalf("err: %v", err)
}
// Now try with the token and it will go through.
args.Token = id
{
reply := structs.IndexedNodes{}
if err := msgpackrpc.CallWithCodec(codec, "Catalog.ListNodes", &args, &reply); err != nil {
t.Fatalf("err: %v", err)
}
if len(reply.Nodes) != 1 {
t.Fatalf("bad: %v", reply.Nodes)
}
}
}
func Benchmark_Catalog_ListNodes(t *testing.B) {
dir1, s1 := testServer(nil)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(nil, s1)
defer codec.Close()
// Just add a node
if err := s1.fsm.State().EnsureNode(1, &structs.Node{Node: "foo", Address: "127.0.0.1"}); err != nil {
t.Fatalf("err: %v", err)
}
args := structs.DCSpecificRequest{
Datacenter: "dc1",
}
for i := 0; i < t.N; i++ {
var out structs.IndexedNodes
if err := msgpackrpc.CallWithCodec(codec, "Catalog.ListNodes", &args, &out); err != nil {
t.Fatalf("err: %v", err)
}
}
}
func TestCatalog_ListServices(t *testing.T) {
t.Parallel()
2013-12-12 19:07:14 +00:00
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
2013-12-12 19:07:14 +00:00
args := structs.DCSpecificRequest{
Datacenter: "dc1",
}
var out structs.IndexedServices
err := msgpackrpc.CallWithCodec(codec, "Catalog.ListServices", &args, &out)
if err != nil {
2013-12-12 19:07:14 +00:00
t.Fatalf("err: %v", err)
}
testrpc.WaitForLeader(t, s1.RPC, "dc1")
2013-12-12 19:07:14 +00:00
// Just add a node
if err := s1.fsm.State().EnsureNode(1, &structs.Node{Node: "foo", Address: "127.0.0.1"}); err != nil {
t.Fatalf("err: %v", err)
}
if err := s1.fsm.State().EnsureService(2, "foo", &structs.NodeService{ID: "db", Service: "db", Tags: []string{"primary"}, Address: "127.0.0.1", Port: 5000}); err != nil {
t.Fatalf("err: %v", err)
}
2013-12-12 19:07:14 +00:00
if err := msgpackrpc.CallWithCodec(codec, "Catalog.ListServices", &args, &out); err != nil {
2013-12-12 19:07:14 +00:00
t.Fatalf("err: %v", err)
}
if len(out.Services) != 2 {
2014-01-10 01:57:13 +00:00
t.Fatalf("bad: %v", out)
}
for _, s := range out.Services {
if s == nil {
t.Fatalf("bad: %v", s)
}
}
2014-01-10 01:57:13 +00:00
// Consul service should auto-register
2014-04-03 19:03:10 +00:00
if _, ok := out.Services["consul"]; !ok {
t.Fatalf("bad: %v", out)
}
if len(out.Services["db"]) != 1 {
t.Fatalf("bad: %v", out)
}
if out.Services["db"][0] != "primary" {
t.Fatalf("bad: %v", out)
}
}
func TestCatalog_ListServices_NodeMetaFilter(t *testing.T) {
t.Parallel()
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
testrpc.WaitForTestAgent(t, s1.RPC, "dc1")
// Add a new node with the right meta k/v pair
node := &structs.Node{Node: "foo", Address: "127.0.0.1", Meta: map[string]string{"somekey": "somevalue"}}
if err := s1.fsm.State().EnsureNode(1, node); err != nil {
t.Fatalf("err: %v", err)
}
// Add a service to the new node
if err := s1.fsm.State().EnsureService(2, "foo", &structs.NodeService{ID: "db", Service: "db", Tags: []string{"primary"}, Address: "127.0.0.1", Port: 5000}); err != nil {
t.Fatalf("err: %v", err)
}
// Filter by a specific meta k/v pair
args := structs.DCSpecificRequest{
Datacenter: "dc1",
NodeMetaFilters: map[string]string{
"somekey": "somevalue",
},
}
var out structs.IndexedServices
if err := msgpackrpc.CallWithCodec(codec, "Catalog.ListServices", &args, &out); err != nil {
t.Fatalf("err: %v", err)
}
if len(out.Services) != 1 {
t.Fatalf("bad: %v", out)
}
if out.Services["db"] == nil {
t.Fatalf("bad: %v", out.Services["db"])
}
if len(out.Services["db"]) != 1 {
t.Fatalf("bad: %v", out)
}
if out.Services["db"][0] != "primary" {
t.Fatalf("bad: %v", out)
}
// Now filter on a nonexistent meta k/v pair
args = structs.DCSpecificRequest{
Datacenter: "dc1",
NodeMetaFilters: map[string]string{
"somekey": "invalid",
},
}
out = structs.IndexedServices{}
err := msgpackrpc.CallWithCodec(codec, "Catalog.ListServices", &args, &out)
if err != nil {
t.Fatalf("err: %v", err)
}
// Should get an empty list of nodes back
if len(out.Services) != 0 {
t.Fatalf("bad: %v", out.Services)
}
}
func TestCatalog_ListServices_Blocking(t *testing.T) {
t.Parallel()
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
args := structs.DCSpecificRequest{
Datacenter: "dc1",
}
var out structs.IndexedServices
testrpc.WaitForLeader(t, s1.RPC, "dc1")
// Run the query
if err := msgpackrpc.CallWithCodec(codec, "Catalog.ListServices", &args, &out); err != nil {
t.Fatalf("err: %v", err)
}
// Setup a blocking query
args.MinQueryIndex = out.Index
args.MaxQueryTime = time.Second
// Async cause a change
idx := out.Index
start := time.Now()
go func() {
2014-05-09 00:36:25 +00:00
time.Sleep(100 * time.Millisecond)
if err := s1.fsm.State().EnsureNode(idx+1, &structs.Node{Node: "foo", Address: "127.0.0.1"}); err != nil {
t.Errorf("err: %v", err)
return
}
if err := s1.fsm.State().EnsureService(idx+2, "foo", &structs.NodeService{ID: "db", Service: "db", Tags: []string{"primary"}, Address: "127.0.0.1", Port: 5000}); err != nil {
t.Errorf("err: %v", err)
}
}()
// Re-run the query
out = structs.IndexedServices{}
if err := msgpackrpc.CallWithCodec(codec, "Catalog.ListServices", &args, &out); err != nil {
t.Fatalf("err: %v", err)
}
// Should block at least 100ms
if time.Since(start) < 100*time.Millisecond {
t.Fatalf("too fast")
}
// Check the indexes
if out.Index != idx+2 {
2013-12-12 19:07:14 +00:00
t.Fatalf("bad: %v", out)
}
// Should find the service
if len(out.Services) != 2 {
2013-12-12 19:07:14 +00:00
t.Fatalf("bad: %v", out)
}
}
func TestCatalog_ListServices_Timeout(t *testing.T) {
t.Parallel()
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
args := structs.DCSpecificRequest{
Datacenter: "dc1",
}
var out structs.IndexedServices
testrpc.WaitForLeader(t, s1.RPC, "dc1")
// Run the query
if err := msgpackrpc.CallWithCodec(codec, "Catalog.ListServices", &args, &out); err != nil {
t.Fatalf("err: %v", err)
}
// Setup a blocking query
args.MinQueryIndex = out.Index
2014-05-09 00:36:25 +00:00
args.MaxQueryTime = 100 * time.Millisecond
// Re-run the query
start := time.Now()
out = structs.IndexedServices{}
if err := msgpackrpc.CallWithCodec(codec, "Catalog.ListServices", &args, &out); err != nil {
t.Fatalf("err: %v", err)
}
// Should block at least 100ms
if time.Since(start) < 100*time.Millisecond {
t.Fatalf("too fast")
}
// Check the indexes, should not change
if out.Index != args.MinQueryIndex {
2013-12-12 19:07:14 +00:00
t.Fatalf("bad: %v", out)
}
}
2013-12-12 19:37:19 +00:00
func TestCatalog_ListServices_Stale(t *testing.T) {
t.Parallel()
dir1, s1 := testServerWithConfig(t, func(c *Config) {
c.ACLDatacenter = "dc1"
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
c.ACLsEnabled = true
})
2014-04-21 18:57:39 +00:00
defer os.RemoveAll(dir1)
defer s1.Shutdown()
testrpc.WaitForTestAgent(t, s1.RPC, "dc1")
dir2, s2 := testServerWithConfig(t, func(c *Config) {
c.ACLDatacenter = "dc1" // Enable ACLs!
