a7fb26f50f
This is like a Möbius strip of code due to the fact that low-level components (serf/memberlist) are connected to high-level components (the catalog and mesh-gateways) in a twisty maze of references which make it hard to dive into. With that in mind here's a high level summary of what you'll find in the patch:
There are several distinct chunks of code that are affected:
* new flags and config options for the server
* retry join WAN is slightly different
* retry join code is shared to discover primary mesh gateways from secondary datacenters
* because retry join logic runs in the *agent* and the results of that
operation for primary mesh gateways are needed in the *server* there are
some methods like `RefreshPrimaryGatewayFallbackAddresses` that must occur
at multiple layers of abstraction just to pass the data down to the right
layer.
* new cache type `FederationStateListMeshGatewaysName` for use in `proxycfg/xds` layers
* the function signature for RPC dialing picked up a new required field (the
node name of the destination)
* several new RPCs for manipulating a FederationState object:
`FederationState:{Apply,Get,List,ListMeshGateways}`
* 3 read-only internal APIs for debugging use to invoke those RPCs from curl
* raft and fsm changes to persist these FederationStates
* replication for FederationStates as they are canonically stored in the
Primary and replicated to the Secondaries.
* a special derivative of anti-entropy that runs in secondaries to snapshot
their local mesh gateway `CheckServiceNodes` and sync them into their upstream
FederationState in the primary (this works in conjunction with the
replication to distribute addresses for all mesh gateways in all DCs to all
other DCs)
* a "gateway locator" convenience object to make use of this data to choose
the addresses of gateways to use for any given RPC or gossip operation to a
remote DC. This gets data from the "retry join" logic in the agent and also
directly calls into the FSM.
* RPC (`:8300`) on the server sniffs the first byte of a new connection to
determine if it's actually doing native TLS. If so it checks the ALPN header
for protocol determination (just like how the existing system uses the
type-byte marker).
* 2 new kinds of protocols are exclusively decoded via this native TLS
mechanism: one for ferrying "packet" operations (udp-like) from the gossip
layer and one for "stream" operations (tcp-like). The packet operations
re-use sockets (using length-prefixing) to cut down on TLS re-negotiation
overhead.
* the server instances specially wrap the `memberlist.NetTransport` when running
with gateway federation enabled (in a `wanfed.Transport`). The general gist is
that if it tries to dial a node in the SAME datacenter (deduced by looking
at the suffix of the node name) there is no change. If dialing a DIFFERENT
datacenter it is wrapped up in a TLS+ALPN blob and sent through some mesh
gateways to eventually end up in a server's :8300 port.
* a new flag when launching a mesh gateway via `consul connect envoy` to
indicate that the servers are to be exposed. This sets a special service
meta when registering the gateway into the catalog.
* `proxycfg/xds` notice this metadata blob to activate additional watches for
the FederationState objects as well as the location of all of the consul
servers in that datacenter.
* `xds:` if the extra metadata is in place additional clusters are defined in a
DC to bulk sink all traffic to another DC's gateways. For the current
datacenter we listen on a wildcard name (`server.<dc>.consul`) that load
balances all servers as well as one mini-cluster per node
(`<node>.server.<dc>.consul`)
* the `consul tls cert create` command got a new flag (`-node`) to help create
an additional SAN in certs that can be used with this flavor of federation.
352 lines
9 KiB
Go
352 lines
9 KiB
Go
package consul
|
|
|
|
import (
|
|
"os"
|
|
"testing"
|
|
"time"
|
|
|
|
"github.com/hashicorp/consul/agent/structs"
|
|
"github.com/hashicorp/consul/agent/token"
|
|
"github.com/hashicorp/consul/api"
|
|
"github.com/hashicorp/consul/sdk/testutil/retry"
|
|
"github.com/hashicorp/consul/testrpc"
|
|
"github.com/stretchr/testify/require"
|
|
)
|
|
|
|
func TestLeader_FederationStateAntiEntropy_BlockingQuery(t *testing.T) {
|
|
t.Parallel()
|
|
|
|
dir1, s1 := testServerWithConfig(t, func(c *Config) {
|
|
c.PrimaryDatacenter = "dc1"
|
|
})
|
|
defer os.RemoveAll(dir1)
|
|
defer s1.Shutdown()
|
|
testrpc.WaitForLeader(t, s1.RPC, "dc1")
|
|
|
|
dir2, s2 := testServerWithConfig(t, func(c *Config) {
|
|
c.Datacenter = "dc2"
|
|
c.PrimaryDatacenter = "dc1"
|
|
c.FederationStateReplicationRate = 100
|
|
c.FederationStateReplicationBurst = 100
|
|
c.FederationStateReplicationApplyLimit = 1000000
|
|
c.DisableFederationStateAntiEntropy = true
|
|
})
|
|
testrpc.WaitForLeader(t, s2.RPC, "dc2")
|
|
defer os.RemoveAll(dir2)
|
|
defer s2.Shutdown()
|
|
|
|
// Try to join.