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
c.ACLsEnabled = true
c.Bootstrap = false // Disable bootstrap
})
defer os.RemoveAll(dir2)
defer s2.Shutdown()
2014-04-21 18:57:39 +00:00
args := structs.DCSpecificRequest{
Datacenter: "dc1",
}
args.AllowStale = true
var out structs.IndexedServices
// Inject a node
if err := s1.fsm.State().EnsureNode(3, &structs.Node{Node: "foo", Address: "127.0.0.1"}); err != nil {
t.Fatalf("err: %v", err)
}
codec := rpcClient(t, s2)
defer codec.Close()
// Run the query, do not wait for leader, never any contact with leader, should fail
if err := msgpackrpc.CallWithCodec(codec, "Catalog.ListServices", &args, &out); err == nil || err.Error() != structs.ErrNoLeader.Error() {
t.Fatalf("expected %v but got err: %v and %v", structs.ErrNoLeader, err, out)
}
// Try to join
joinLAN(t, s2, s1)
retry.Run(t, func(r *retry.R) { r.Check(wantRaft([]*Server{s1, s2})) })
waitForLeader(s1, s2)
testrpc.WaitForLeader(t, s2.RPC, "dc1")
2014-04-21 18:57:39 +00:00
2019-07-12 15:52:26 +00:00
retry.Run(t, func(r *retry.R) {
if err := msgpackrpc.CallWithCodec(codec, "Catalog.ListServices", &args, &out); err != nil {
r.Fatalf("err: %v", err)
}
// Should find the services
if len(out.Services) != 1 {
r.Fatalf("bad: %#v", out.Services)
}
if !out.KnownLeader {
r.Fatalf("should have a leader: %v", out)
}
})
s1.Leave()
s1.Shutdown()
testrpc.WaitUntilNoLeader(t, s2.RPC, "dc1")
args.AllowStale = false
// Since the leader is now down, non-stale query should fail now
if err := msgpackrpc.CallWithCodec(codec, "Catalog.ListServices", &args, &out); err == nil || err.Error() != structs.ErrNoLeader.Error() {
t.Fatalf("expected %v but got err: %v and %v", structs.ErrNoLeader, err, out)
}
// With stale, request should still work
args.AllowStale = true
if err := msgpackrpc.CallWithCodec(codec, "Catalog.ListServices", &args, &out); err != nil {
2014-04-21 18:57:39 +00:00
t.Fatalf("err: %v", err)
}
// Should find old service
2014-04-21 18:57:39 +00:00
if len(out.Services) != 1 {
t.Fatalf("bad: %#v", out)
2014-04-21 18:57:39 +00:00
}
if out.KnownLeader {
t.Fatalf("should not have a leader anymore: %#v", out)
2014-04-21 18:57:39 +00:00
}
}
func TestCatalog_ListServiceNodes(t *testing.T) {
t.Parallel()
2013-12-12 19:37:19 +00:00
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
2013-12-12 19:37:19 +00:00
2014-01-08 21:52:09 +00:00
args := structs.ServiceSpecificRequest{
2013-12-12 19:37:19 +00:00
Datacenter: "dc1",
ServiceName: "db",
ServiceTags: []string{"slave"},
2013-12-12 19:37:19 +00:00
TagFilter: false,
}
var out structs.IndexedServiceNodes
err := msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &args, &out)
if err != nil {
2013-12-12 19:37:19 +00:00
t.Fatalf("err: %v", err)
}
testrpc.WaitForLeader(t, s1.RPC, "dc1")
2013-12-12 19:37:19 +00:00
// Just add a node
if err := s1.fsm.State().EnsureNode(1, &structs.Node{Node: "foo", Address: "127.0.0.1"}); err != nil {
t.Fatalf("err: %v", err)
}
if err := s1.fsm.State().EnsureService(2, "foo", &structs.NodeService{ID: "db", Service: "db", Tags: []string{"primary"}, Address: "127.0.0.1", Port: 5000}); err != nil {
t.Fatalf("err: %v", err)
}
2013-12-12 19:37:19 +00:00
if err := msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &args, &out); err != nil {
2013-12-12 19:37:19 +00:00
t.Fatalf("err: %v", err)
}
if len(out.ServiceNodes) != 1 {
2013-12-12 19:37:19 +00:00
t.Fatalf("bad: %v", out)
}
// Try with a filter
args.TagFilter = true
out = structs.IndexedServiceNodes{}
2013-12-12 19:37:19 +00:00
if err := msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &args, &out); err != nil {
2013-12-12 19:37:19 +00:00
t.Fatalf("err: %v", err)
}
if len(out.ServiceNodes) != 0 {
2013-12-12 19:37:19 +00:00
t.Fatalf("bad: %v", out)
}
}
2013-12-12 19:46:25 +00:00
func TestCatalog_ListServiceNodes_ByAddress(t *testing.T) {
t.Parallel()
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
testrpc.WaitForLeader(t, s1.RPC, "dc1")
args := structs.ServiceSpecificRequest{
Datacenter: "dc1",
ServiceName: "db",
}
var out structs.IndexedServiceNodes
require.NoError(t, msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &args, &out))
fooAddress := "10.1.2.3"
fooPort := 1111
fooTaggedAddresses := map[string]structs.ServiceAddress{
"lan": structs.ServiceAddress{
Address: "10.1.2.3",
Port: fooPort,
},
"wan": structs.ServiceAddress{
Address: "198.18.1.2",
Port: fooPort,
},
}
barAddress := "10.1.2.3"
barPort := 2222
barTaggedAddresses := map[string]structs.ServiceAddress{
"lan": structs.ServiceAddress{
Address: "10.1.2.3",
Port: barPort,
},
"wan": structs.ServiceAddress{
Address: "198.18.2.3",
Port: barPort,
},
}
bazAddress := "192.168.1.35"
bazPort := 2222
bazTaggedAddresses := map[string]structs.ServiceAddress{
"lan": structs.ServiceAddress{
Address: "192.168.1.35",
Port: barPort,
},
"wan": structs.ServiceAddress{
Address: "198.18.2.4",
Port: barPort,
},
}
// Just add a node
require.NoError(t, s1.fsm.State().EnsureNode(1, &structs.Node{Node: "node", Address: "127.0.0.1"}))
require.NoError(t, s1.fsm.State().EnsureService(2, "node", &structs.NodeService{ID: "foo", Service: "db", Address: fooAddress, TaggedAddresses: fooTaggedAddresses, Port: fooPort}))
require.NoError(t, s1.fsm.State().EnsureService(2, "node", &structs.NodeService{ID: "bar", Service: "db", Address: barAddress, TaggedAddresses: barTaggedAddresses, Port: barPort}))
require.NoError(t, s1.fsm.State().EnsureService(2, "node", &structs.NodeService{ID: "baz", Service: "db", Address: bazAddress, TaggedAddresses: bazTaggedAddresses, Port: bazPort}))
require.NoError(t, msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &args, &out))
require.Len(t, out.ServiceNodes, 3)
// Try with an address that would match foo & bar
args.ServiceAddress = "10.1.2.3"
out = structs.IndexedServiceNodes{}
require.NoError(t, msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &args, &out))
require.Len(t, out.ServiceNodes, 2)
for _, sn := range out.ServiceNodes {
require.True(t, sn.ServiceID == "foo" || sn.ServiceID == "bar")
}
// Try with an address that would match just bar
args.ServiceAddress = "198.18.2.3"
out = structs.IndexedServiceNodes{}
require.NoError(t, msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &args, &out))
require.Len(t, out.ServiceNodes, 1)
require.Equal(t, "bar", out.ServiceNodes[0].ServiceID)
}
// TestCatalog_ListServiceNodes_ServiceTags_V1_2_3Compat asserts the compatibility between <=v1.2.3 agents and >=v1.3.0 servers
// see https://github.com/hashicorp/consul/issues/4922
func TestCatalog_ListServiceNodes_ServiceTags_V1_2_3Compat(t *testing.T) {
t.Parallel()
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
testrpc.WaitForTestAgent(t, s1.RPC, "dc1")
err := s1.fsm.State().EnsureNode(1, &structs.Node{Node: "foo", Address: "127.0.0.1"})
require.NoError(t, err)
// register two service instances with different tags
err = s1.fsm.State().EnsureService(2, "foo", &structs.NodeService{ID: "db1", Service: "db", Tags: []string{"primary"}, Address: "127.0.0.1", Port: 5000})
require.NoError(t, err)
err = s1.fsm.State().EnsureService(2, "foo", &structs.NodeService{ID: "db2", Service: "db", Tags: []string{"secondary"}, Address: "127.0.0.1", Port: 5001})
require.NoError(t, err)
// DEPRECATED (singular-service-tag) - remove this when backwards RPC compat
// with 1.2.x is not required.
// make a request with the <=1.2.3 ServiceTag tag field (vs ServiceTags)
args := structs.ServiceSpecificRequest{
Datacenter: "dc1",
ServiceName: "db",
ServiceTag: "primary",
TagFilter: true,
}
var out structs.IndexedServiceNodes
err = msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &args, &out)
require.NoError(t, err)
// nodes should be filtered, even when using the old ServiceTag field
require.Equal(t, 1, len(out.ServiceNodes))
require.Equal(t, "db1", out.ServiceNodes[0].ServiceID)
// DEPRECATED (singular-service-tag) - remove this when backwards RPC compat
// with 1.2.x is not required.
// test with the other tag
args = structs.ServiceSpecificRequest{
Datacenter: "dc1",
ServiceName: "db",
ServiceTag: "secondary",
TagFilter: true,
}
out = structs.IndexedServiceNodes{}
err = msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &args, &out)
require.NoError(t, err)
require.Equal(t, 1, len(out.ServiceNodes))
require.Equal(t, "db2", out.ServiceNodes[0].ServiceID)
// no tag, both instances
args = structs.ServiceSpecificRequest{
Datacenter: "dc1",
ServiceName: "db",
}
out = structs.IndexedServiceNodes{}
err = msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &args, &out)
require.NoError(t, err)
require.Equal(t, 2, len(out.ServiceNodes))
// DEPRECATED (singular-service-tag) - remove this when backwards RPC compat
// with 1.2.x is not required.
// when both ServiceTag and ServiceTags fields are populated, use ServiceTag
args = structs.ServiceSpecificRequest{
Datacenter: "dc1",
ServiceName: "db",
ServiceTag: "primary",
ServiceTags: []string{"secondary"},
TagFilter: true,
}
out = structs.IndexedServiceNodes{}
err = msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &args, &out)
require.NoError(t, err)
require.Equal(t, "db1", out.ServiceNodes[0].ServiceID)
}
func TestCatalog_ListServiceNodes_NodeMetaFilter(t *testing.T) {
t.Parallel()
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
testrpc.WaitForLeader(t, s1.RPC, "dc1")
// Add 2 nodes with specific meta maps
node := &structs.Node{Node: "foo", Address: "127.0.0.1", Meta: map[string]string{"somekey": "somevalue", "common": "1"}}
if err := s1.fsm.State().EnsureNode(1, node); err != nil {
t.Fatalf("err: %v", err)
}
node2 := &structs.Node{Node: "bar", Address: "127.0.0.2", Meta: map[string]string{"common": "1"}}
if err := s1.fsm.State().EnsureNode(2, node2); err != nil {
t.Fatalf("err: %v", err)
}
if err := s1.fsm.State().EnsureService(3, "foo", &structs.NodeService{ID: "db", Service: "db", Tags: []string{"primary", "v2"}, Address: "127.0.0.1", Port: 5000}); err != nil {
t.Fatalf("err: %v", err)
}
if err := s1.fsm.State().EnsureService(4, "bar", &structs.NodeService{ID: "db2", Service: "db", Tags: []string{"secondary", "v2"}, Address: "127.0.0.2", Port: 5000}); err != nil {
t.Fatalf("err: %v", err)
}
cases := []struct {
filters map[string]string
tags []string
services structs.ServiceNodes
}{
// Basic meta filter
{
filters: map[string]string{"somekey": "somevalue"},
services: structs.ServiceNodes{&structs.ServiceNode{Node: "foo", ServiceID: "db"}},
},
// Basic meta filter, tag
{
filters: map[string]string{"somekey": "somevalue"},
tags: []string{"primary"},
services: structs.ServiceNodes{&structs.ServiceNode{Node: "foo", ServiceID: "db"}},
},
// Common meta filter
{
filters: map[string]string{"common": "1"},
services: structs.ServiceNodes{
&structs.ServiceNode{Node: "bar", ServiceID: "db2"},
&structs.ServiceNode{Node: "foo", ServiceID: "db"},
},
},
// Common meta filter, tag
{
filters: map[string]string{"common": "1"},
tags: []string{"secondary"},
services: structs.ServiceNodes{
&structs.ServiceNode{Node: "bar", ServiceID: "db2"},
},
},
// Invalid meta filter
{
filters: map[string]string{"invalid": "nope"},
services: structs.ServiceNodes{},
},
// Multiple filter values
{
filters: map[string]string{"somekey": "somevalue", "common": "1"},
services: structs.ServiceNodes{&structs.ServiceNode{Node: "foo", ServiceID: "db"}},
},
// Multiple filter values, tag
{
filters: map[string]string{"somekey": "somevalue", "common": "1"},
tags: []string{"primary"},
services: structs.ServiceNodes{&structs.ServiceNode{Node: "foo", ServiceID: "db"}},
},
// Common meta filter, single tag
{
filters: map[string]string{"common": "1"},
tags: []string{"v2"},
services: structs.ServiceNodes{
&structs.ServiceNode{Node: "bar", ServiceID: "db2"},
&structs.ServiceNode{Node: "foo", ServiceID: "db"},
},
},
// Common meta filter, multiple tags
{
filters: map[string]string{"common": "1"},
tags: []string{"v2", "primary"},
services: structs.ServiceNodes{&structs.ServiceNode{Node: "foo", ServiceID: "db"}},
},
}
for _, tc := range cases {
args := structs.ServiceSpecificRequest{
Datacenter: "dc1",
NodeMetaFilters: tc.filters,
ServiceName: "db",
ServiceTags: tc.tags,
TagFilter: len(tc.tags) > 0,
}
var out structs.IndexedServiceNodes
require.NoError(t, msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &args, &out))
require.Len(t, out.ServiceNodes, len(tc.services))
for i, serviceNode := range out.ServiceNodes {
if serviceNode.Node != tc.services[i].Node || serviceNode.ServiceID != tc.services[i].ServiceID {
t.Fatalf("bad: %v, %v filters: %v", serviceNode, tc.services[i], tc.filters)
}
}
}
}
func TestCatalog_ListServiceNodes_DistanceSort(t *testing.T) {
t.Parallel()
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
2015-10-15 23:07:16 +00:00
codec := rpcClient(t, s1)
defer codec.Close()
args := structs.ServiceSpecificRequest{
Datacenter: "dc1",
ServiceName: "db",
}
var out structs.IndexedServiceNodes
2015-10-15 23:07:16 +00:00
err := msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &args, &out)
if err != nil {
t.Fatalf("err: %v", err)
}
testrpc.WaitForLeader(t, s1.RPC, "dc1")
// Add a few nodes for the associated services.
s1.fsm.State().EnsureNode(1, &structs.Node{Node: "aaa", Address: "127.0.0.1"})
s1.fsm.State().EnsureService(2, "aaa", &structs.NodeService{ID: "db", Service: "db", Tags: []string{"primary"}, Address: "127.0.0.1", Port: 5000})
s1.fsm.State().EnsureNode(3, &structs.Node{Node: "foo", Address: "127.0.0.2"})
s1.fsm.State().EnsureService(4, "foo", &structs.NodeService{ID: "db", Service: "db", Tags: []string{"primary"}, Address: "127.0.0.2", Port: 5000})
s1.fsm.State().EnsureNode(5, &structs.Node{Node: "bar", Address: "127.0.0.3"})
s1.fsm.State().EnsureService(6, "bar", &structs.NodeService{ID: "db", Service: "db", Tags: []string{"primary"}, Address: "127.0.0.3", Port: 5000})
s1.fsm.State().EnsureNode(7, &structs.Node{Node: "baz", Address: "127.0.0.4"})
s1.fsm.State().EnsureService(8, "baz", &structs.NodeService{ID: "db", Service: "db", Tags: []string{"primary"}, Address: "127.0.0.4", Port: 5000})
// Set all but one of the nodes to known coordinates.
updates := structs.Coordinates{
{Node: "foo", Coord: lib.GenerateCoordinate(2 * time.Millisecond)},
{Node: "bar", Coord: lib.GenerateCoordinate(5 * time.Millisecond)},
{Node: "baz", Coord: lib.GenerateCoordinate(1 * time.Millisecond)},
}
if err := s1.fsm.State().CoordinateBatchUpdate(9, updates); err != nil {
t.Fatalf("err: %v", err)
}
// Query with no given source node, should get the natural order from
// the index.