|
|
joinWAN(t, s2, s1)
|
|
testrpc.WaitForLeader(t, s1.RPC, "dc1")
|
|
testrpc.WaitForLeader(t, s1.RPC, "dc2")
|
|
|
|
checkSame := func(t *testing.T, expectN, expectGatewaysInDC2 int) {
|
|
t.Helper()
|
|
retry.Run(t, func(r *retry.R) {
|
|
_, remote, err := s1.fsm.State().FederationStateList(nil)
|
|
require.NoError(r, err)
|
|
require.Len(r, remote, expectN)
|
|
|
|
_, local, err := s2.fsm.State().FederationStateList(nil)
|
|
require.NoError(r, err)
|
|
require.Len(r, local, expectN)
|
|
|
|
var fs2 *structs.FederationState
|
|
for _, fs := range local {
|
|
if fs.Datacenter == "dc2" {
|
|
fs2 = fs
|
|
break
|
|
}
|
|
}
|
|
if expectGatewaysInDC2 < 0 {
|
|
require.Nil(r, fs2)
|
|
} else {
|
|
require.NotNil(r, fs2)
|
|
require.Len(r, fs2.MeshGateways, expectGatewaysInDC2)
|
|
}
|
|
})
|
|
}
|
|
|
|
gatewayCSN1 := newTestMeshGatewayNode(
|
|
"dc2", "gateway1", "1.2.3.4", 443, map[string]string{structs.MetaWANFederationKey: "1"}, api.HealthPassing,
|
|
)
|
|
gatewayCSN2 := newTestMeshGatewayNode(
|
|
"dc2", "gateway2", "4.3.2.1", 443, map[string]string{structs.MetaWANFederationKey: "1"}, api.HealthPassing,
|
|
)
|
|
|
|
// populate with some stuff
|
|
makeFedState := func(t *testing.T, dc string, csn ...structs.CheckServiceNode) {
|
|
t.Helper()
|
|
arg := structs.FederationStateRequest{
|
|
Datacenter: "dc1",
|
|
Op: structs.FederationStateUpsert,
|
|
State: &structs.FederationState{
|
|
Datacenter: dc,
|
|
MeshGateways: csn,
|
|
UpdatedAt: time.Now().UTC(),
|
|
},
|
|
}
|
|
|
|
out := false
|
|
require.NoError(t, s1.RPC("FederationState.Apply", &arg, &out))
|
|
}
|
|
|
|
makeGateways := func(t *testing.T, csn structs.CheckServiceNode) {
|
|
t.Helper()
|
|
const dc = "dc2"
|
|
|
|
arg := structs.RegisterRequest{
|
|
Datacenter: csn.Node.Datacenter,
|
|
Node: csn.Node.Node,
|
|
Address: csn.Node.Address,
|
|
Service: csn.Service,
|
|
Checks: csn.Checks,
|
|
}
|
|
var out struct{}
|
|
require.NoError(t, s2.RPC("Catalog.Register", &arg, &out))
|
|
}
|
|
|
|
type result struct {
|
|
idx uint64
|
|
prev, curr *structs.FederationState
|
|
err error
|
|
}
|
|
|
|
blockAgain := func(last uint64) <-chan result {
|
|
ch := make(chan result, 1)
|
|
go func() {
|
|
var res result
|
|
res.idx, res.prev, res.curr, res.err = s2.fetchFederationStateAntiEntropyDetails(&structs.QueryOptions{
|
|
MinQueryIndex: last,
|
|
RequireConsistent: true,
|
|
})
|
|
ch <- res
|
|
}()
|
|
return ch
|
|
}
|
|
|
|
// wait for the primary to do one round of AE and replicate it
|
|
checkSame(t, 1, -1)
|
|
|
|
// // wait for change to be reflected as well
|
|
// makeFedState(t, "dc2")
|
|
// checkSame(t, 1)
|
|
|
|
// Do the initial fetch (len0 local gateways, upstream has nil fedstate)
|
|
res0 := <-blockAgain(0)
|
|
require.NoError(t, res0.err)
|
|
|
|
ch := blockAgain(res0.idx)
|
|
|
|
// bump the local mesh gateways; should unblock query
|
|
makeGateways(t, gatewayCSN1)
|
|
|
|
res1 := <-ch
|
|
require.NoError(t, res1.err)
|
|