2015-10-15 23:07:16 +00:00
if err := msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &args, &out); err != nil {
t.Fatalf("err: %v", err)
}
if len(out.ServiceNodes) != 4 {
t.Fatalf("bad: %v", out)
}
if out.ServiceNodes[0].Node != "aaa" {
t.Fatalf("bad: %v", out)
}
if out.ServiceNodes[1].Node != "bar" {
t.Fatalf("bad: %v", out)
}
if out.ServiceNodes[2].Node != "baz" {
t.Fatalf("bad: %v", out)
}
if out.ServiceNodes[3].Node != "foo" {
t.Fatalf("bad: %v", out)
}
// Query relative to foo, note that there's no known coordinate for "aaa"
// so it will go at the end.
args = structs.ServiceSpecificRequest{
Datacenter: "dc1",
ServiceName: "db",
Source: structs.QuerySource{Datacenter: "dc1", Node: "foo"},
}
2015-10-15 23:07:16 +00:00
if err := msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &args, &out); err != nil {
t.Fatalf("err: %v", err)
}
if len(out.ServiceNodes) != 4 {
t.Fatalf("bad: %v", out)
}
if out.ServiceNodes[0].Node != "foo" {
t.Fatalf("bad: %v", out)
}
if out.ServiceNodes[1].Node != "baz" {
t.Fatalf("bad: %v", out)
}
if out.ServiceNodes[2].Node != "bar" {
t.Fatalf("bad: %v", out)
}
if out.ServiceNodes[3].Node != "aaa" {
t.Fatalf("bad: %v", out)
}
}
func TestCatalog_ListServiceNodes_ConnectProxy(t *testing.T) {
t.Parallel()
assert := assert.New(t)
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
testrpc.WaitForLeader(t, s1.RPC, "dc1")
// Register the service
args := structs.TestRegisterRequestProxy(t)
var out struct{}
assert.Nil(msgpackrpc.CallWithCodec(codec, "Catalog.Register", args, &out))
// List
req := structs.ServiceSpecificRequest{
Datacenter: "dc1",
ServiceName: args.Service.Service,
TagFilter: false,
}
var resp structs.IndexedServiceNodes
assert.Nil(msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &req, &resp))
assert.Len(resp.ServiceNodes, 1)
v := resp.ServiceNodes[0]
assert.Equal(structs.ServiceKindConnectProxy, v.ServiceKind)
Add Proxy Upstreams to Service Definition (#4639) * Refactor Service Definition ProxyDestination. This includes: - Refactoring all internal structs used - Updated tests for both deprecated and new input for: - Agent Services endpoint response - Agent Service endpoint response - Agent Register endpoint - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Register - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Services endpoint response - Catalog Node endpoint response - Catalog Service endpoint response - Updated API tests for all of the above too (both deprecated and new forms of register) TODO: - config package changes for on-disk service definitions - proxy config endpoint - built-in proxy support for new fields * Agent proxy config endpoint updated with upstreams * Config file changes for upstreams. * Add upstream opaque config and update all tests to ensure it works everywhere. * Built in proxy working with new Upstreams config * Command fixes and deprecations * Fix key translation, upstream type defaults and a spate of other subtele bugs found with ned to end test scripts... TODO: tests still failing on one case that needs a fix. I think it's key translation for upstreams nested in Managed proxy struct. * Fix translated keys in API registration. ≈ * Fixes from docs - omit some empty undocumented fields in API - Bring back ServiceProxyDestination in Catalog responses to not break backwards compat - this was removed assuming it was only used internally. * Documentation updates for Upstreams in service definition * Fixes for tests broken by many refactors. * Enable travis on f-connect branch in this branch too. * Add consistent Deprecation comments to ProxyDestination uses * Update version number on deprecation notices, and correct upstream datacenter field with explanation in docs
2018-09-12 16:07:47 +00:00
assert.Equal(args.Service.Proxy.DestinationServiceName, v.ServiceProxy.DestinationServiceName)
}
func TestCatalog_ServiceNodes_Gateway(t *testing.T) {
t.Parallel()
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
testrpc.WaitForTestAgent(t, s1.RPC, "dc1")
{
var out struct{}
// Register a service "api"
args := structs.TestRegisterRequest(t)
args.Service.Service = "api"
args.Check = &structs.HealthCheck{
Name: "api",
Status: api.HealthPassing,
ServiceID: args.Service.Service,
}
assert.Nil(t, msgpackrpc.CallWithCodec(codec, "Catalog.Register", &args, &out))
// Register a proxy for api
args = structs.TestRegisterRequestProxy(t)
args.Service.Service = "api-proxy"
args.Service.Proxy.DestinationServiceName = "api"
args.Check = &structs.HealthCheck{
Name: "api-proxy",
Status: api.HealthPassing,
ServiceID: args.Service.Service,
}
assert.Nil(t, msgpackrpc.CallWithCodec(codec, "Catalog.Register", &args, &out))
// Register a service "web"
args = structs.TestRegisterRequest(t)
args.Check = &structs.HealthCheck{
Name: "web",
Status: api.HealthPassing,
ServiceID: args.Service.Service,
}
assert.Nil(t, msgpackrpc.CallWithCodec(codec, "Catalog.Register", &args, &out))
// Register a proxy for web
args = structs.TestRegisterRequestProxy(t)
args.Check = &structs.HealthCheck{
Name: "web-proxy",
Status: api.HealthPassing,
ServiceID: args.Service.Service,
}
assert.Nil(t, msgpackrpc.CallWithCodec(codec, "Catalog.Register", &args, &out))
// Register a gateway for web
args = &structs.RegisterRequest{
Datacenter: "dc1",
Node: "foo",
Address: "127.0.0.1",
Service: &structs.NodeService{
Kind: structs.ServiceKindTerminatingGateway,
Service: "gateway",
Port: 443,
},
Check: &structs.HealthCheck{
Name: "gateway",
Status: api.HealthPassing,
ServiceID: args.Service.Service,
},
}
assert.Nil(t, msgpackrpc.CallWithCodec(codec, "Catalog.Register", &args, &out))
entryArgs := &structs.ConfigEntryRequest{
Op: structs.ConfigEntryUpsert,
Datacenter: "dc1",
Entry: &structs.TerminatingGatewayConfigEntry{
Kind: "terminating-gateway",
Name: "gateway",
Services: []structs.LinkedService{
{
Name: "web",
},
},
},
}
var entryResp bool
assert.Nil(t, msgpackrpc.CallWithCodec(codec, "ConfigEntry.Apply", &entryArgs, &entryResp))
}
retry.Run(t, func(r *retry.R) {
// List should return both the terminating-gateway and the connect-proxy associated with web
req := structs.ServiceSpecificRequest{
Connect: true,
Datacenter: "dc1",
ServiceName: "web",
}
var resp structs.IndexedServiceNodes
assert.Nil(r, msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &req, &resp))
assert.Len(r, resp.ServiceNodes, 2)
// Check sidecar
assert.Equal(r, structs.ServiceKindConnectProxy, resp.ServiceNodes[0].ServiceKind)
assert.Equal(r, "foo", resp.ServiceNodes[0].Node)
assert.Equal(r, "web-proxy", resp.ServiceNodes[0].ServiceName)
assert.Equal(r, "web-proxy", resp.ServiceNodes[0].ServiceID)
assert.Equal(r, "web", resp.ServiceNodes[0].ServiceProxy.DestinationServiceName)
assert.Equal(r, 2222, resp.ServiceNodes[0].ServicePort)
// Check gateway
assert.Equal(r, structs.ServiceKindTerminatingGateway, resp.ServiceNodes[1].ServiceKind)
assert.Equal(r, "foo", resp.ServiceNodes[1].Node)
assert.Equal(r, "gateway", resp.ServiceNodes[1].ServiceName)
assert.Equal(r, "gateway", resp.ServiceNodes[1].ServiceID)
assert.Equal(r, 443, resp.ServiceNodes[1].ServicePort)
})
}
func TestCatalog_ListServiceNodes_ConnectDestination(t *testing.T) {
t.Parallel()
assert := assert.New(t)
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
testrpc.WaitForLeader(t, s1.RPC, "dc1")
// Register the proxy service
args := structs.TestRegisterRequestProxy(t)
var out struct{}
assert.Nil(msgpackrpc.CallWithCodec(codec, "Catalog.Register", args, &out))
// Register the service
{
Add Proxy Upstreams to Service Definition (#4639) * Refactor Service Definition ProxyDestination. This includes: - Refactoring all internal structs used - Updated tests for both deprecated and new input for: - Agent Services endpoint response - Agent Service endpoint response - Agent Register endpoint - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Register - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Services endpoint response - Catalog Node endpoint response - Catalog Service endpoint response - Updated API tests for all of the above too (both deprecated and new forms of register) TODO: - config package changes for on-disk service definitions - proxy config endpoint - built-in proxy support for new fields * Agent proxy config endpoint updated with upstreams * Config file changes for upstreams. * Add upstream opaque config and update all tests to ensure it works everywhere. * Built in proxy working with new Upstreams config * Command fixes and deprecations * Fix key translation, upstream type defaults and a spate of other subtele bugs found with ned to end test scripts... TODO: tests still failing on one case that needs a fix. I think it's key translation for upstreams nested in Managed proxy struct. * Fix translated keys in API registration. ≈ * Fixes from docs - omit some empty undocumented fields in API - Bring back ServiceProxyDestination in Catalog responses to not break backwards compat - this was removed assuming it was only used internally. * Documentation updates for Upstreams in service definition * Fixes for tests broken by many refactors. * Enable travis on f-connect branch in this branch too. * Add consistent Deprecation comments to ProxyDestination uses * Update version number on deprecation notices, and correct upstream datacenter field with explanation in docs
2018-09-12 16:07:47 +00:00
dst := args.Service.Proxy.DestinationServiceName
args := structs.TestRegisterRequest(t)
args.Service.Service = dst
var out struct{}
assert.Nil(msgpackrpc.CallWithCodec(codec, "Catalog.Register", args, &out))
}
// List
req := structs.ServiceSpecificRequest{
Connect: true,
Datacenter: "dc1",
Add Proxy Upstreams to Service Definition (#4639) * Refactor Service Definition ProxyDestination. This includes: - Refactoring all internal structs used - Updated tests for both deprecated and new input for: - Agent Services endpoint response - Agent Service endpoint response - Agent Register endpoint - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Register - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Services endpoint response - Catalog Node endpoint response - Catalog Service endpoint response - Updated API tests for all of the above too (both deprecated and new forms of register) TODO: - config package changes for on-disk service definitions - proxy config endpoint - built-in proxy support for new fields * Agent proxy config endpoint updated with upstreams * Config file changes for upstreams. * Add upstream opaque config and update all tests to ensure it works everywhere. * Built in proxy working with new Upstreams config * Command fixes and deprecations * Fix key translation, upstream type defaults and a spate of other subtele bugs found with ned to end test scripts... TODO: tests still failing on one case that needs a fix. I think it's key translation for upstreams nested in Managed proxy struct. * Fix translated keys in API registration. ≈ * Fixes from docs - omit some empty undocumented fields in API - Bring back ServiceProxyDestination in Catalog responses to not break backwards compat - this was removed assuming it was only used internally. * Documentation updates for Upstreams in service definition * Fixes for tests broken by many refactors. * Enable travis on f-connect branch in this branch too. * Add consistent Deprecation comments to ProxyDestination uses * Update version number on deprecation notices, and correct upstream datacenter field with explanation in docs
2018-09-12 16:07:47 +00:00
ServiceName: args.Service.Proxy.DestinationServiceName,
}
var resp structs.IndexedServiceNodes
assert.Nil(msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &req, &resp))
assert.Len(resp.ServiceNodes, 1)
v := resp.ServiceNodes[0]
assert.Equal(structs.ServiceKindConnectProxy, v.ServiceKind)
Add Proxy Upstreams to Service Definition (#4639) * Refactor Service Definition ProxyDestination. This includes: - Refactoring all internal structs used - Updated tests for both deprecated and new input for: - Agent Services endpoint response - Agent Service endpoint response - Agent Register endpoint - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Register - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Services endpoint response - Catalog Node endpoint response - Catalog Service endpoint response - Updated API tests for all of the above too (both deprecated and new forms of register) TODO: - config package changes for on-disk service definitions - proxy config endpoint - built-in proxy support for new fields * Agent proxy config endpoint updated with upstreams * Config file changes for upstreams. * Add upstream opaque config and update all tests to ensure it works everywhere. * Built in proxy working with new Upstreams config * Command fixes and deprecations * Fix key translation, upstream type defaults and a spate of other subtele bugs found with ned to end test scripts... TODO: tests still failing on one case that needs a fix. I think it's key translation for upstreams nested in Managed proxy struct. * Fix translated keys in API registration. ≈ * Fixes from docs - omit some empty undocumented fields in API - Bring back ServiceProxyDestination in Catalog responses to not break backwards compat - this was removed assuming it was only used internally. * Documentation updates for Upstreams in service definition * Fixes for tests broken by many refactors. * Enable travis on f-connect branch in this branch too. * Add consistent Deprecation comments to ProxyDestination uses * Update version number on deprecation notices, and correct upstream datacenter field with explanation in docs
2018-09-12 16:07:47 +00:00
assert.Equal(args.Service.Proxy.DestinationServiceName, v.ServiceProxy.DestinationServiceName)
// List by non-Connect
req = structs.ServiceSpecificRequest{
Datacenter: "dc1",
Add Proxy Upstreams to Service Definition (#4639) * Refactor Service Definition ProxyDestination. This includes: - Refactoring all internal structs used - Updated tests for both deprecated and new input for: - Agent Services endpoint response - Agent Service endpoint response - Agent Register endpoint - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Register - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Services endpoint response - Catalog Node endpoint response - Catalog Service endpoint response - Updated API tests for all of the above too (both deprecated and new forms of register) TODO: - config package changes for on-disk service definitions - proxy config endpoint - built-in proxy support for new fields * Agent proxy config endpoint updated with upstreams * Config file changes for upstreams. * Add upstream opaque config and update all tests to ensure it works everywhere. * Built in proxy working with new Upstreams config * Command fixes and deprecations * Fix key translation, upstream type defaults and a spate of other subtele bugs found with ned to end test scripts... TODO: tests still failing on one case that needs a fix. I think it's key translation for upstreams nested in Managed proxy struct. * Fix translated keys in API registration. ≈ * Fixes from docs - omit some empty undocumented fields in API - Bring back ServiceProxyDestination in Catalog responses to not break backwards compat - this was removed assuming it was only used internally. * Documentation updates for Upstreams in service definition * Fixes for tests broken by many refactors. * Enable travis on f-connect branch in this branch too. * Add consistent Deprecation comments to ProxyDestination uses * Update version number on deprecation notices, and correct upstream datacenter field with explanation in docs
2018-09-12 16:07:47 +00:00
ServiceName: args.Service.Proxy.DestinationServiceName,
}
assert.Nil(msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &req, &resp))
assert.Len(resp.ServiceNodes, 1)
v = resp.ServiceNodes[0]
Add Proxy Upstreams to Service Definition (#4639) * Refactor Service Definition ProxyDestination. This includes: - Refactoring all internal structs used - Updated tests for both deprecated and new input for: - Agent Services endpoint response - Agent Service endpoint response - Agent Register endpoint - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Register - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Services endpoint response - Catalog Node endpoint response - Catalog Service endpoint response - Updated API tests for all of the above too (both deprecated and new forms of register) TODO: - config package changes for on-disk service definitions - proxy config endpoint - built-in proxy support for new fields * Agent proxy config endpoint updated with upstreams * Config file changes for upstreams. * Add upstream opaque config and update all tests to ensure it works everywhere. * Built in proxy working with new Upstreams config * Command fixes and deprecations * Fix key translation, upstream type defaults and a spate of other subtele bugs found with ned to end test scripts... TODO: tests still failing on one case that needs a fix. I think it's key translation for upstreams nested in Managed proxy struct. * Fix translated keys in API registration. ≈ * Fixes from docs - omit some empty undocumented fields in API - Bring back ServiceProxyDestination in Catalog responses to not break backwards compat - this was removed assuming it was only used internally. * Documentation updates for Upstreams in service definition * Fixes for tests broken by many refactors. * Enable travis on f-connect branch in this branch too. * Add consistent Deprecation comments to ProxyDestination uses * Update version number on deprecation notices, and correct upstream datacenter field with explanation in docs
2018-09-12 16:07:47 +00:00
assert.Equal(args.Service.Proxy.DestinationServiceName, v.ServiceName)
assert.Equal("", v.ServiceProxy.DestinationServiceName)
}
// Test that calling ServiceNodes with Connect: true will return
// Connect native services.