require.NotEqual(t, res1.idx, res0.idx)
|
|
require.Nil(t, res1.prev)
|
|
require.Len(t, res1.curr.MeshGateways, 1)
|
|
|
|
checkSame(t, 1, -1) // no fed state update yet
|
|
|
|
ch = blockAgain(res1.idx)
|
|
|
|
// do manual AE
|
|
makeFedState(t, "dc2", gatewayCSN1)
|
|
|
|
res2 := <-ch
|
|
require.NoError(t, res2.err)
|
|
require.NotEqual(t, res2.idx, res1.idx)
|
|
require.Len(t, res2.prev.MeshGateways, 1)
|
|
require.Len(t, res2.curr.MeshGateways, 1)
|
|
|
|
checkSame(t, 2, 1)
|
|
|
|
ch = blockAgain(res2.idx)
|
|
|
|
// add another local mesh gateway
|
|
makeGateways(t, gatewayCSN2)
|
|
|
|
res3 := <-ch
|
|
require.NoError(t, res3.err)
|
|
require.NotEqual(t, res3.idx, res2.idx)
|
|
require.Len(t, res3.prev.MeshGateways, 1)
|
|
require.Len(t, res3.curr.MeshGateways, 2)
|
|
|
|
checkSame(t, 2, 1)
|
|
|
|
ch = blockAgain(res3.idx)
|
|
|
|
// do manual AE
|
|
makeFedState(t, "dc2", gatewayCSN1, gatewayCSN2)
|
|
|
|
res4 := <-ch
|
|
require.NoError(t, res4.err)
|
|
require.NotEqual(t, res4.idx, res3.idx)
|
|
require.Len(t, res4.prev.MeshGateways, 2)
|
|
require.Len(t, res4.curr.MeshGateways, 2)
|
|
|
|
checkSame(t, 2, 2)
|
|
}
|
|
|
|
func TestLeader_FederationStateAntiEntropyPruning(t *testing.T) {
|
|
t.Parallel()
|
|
|
|
dir1, s1 := testServerWithConfig(t, func(c *Config) {
|
|
c.PrimaryDatacenter = "dc1"
|
|
})
|
|
defer os.RemoveAll(dir1)
|
|
defer s1.Shutdown()
|
|
testrpc.WaitForLeader(t, s1.RPC, "dc1")
|
|
client := rpcClient(t, s1)
|
|
defer client.Close()
|
|
|
|
dir2, s2 := testServerWithConfig(t, func(c *Config) {
|
|
c.Datacenter = "dc2"
|
|
c.PrimaryDatacenter = "dc1"
|
|
})
|
|
testrpc.WaitForLeader(t, s2.RPC, "dc2")
|
|
defer os.RemoveAll(dir2)
|
|
defer s2.Shutdown()
|
|
|
|
// Try to join.
|
|
joinWAN(t, s2, s1)
|
|
testrpc.WaitForLeader(t, s1.RPC, "dc1")
|
|
testrpc.WaitForLeader(t, s1.RPC, "dc2")
|
|
|
|
checkSame := func(r *retry.R) error {
|
|
_, remote, err := s1.fsm.State().FederationStateList(nil)
|
|
require.NoError(r, err)
|
|
_, local, err := s2.fsm.State().FederationStateList(nil)
|
|
require.NoError(r, err)
|
|
|
|
require.Len(r, remote, 2)
|
|
require.Len(r, local, 2)
|
|
for i, _ := range remote {
|
|
// zero out the raft data for future comparisons
|
|
remote[i].RaftIndex = structs.RaftIndex{}
|
|
local[i].RaftIndex = structs.RaftIndex{}
|
|
require.Equal(r, remote[i], local[i])
|
|
}
|
|
return nil
|
|
}
|
|
|
|
// Wait for the replica to converge.
|
|
retry.Run(t, func(r *retry.R) {
|
|
checkSame(r)
|
|
})
|
|
|
|
// Now leave and shutdown dc2.
|
|
require.NoError(t, s2.Leave())
|
|
require.NoError(t, s2.Shutdown())
|
|
|
|
// Wait until we know the router is updated.
|
|
retry.Run(t, func(r *retry.R) {
|
|
dcs := s1.router.GetDatacenters()
|
|
require.Len(r, dcs, 1)
|
|
require.Equal(r, "dc1", dcs[0])
|
|
})
|
|
|
|
// Since the background routine is going to run every hour, it likely is
|
|
// not going to run during this test, so it's safe to directly invoke the
|
|
// core method.