func TestCatalog_ListServiceNodes_ConnectDestinationNative(t *testing.T) {
t.Parallel()
require := require.New(t)
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
testrpc.WaitForLeader(t, s1.RPC, "dc1")
// Register the native service
args := structs.TestRegisterRequest(t)
args.Service.Connect.Native = true
var out struct{}
require.Nil(msgpackrpc.CallWithCodec(codec, "Catalog.Register", args, &out))
// List
req := structs.ServiceSpecificRequest{
Connect: true,
Datacenter: "dc1",
ServiceName: args.Service.Service,
}
var resp structs.IndexedServiceNodes
require.Nil(msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &req, &resp))
require.Len(resp.ServiceNodes, 1)
v := resp.ServiceNodes[0]
require.Equal(args.Service.Service, v.ServiceName)
// List by non-Connect
req = structs.ServiceSpecificRequest{
Datacenter: "dc1",
ServiceName: args.Service.Service,
}
require.Nil(msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &req, &resp))
require.Len(resp.ServiceNodes, 1)
v = resp.ServiceNodes[0]
require.Equal(args.Service.Service, v.ServiceName)
}
func TestCatalog_ListServiceNodes_ConnectProxy_ACL(t *testing.T) {
t.Parallel()
dir1, s1 := testServerWithConfig(t, func(c *Config) {
c.ACLDatacenter = "dc1"
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
c.ACLsEnabled = true
c.ACLMasterToken = "root"
c.ACLDefaultPolicy = "deny"
})
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
testrpc.WaitForLeader(t, s1.RPC, "dc1")
// Create the ACL.
arg := structs.ACLRequest{
Datacenter: "dc1",
Op: structs.ACLSet,
ACL: structs.ACL{
Name: "User token",
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
Type: structs.ACLTokenTypeClient,
Rules: `
service "foo" {
policy = "write"
}
node "" { policy = "read" }
`,
},
WriteRequest: structs.WriteRequest{Token: "root"},
}
var token string
require.NoError(t, msgpackrpc.CallWithCodec(codec, "ACL.Apply", &arg, &token))
{
// Register a proxy
args := structs.TestRegisterRequestProxy(t)
args.Service.Service = "foo-proxy"
Add Proxy Upstreams to Service Definition (#4639) * Refactor Service Definition ProxyDestination. This includes: - Refactoring all internal structs used - Updated tests for both deprecated and new input for: - Agent Services endpoint response - Agent Service endpoint response - Agent Register endpoint - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Register - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Services endpoint response - Catalog Node endpoint response - Catalog Service endpoint response - Updated API tests for all of the above too (both deprecated and new forms of register) TODO: - config package changes for on-disk service definitions - proxy config endpoint - built-in proxy support for new fields * Agent proxy config endpoint updated with upstreams * Config file changes for upstreams. * Add upstream opaque config and update all tests to ensure it works everywhere. * Built in proxy working with new Upstreams config * Command fixes and deprecations * Fix key translation, upstream type defaults and a spate of other subtele bugs found with ned to end test scripts... TODO: tests still failing on one case that needs a fix. I think it's key translation for upstreams nested in Managed proxy struct. * Fix translated keys in API registration. ≈ * Fixes from docs - omit some empty undocumented fields in API - Bring back ServiceProxyDestination in Catalog responses to not break backwards compat - this was removed assuming it was only used internally. * Documentation updates for Upstreams in service definition * Fixes for tests broken by many refactors. * Enable travis on f-connect branch in this branch too. * Add consistent Deprecation comments to ProxyDestination uses * Update version number on deprecation notices, and correct upstream datacenter field with explanation in docs
2018-09-12 16:07:47 +00:00
args.Service.Proxy.DestinationServiceName = "bar"
args.WriteRequest.Token = "root"
var out struct{}
require.NoError(t, msgpackrpc.CallWithCodec(codec, "Catalog.Register", &args, &out))
// Register a proxy
args = structs.TestRegisterRequestProxy(t)
args.Service.Service = "foo-proxy"
Add Proxy Upstreams to Service Definition (#4639) * Refactor Service Definition ProxyDestination. This includes: - Refactoring all internal structs used - Updated tests for both deprecated and new input for: - Agent Services endpoint response - Agent Service endpoint response - Agent Register endpoint - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Register - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Services endpoint response - Catalog Node endpoint response - Catalog Service endpoint response - Updated API tests for all of the above too (both deprecated and new forms of register) TODO: - config package changes for on-disk service definitions - proxy config endpoint - built-in proxy support for new fields * Agent proxy config endpoint updated with upstreams * Config file changes for upstreams. * Add upstream opaque config and update all tests to ensure it works everywhere. * Built in proxy working with new Upstreams config * Command fixes and deprecations * Fix key translation, upstream type defaults and a spate of other subtele bugs found with ned to end test scripts... TODO: tests still failing on one case that needs a fix. I think it's key translation for upstreams nested in Managed proxy struct. * Fix translated keys in API registration. ≈ * Fixes from docs - omit some empty undocumented fields in API - Bring back ServiceProxyDestination in Catalog responses to not break backwards compat - this was removed assuming it was only used internally. * Documentation updates for Upstreams in service definition * Fixes for tests broken by many refactors. * Enable travis on f-connect branch in this branch too. * Add consistent Deprecation comments to ProxyDestination uses * Update version number on deprecation notices, and correct upstream datacenter field with explanation in docs
2018-09-12 16:07:47 +00:00
args.Service.Proxy.DestinationServiceName = "foo"
args.WriteRequest.Token = "root"
require.NoError(t, msgpackrpc.CallWithCodec(codec, "Catalog.Register", &args, &out))
// Register a proxy
args = structs.TestRegisterRequestProxy(t)
args.Service.Service = "another-proxy"
Add Proxy Upstreams to Service Definition (#4639) * Refactor Service Definition ProxyDestination. This includes: - Refactoring all internal structs used - Updated tests for both deprecated and new input for: - Agent Services endpoint response - Agent Service endpoint response - Agent Register endpoint - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Register - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Services endpoint response - Catalog Node endpoint response - Catalog Service endpoint response - Updated API tests for all of the above too (both deprecated and new forms of register) TODO: - config package changes for on-disk service definitions - proxy config endpoint - built-in proxy support for new fields * Agent proxy config endpoint updated with upstreams * Config file changes for upstreams. * Add upstream opaque config and update all tests to ensure it works everywhere. * Built in proxy working with new Upstreams config * Command fixes and deprecations * Fix key translation, upstream type defaults and a spate of other subtele bugs found with ned to end test scripts... TODO: tests still failing on one case that needs a fix. I think it's key translation for upstreams nested in Managed proxy struct. * Fix translated keys in API registration. ≈ * Fixes from docs - omit some empty undocumented fields in API - Bring back ServiceProxyDestination in Catalog responses to not break backwards compat - this was removed assuming it was only used internally. * Documentation updates for Upstreams in service definition * Fixes for tests broken by many refactors. * Enable travis on f-connect branch in this branch too. * Add consistent Deprecation comments to ProxyDestination uses * Update version number on deprecation notices, and correct upstream datacenter field with explanation in docs
2018-09-12 16:07:47 +00:00
args.Service.Proxy.DestinationServiceName = "foo"
args.WriteRequest.Token = "root"
require.NoError(t, msgpackrpc.CallWithCodec(codec, "Catalog.Register", &args, &out))
}
// List w/ token. This should disallow because we don't have permission
// to read "bar"
req := structs.ServiceSpecificRequest{
Connect: true,
Datacenter: "dc1",
ServiceName: "bar",
QueryOptions: structs.QueryOptions{Token: token},
}
var resp structs.IndexedServiceNodes
require.NoError(t, msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &req, &resp))
require.Len(t, resp.ServiceNodes, 0)
// List w/ token. This should work since we're requesting "foo", but should
// also only contain the proxies with names that adhere to our ACL.
req = structs.ServiceSpecificRequest{
Connect: true,
Datacenter: "dc1",
ServiceName: "foo",
QueryOptions: structs.QueryOptions{Token: token},
}
resp = structs.IndexedServiceNodes{}
require.NoError(t, msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &req, &resp))
require.Len(t, resp.ServiceNodes, 1)
v := resp.ServiceNodes[0]
require.Equal(t, "foo-proxy", v.ServiceName)
}
func TestCatalog_ListServiceNodes_ConnectNative(t *testing.T) {
t.Parallel()
assert := assert.New(t)
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
testrpc.WaitForLeader(t, s1.RPC, "dc1")
// Register the service
args := structs.TestRegisterRequest(t)
args.Service.Connect.Native = true
var out struct{}
assert.Nil(msgpackrpc.CallWithCodec(codec, "Catalog.Register", args, &out))
// List
req := structs.ServiceSpecificRequest{
Datacenter: "dc1",
ServiceName: args.Service.Service,
TagFilter: false,
}
var resp structs.IndexedServiceNodes
assert.Nil(msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &req, &resp))
assert.Len(resp.ServiceNodes, 1)
v := resp.ServiceNodes[0]
assert.Equal(args.Service.Connect.Native, v.ServiceConnect.Native)
}
func TestCatalog_NodeServices(t *testing.T) {
t.Parallel()
2013-12-12 19:46:25 +00:00
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
testrpc.WaitForTestAgent(t, s1.RPC, "dc1")
2013-12-12 19:46:25 +00:00
2014-01-08 21:52:09 +00:00
args := structs.NodeSpecificRequest{
2013-12-12 19:46:25 +00:00
Datacenter: "dc1",
Node: "foo",
}
var out structs.IndexedNodeServices
err := msgpackrpc.CallWithCodec(codec, "Catalog.NodeServices", &args, &out)
if err != nil {
2013-12-12 19:46:25 +00:00
t.Fatalf("err: %v", err)
}
testrpc.WaitForLeader(t, s1.RPC, "dc1")
2013-12-12 19:46:25 +00:00
// Just add a node
if err := s1.fsm.State().EnsureNode(1, &structs.Node{Node: "foo", Address: "127.0.0.1"}); err != nil {
t.Fatalf("err: %v", err)
}
if err := s1.fsm.State().EnsureService(2, "foo", &structs.NodeService{ID: "db", Service: "db", Tags: []string{"primary"}, Address: "127.0.0.1", Port: 5000}); err != nil {
t.Fatalf("err: %v", err)
}
if err := s1.fsm.State().EnsureService(3, "foo", &structs.NodeService{ID: "web", Service: "web", Tags: nil, Address: "127.0.0.1", Port: 80}); err != nil {
t.Fatalf("err: %v", err)
}
2013-12-12 19:46:25 +00:00
if err := msgpackrpc.CallWithCodec(codec, "Catalog.NodeServices", &args, &out); err != nil {
2013-12-12 19:46:25 +00:00
t.Fatalf("err: %v", err)
}
if out.NodeServices.Node.Address != "127.0.0.1" {
t.Fatalf("bad: %v", out)
}
if len(out.NodeServices.Services) != 2 {
2013-12-12 19:46:25 +00:00
t.Fatalf("bad: %v", out)
}
services := out.NodeServices.Services
2020-05-27 16:47:32 +00:00
if !stringslice.Contains(services["db"].Tags, "primary") || services["db"].Port != 5000 {
2013-12-12 19:46:25 +00:00
t.Fatalf("bad: %v", out)
}
if len(services["web"].Tags) != 0 || services["web"].Port != 80 {
2013-12-12 19:46:25 +00:00
t.Fatalf("bad: %v", out)
}
}
2014-01-01 02:31:17 +00:00
func TestCatalog_NodeServices_ConnectProxy(t *testing.T) {
t.Parallel()
assert := assert.New(t)
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
testrpc.WaitForTestAgent(t, s1.RPC, "dc1")
// Register the service
args := structs.TestRegisterRequestProxy(t)
var out struct{}
assert.Nil(msgpackrpc.CallWithCodec(codec, "Catalog.Register", args, &out))
// List
req := structs.NodeSpecificRequest{
Datacenter: "dc1",
Node: args.Node,
}
var resp structs.IndexedNodeServices
assert.Nil(msgpackrpc.CallWithCodec(codec, "Catalog.NodeServices", &req, &resp))
assert.Len(resp.NodeServices.Services, 1)
v := resp.NodeServices.Services[args.Service.Service]
assert.Equal(structs.ServiceKindConnectProxy, v.Kind)
Add Proxy Upstreams to Service Definition (#4639) * Refactor Service Definition ProxyDestination. This includes: - Refactoring all internal structs used - Updated tests for both deprecated and new input for: - Agent Services endpoint response - Agent Service endpoint response - Agent Register endpoint - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Register - Unmanaged deprecated field - Unmanaged new fields - Managed deprecated upstreams - Managed new - Catalog Services endpoint response - Catalog Node endpoint response - Catalog Service endpoint response - Updated API tests for all of the above too (both deprecated and new forms of register) TODO: - config package changes for on-disk service definitions - proxy config endpoint - built-in proxy support for new fields * Agent proxy config endpoint updated with upstreams * Config file changes for upstreams. * Add upstream opaque config and update all tests to ensure it works everywhere. * Built in proxy working with new Upstreams config * Command fixes and deprecations * Fix key translation, upstream type defaults and a spate of other subtele bugs found with ned to end test scripts... TODO: tests still failing on one case that needs a fix. I think it's key translation for upstreams nested in Managed proxy struct. * Fix translated keys in API registration. ≈ * Fixes from docs - omit some empty undocumented fields in API - Bring back ServiceProxyDestination in Catalog responses to not break backwards compat - this was removed assuming it was only used internally. * Documentation updates for Upstreams in service definition * Fixes for tests broken by many refactors. * Enable travis on f-connect branch in this branch too. * Add consistent Deprecation comments to ProxyDestination uses * Update version number on deprecation notices, and correct upstream datacenter field with explanation in docs
2018-09-12 16:07:47 +00:00
assert.Equal(args.Service.Proxy.DestinationServiceName, v.Proxy.DestinationServiceName)
}
func TestCatalog_NodeServices_ConnectNative(t *testing.T) {
t.Parallel()
assert := assert.New(t)
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
testrpc.WaitForTestAgent(t, s1.RPC, "dc1")
// Register the service
args := structs.TestRegisterRequest(t)
var out struct{}
assert.Nil(msgpackrpc.CallWithCodec(codec, "Catalog.Register", args, &out))
// List
req := structs.NodeSpecificRequest{
Datacenter: "dc1",
Node: args.Node,
}
var resp structs.IndexedNodeServices
assert.Nil(msgpackrpc.CallWithCodec(codec, "Catalog.NodeServices", &req, &resp))
assert.Len(resp.NodeServices.Services, 1)
v := resp.NodeServices.Services[args.Service.Service]
assert.Equal(args.Service.Connect.Native, v.Connect.Native)
}
2014-01-01 02:31:17 +00:00
// Used to check for a regression against a known bug
func TestCatalog_Register_FailedCase1(t *testing.T) {
t.Parallel()
2014-01-01 02:31:17 +00:00
dir1, s1 := testServer(t)
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
2014-01-01 02:31:17 +00:00
arg := structs.RegisterRequest{
Datacenter: "dc1",
Node: "bar",
Address: "127.0.0.2",
Service: &structs.NodeService{
Service: "web",
2014-04-03 19:03:10 +00:00
Tags: nil,
Port: 8000,
},
2014-01-01 02:31:17 +00:00
}
var out struct{}
err := msgpackrpc.CallWithCodec(codec, "Catalog.Register", &arg, &out)
if err != nil {
2014-01-01 02:31:17 +00:00
t.Fatalf("err: %v", err)
}
testrpc.WaitForLeader(t, s1.RPC, "dc1")
2014-01-01 02:31:17 +00:00
if err := msgpackrpc.CallWithCodec(codec, "Catalog.Register", &arg, &out); err != nil {
2014-01-01 02:31:17 +00:00
t.Fatalf("err: %v", err)
}
// Check we can get this back
2014-01-08 21:52:09 +00:00