|
|
require.NoError(t, s1.pruneStaleFederationStates())
|
|
|
|
// Wait for dc2 to drop out.
|
|
retry.Run(t, func(r *retry.R) {
|
|
_, mine, err := s1.fsm.State().FederationStateList(nil)
|
|
require.NoError(r, err)
|
|
|
|
require.Len(r, mine, 1)
|
|
require.Equal(r, "dc1", mine[0].Datacenter)
|
|
})
|
|
}
|
|
|
|
func TestLeader_FederationStateAntiEntropyPruning_ACLDeny(t *testing.T) {
|
|
t.Parallel()
|
|
|
|
dir1, s1 := testServerWithConfig(t, func(c *Config) {
|
|
c.PrimaryDatacenter = "dc1"
|
|
c.ACLDatacenter = "dc1"
|
|
c.ACLsEnabled = true
|
|
c.ACLMasterToken = "root"
|
|
c.ACLDefaultPolicy = "deny"
|
|
})
|
|
defer os.RemoveAll(dir1)
|
|
defer s1.Shutdown()
|
|
testrpc.WaitForLeader(t, s1.RPC, "dc1")
|
|
client := rpcClient(t, s1)
|
|
defer client.Close()
|
|
|
|
dir2, s2 := testServerWithConfig(t, func(c *Config) {
|
|
c.Datacenter = "dc2"
|
|
c.PrimaryDatacenter = "dc1"
|
|
c.ACLDatacenter = "dc1"
|
|
c.ACLsEnabled = true
|
|
c.ACLMasterToken = "root"
|
|
c.ACLDefaultPolicy = "deny"
|
|
})
|
|
testrpc.WaitForLeader(t, s2.RPC, "dc2")
|
|
defer os.RemoveAll(dir2)
|
|
defer s2.Shutdown()
|
|
|
|
// Try to join.
|
|
joinWAN(t, s2, s1)
|
|
testrpc.WaitForLeader(t, s1.RPC, "dc1")
|
|
testrpc.WaitForLeader(t, s1.RPC, "dc2")
|
|
|
|
// Create the ACL token.
|
|
opWriteToken, err := upsertTestTokenWithPolicyRules(client, "root", "dc1", `operator = "write"`)
|
|
require.NoError(t, err)
|
|
|
|
require.True(t, s1.tokens.UpdateReplicationToken(opWriteToken.SecretID, token.TokenSourceAPI))
|
|
require.True(t, s2.tokens.UpdateReplicationToken(opWriteToken.SecretID, token.TokenSourceAPI))
|
|
|
|
checkSame := func(r *retry.R) error {
|
|
_, remote, err := s1.fsm.State().FederationStateList(nil)
|
|
require.NoError(r, err)
|
|
_, local, err := s2.fsm.State().FederationStateList(nil)
|
|
require.NoError(r, err)
|
|
|
|
require.Len(r, remote, 2)
|
|
require.Len(r, local, 2)
|
|
for i, _ := range remote {
|
|
// zero out the raft data for future comparisons
|
|
remote[i].RaftIndex = structs.RaftIndex{}
|
|
local[i].RaftIndex = structs.RaftIndex{}
|
|
require.Equal(r, remote[i], local[i])
|
|
}
|
|
return nil
|
|
}
|
|
|
|
// Wait for the replica to converge.
|
|
retry.Run(t, func(r *retry.R) {
|
|
checkSame(r)
|
|
})
|
|
|
|
// Now leave and shutdown dc2.
|
|
require.NoError(t, s2.Leave())
|
|
require.NoError(t, s2.Shutdown())
|
|
|
|
// Wait until we know the router is updated.
|
|
retry.Run(t, func(r *retry.R) {
|
|
dcs := s1.router.GetDatacenters()
|
|
require.Len(r, dcs, 1)
|
|
require.Equal(r, "dc1", dcs[0])
|
|
})
|
|
|
|
// Since the background routine is going to run every hour, it likely is
|
|
// not going to run during this test, so it's safe to directly invoke the
|
|
// core method.
|
|
require.NoError(t, s1.pruneStaleFederationStates())
|
|
|
|
// Wait for dc2 to drop out.
|
|
retry.Run(t, func(r *retry.R) {
|
|
_, mine, err := s1.fsm.State().FederationStateList(nil)
|
|
require.NoError(r, err)
|
|
|
|
require.Len(r, mine, 1)
|
|
require.Equal(r, "dc1", mine[0].Datacenter)
|
|
})
|
|
}
|