query := &structs.ServiceSpecificRequest{
2014-01-01 02:31:17 +00:00
Datacenter: "dc1",
ServiceName: "web",
}
var out2 structs.IndexedServiceNodes
if err := msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", query, &out2); err != nil {
2014-01-01 02:31:17 +00:00
t.Fatalf("err: %v", err)
}
// Check the output
if len(out2.ServiceNodes) != 1 {
t.Fatalf("Bad: %v", out2)
2014-01-01 02:31:17 +00:00
}
}
func testACLFilterServer(t *testing.T) (dir, token string, srv *Server, codec rpc.ClientCodec) {
2015-06-11 20:23:49 +00:00
dir, srv = testServerWithConfig(t, func(c *Config) {
c.ACLDatacenter = "dc1"
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
c.ACLsEnabled = true
c.ACLMasterToken = "root"
c.ACLDefaultPolicy = "deny"
})
codec = rpcClient(t, srv)
testrpc.WaitForTestAgent(t, srv.RPC, "dc1", testrpc.WithToken("root"))
// Create a new token
arg := structs.ACLRequest{
Datacenter: "dc1",
Op: structs.ACLSet,
ACL: structs.ACL{
Name: "User token",
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
Type: structs.ACLTokenTypeClient,
Rules: `
service "foo" {
policy = "write"
}
node "" {
policy = "read"
}
`,
},
WriteRequest: structs.WriteRequest{Token: "root"},
}
if err := msgpackrpc.CallWithCodec(codec, "ACL.Apply", &arg, &token); err != nil {
t.Fatalf("err: %v", err)
}
// Register a service
regArg := structs.RegisterRequest{
Datacenter: "dc1",
Node: srv.config.NodeName,
Address: "127.0.0.1",
Service: &structs.NodeService{
ID: "foo",
Service: "foo",
},
Check: &structs.HealthCheck{
CheckID: "service:foo",
Name: "service:foo",
ServiceID: "foo",
Status: api.HealthPassing,
},
WriteRequest: structs.WriteRequest{Token: "root"},
}
if err := msgpackrpc.CallWithCodec(codec, "Catalog.Register", &regArg, nil); err != nil {
t.Fatalf("err: %s", err)
}
// Register a service which should be denied
regArg = structs.RegisterRequest{
Datacenter: "dc1",
Node: srv.config.NodeName,
Address: "127.0.0.1",
Service: &structs.NodeService{
ID: "bar",
Service: "bar",
},
Check: &structs.HealthCheck{
CheckID: "service:bar",
Name: "service:bar",
ServiceID: "bar",
},
WriteRequest: structs.WriteRequest{Token: "root"},
}
if err := msgpackrpc.CallWithCodec(codec, "Catalog.Register", &regArg, nil); err != nil {
t.Fatalf("err: %s", err)
}
2015-06-11 20:23:49 +00:00
return
}
2015-06-11 20:23:49 +00:00
func TestCatalog_ListServices_FilterACL(t *testing.T) {
t.Parallel()
dir, token, srv, codec := testACLFilterServer(t)
2015-06-11 20:23:49 +00:00
defer os.RemoveAll(dir)
defer srv.Shutdown()
defer codec.Close()
testrpc.WaitForTestAgent(t, srv.RPC, "dc1", testrpc.WithToken("root"))
2015-06-11 20:23:49 +00:00
opt := structs.DCSpecificRequest{
Datacenter: "dc1",
QueryOptions: structs.QueryOptions{Token: token},
}
2015-06-11 20:23:49 +00:00
reply := structs.IndexedServices{}
if err := msgpackrpc.CallWithCodec(codec, "Catalog.ListServices", &opt, &reply); err != nil {
2015-06-11 20:23:49 +00:00
t.Fatalf("err: %s", err)
}
if _, ok := reply.Services["foo"]; !ok {
t.Fatalf("bad: %#v", reply.Services)
}
if _, ok := reply.Services["bar"]; ok {
t.Fatalf("bad: %#v", reply.Services)
}
}
2015-06-11 20:23:49 +00:00
func TestCatalog_ServiceNodes_FilterACL(t *testing.T) {
t.Parallel()
dir, token, srv, codec := testACLFilterServer(t)
2015-06-11 20:23:49 +00:00
defer os.RemoveAll(dir)
defer srv.Shutdown()
defer codec.Close()
2015-06-11 20:23:49 +00:00
opt := structs.ServiceSpecificRequest{
Datacenter: "dc1",
ServiceName: "foo",
QueryOptions: structs.QueryOptions{Token: token},
}
reply := structs.IndexedServiceNodes{}
if err := msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &opt, &reply); err != nil {
2015-06-11 20:23:49 +00:00
t.Fatalf("err: %s", err)
}
found := false
for _, sn := range reply.ServiceNodes {
if sn.ServiceID == "foo" {
found = true
break
}
}
2015-06-11 20:23:49 +00:00
if !found {
t.Fatalf("bad: %#v", reply.ServiceNodes)
}
2015-06-11 20:23:49 +00:00
// Filters services we can't access
opt = structs.ServiceSpecificRequest{
Datacenter: "dc1",
ServiceName: "bar",
QueryOptions: structs.QueryOptions{Token: token},
}
reply = structs.IndexedServiceNodes{}
if err := msgpackrpc.CallWithCodec(codec, "Catalog.ServiceNodes", &opt, &reply); err != nil {
2015-06-11 20:23:49 +00:00
t.Fatalf("err: %s", err)
}
for _, sn := range reply.ServiceNodes {
if sn.ServiceID == "bar" {
t.Fatalf("bad: %#v", reply.ServiceNodes)
}
}
// We've already proven that we call the ACL filtering function so we
// test node filtering down in acl.go for node cases. This also proves
// that we respect the version 8 ACL flag, since the test server sets
// that to false (the regression value of *not* changing this is better
// for now until we change the sense of the version 8 ACL flag).
2015-06-11 20:23:49 +00:00
}
func TestCatalog_NodeServices_ACLDeny(t *testing.T) {
t.Parallel()
dir1, s1 := testServerWithConfig(t, func(c *Config) {
c.ACLDatacenter = "dc1"
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
c.ACLsEnabled = true
c.ACLMasterToken = "root"
c.ACLDefaultPolicy = "deny"
})
defer os.RemoveAll(dir1)
defer s1.Shutdown()
codec := rpcClient(t, s1)
defer codec.Close()
testrpc.WaitForTestAgent(t, s1.RPC, "dc1", testrpc.WithToken("root"))
// The node policy should not be ignored.
args := structs.NodeSpecificRequest{
Datacenter: "dc1",
Node: s1.config.NodeName,
}
reply := structs.IndexedNodeServices{}
if err := msgpackrpc.CallWithCodec(codec, "Catalog.NodeServices", &args, &reply); err != nil {
t.Fatalf("err: %v", err)
}
if reply.NodeServices != nil {
t.Fatalf("should not nil")
}
// Create an ACL that can read the node.
arg := structs.ACLRequest{
Datacenter: "dc1",
Op: structs.ACLSet,
ACL: structs.ACL{
Name: "User token",
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
Type: structs.ACLTokenTypeClient,
Rules: fmt.Sprintf(`
node "%s" {
policy = "read"
}
`, s1.config.NodeName),
},
WriteRequest: structs.WriteRequest{Token: "root"},
}
var id string
if err := msgpackrpc.CallWithCodec(codec, "ACL.Apply", &arg, &id); err != nil {
t.Fatalf("err: %v", err)
}
// Now try with the token and it will go through.
args.Token = id
if err := msgpackrpc.CallWithCodec(codec, "Catalog.NodeServices", &args, &reply); err != nil {
t.Fatalf("err: %v", err)
}
if reply.NodeServices == nil {
t.Fatalf("should not be nil")
}
// Make sure an unknown node doesn't cause trouble.
args.Node = "nope"
if err := msgpackrpc.CallWithCodec(codec, "Catalog.NodeServices", &args, &reply); err != nil {
t.Fatalf("err: %v", err)
}
if reply.NodeServices != nil {
t.Fatalf("should not nil")
}
}
2015-06-11 20:23:49 +00:00
func TestCatalog_NodeServices_FilterACL(t *testing.T) {
t.Parallel()
dir, token, srv, codec := testACLFilterServer(t)
2015-06-11 20:23:49 +00:00
defer os.RemoveAll(dir)
defer srv.Shutdown()
defer codec.Close()
testrpc.WaitForTestAgent(t, srv.RPC, "dc1", testrpc.WithToken("root"))
2015-06-11 20:23:49 +00:00
opt := structs.NodeSpecificRequest{
Datacenter: "dc1",
Node: srv.config.NodeName,
QueryOptions: structs.QueryOptions{Token: token},
}
var reply structs.IndexedNodeServices
require.NoError(t, msgpackrpc.CallWithCodec(codec, "Catalog.NodeServices", &opt, &reply))
require.NotNil(t, reply.NodeServices)
require.Len(t, reply.NodeServices.Services, 1)
svc, ok := reply.NodeServices.Services["foo"]
require.True(t, ok)
require.Equal(t, "foo", svc.ID)
}