consul/agent/proxycfg_test.go

144 lines
3.9 KiB
Go
Raw Normal View History

// Copyright (c) HashiCorp, Inc.
[COMPLIANCE] License changes (#18443) * Adding explicit MPL license for sub-package This directory and its subdirectories (packages) contain files licensed with the MPLv2 `LICENSE` file in this directory and are intentionally licensed separately from the BSL `LICENSE` file at the root of this repository. * Adding explicit MPL license for sub-package This directory and its subdirectories (packages) contain files licensed with the MPLv2 `LICENSE` file in this directory and are intentionally licensed separately from the BSL `LICENSE` file at the root of this repository. * Updating the license from MPL to Business Source License Going forward, this project will be licensed under the Business Source License v1.1. Please see our blog post for more details at <Blog URL>, FAQ at www.hashicorp.com/licensing-faq, and details of the license at www.hashicorp.com/bsl. * add missing license headers * Update copyright file headers to BUSL-1.1 * Update copyright file headers to BUSL-1.1 * Update copyright file headers to BUSL-1.1 * Update copyright file headers to BUSL-1.1 * Update copyright file headers to BUSL-1.1 * Update copyright file headers to BUSL-1.1 * Update copyright file headers to BUSL-1.1 * Update copyright file headers to BUSL-1.1 * Update copyright file headers to BUSL-1.1 * Update copyright file headers to BUSL-1.1 * Update copyright file headers to BUSL-1.1 * Update copyright file headers to BUSL-1.1 * Update copyright file headers to BUSL-1.1 * Update copyright file headers to BUSL-1.1 * Update copyright file headers to BUSL-1.1 --------- Co-authored-by: hashicorp-copywrite[bot] <110428419+hashicorp-copywrite[bot]@users.noreply.github.com>
2023-08-11 13:12:13 +00:00
// SPDX-License-Identifier: BUSL-1.1
proxycfg: ensure that an irrecoverable error in proxycfg closes the xds session and triggers a replacement proxycfg watcher (#16497) Receiving an "acl not found" error from an RPC in the agent cache and the streaming/event components will cause any request loops to cease under the assumption that they will never work again if the token was destroyed. This prevents log spam (#14144, #9738). Unfortunately due to things like: - authz requests going to stale servers that may not have witnessed the token creation yet - authz requests in a secondary datacenter happening before the tokens get replicated to that datacenter - authz requests from a primary TO a secondary datacenter happening before the tokens get replicated to that datacenter The caller will get an "acl not found" *before* the token exists, rather than just after. The machinery added above in the linked PRs will kick in and prevent the request loop from looping around again once the tokens actually exist. For `consul-dataplane` usages, where xDS is served by the Consul servers rather than the clients ultimately this is not a problem because in that scenario the `agent/proxycfg` machinery is on-demand and launched by a new xDS stream needing data for a specific service in the catalog. If the watching goroutines are terminated it ripples down and terminates the xDS stream, which CDP will eventually re-establish and restart everything. For Consul client usages, the `agent/proxycfg` machinery is ahead-of-time launched at service registration time (called "local" in some of the proxycfg machinery) so when the xDS stream comes in the data is already ready to go. If the watching goroutines terminate it should terminate the xDS stream, but there's no mechanism to re-spawn the watching goroutines. If the xDS stream reconnects it will see no `ConfigSnapshot` and will not get one again until the client agent is restarted, or the service is re-registered with something changed in it. This PR fixes a few things in the machinery: - there was an inadvertent deadlock in fetching snapshot from the proxycfg machinery by xDS, such that when the watching goroutine terminated the snapshots would never be fetched. This caused some of the xDS machinery to get indefinitely paused and not finish the teardown properly. - Every 30s we now attempt to re-insert all locally registered services into the proxycfg machinery. - When services are re-inserted into the proxycfg machinery we special case "dead" ones such that we unilaterally replace them rather that doing that conditionally.
2023-03-03 20:27:53 +00:00
package agent
import (
"encoding/json"
"net/http"
"net/http/httptest"
"testing"
"time"
"github.com/stretchr/testify/require"
"github.com/hashicorp/consul/agent/grpc-external/limiter"
proxycfg: ensure that an irrecoverable error in proxycfg closes the xds session and triggers a replacement proxycfg watcher (#16497) Receiving an "acl not found" error from an RPC in the agent cache and the streaming/event components will cause any request loops to cease under the assumption that they will never work again if the token was destroyed. This prevents log spam (#14144, #9738). Unfortunately due to things like: - authz requests going to stale servers that may not have witnessed the token creation yet - authz requests in a secondary datacenter happening before the tokens get replicated to that datacenter - authz requests from a primary TO a secondary datacenter happening before the tokens get replicated to that datacenter The caller will get an "acl not found" *before* the token exists, rather than just after. The machinery added above in the linked PRs will kick in and prevent the request loop from looping around again once the tokens actually exist. For `consul-dataplane` usages, where xDS is served by the Consul servers rather than the clients ultimately this is not a problem because in that scenario the `agent/proxycfg` machinery is on-demand and launched by a new xDS stream needing data for a specific service in the catalog. If the watching goroutines are terminated it ripples down and terminates the xDS stream, which CDP will eventually re-establish and restart everything. For Consul client usages, the `agent/proxycfg` machinery is ahead-of-time launched at service registration time (called "local" in some of the proxycfg machinery) so when the xDS stream comes in the data is already ready to go. If the watching goroutines terminate it should terminate the xDS stream, but there's no mechanism to re-spawn the watching goroutines. If the xDS stream reconnects it will see no `ConfigSnapshot` and will not get one again until the client agent is restarted, or the service is re-registered with something changed in it. This PR fixes a few things in the machinery: - there was an inadvertent deadlock in fetching snapshot from the proxycfg machinery by xDS, such that when the watching goroutine terminated the snapshots would never be fetched. This caused some of the xDS machinery to get indefinitely paused and not finish the teardown properly. - Every 30s we now attempt to re-insert all locally registered services into the proxycfg machinery. - When services are re-inserted into the proxycfg machinery we special case "dead" ones such that we unilaterally replace them rather that doing that conditionally.
2023-03-03 20:27:53 +00:00
"github.com/hashicorp/consul/agent/structs"
"github.com/hashicorp/consul/api"
proxysnapshot "github.com/hashicorp/consul/internal/mesh/proxy-snapshot"
rtest "github.com/hashicorp/consul/internal/resource/resourcetest"
pbmesh "github.com/hashicorp/consul/proto-public/pbmesh/v2beta1"
proxycfg: ensure that an irrecoverable error in proxycfg closes the xds session and triggers a replacement proxycfg watcher (#16497) Receiving an "acl not found" error from an RPC in the agent cache and the streaming/event components will cause any request loops to cease under the assumption that they will never work again if the token was destroyed. This prevents log spam (#14144, #9738). Unfortunately due to things like: - authz requests going to stale servers that may not have witnessed the token creation yet - authz requests in a secondary datacenter happening before the tokens get replicated to that datacenter - authz requests from a primary TO a secondary datacenter happening before the tokens get replicated to that datacenter The caller will get an "acl not found" *before* the token exists, rather than just after. The machinery added above in the linked PRs will kick in and prevent the request loop from looping around again once the tokens actually exist. For `consul-dataplane` usages, where xDS is served by the Consul servers rather than the clients ultimately this is not a problem because in that scenario the `agent/proxycfg` machinery is on-demand and launched by a new xDS stream needing data for a specific service in the catalog. If the watching goroutines are terminated it ripples down and terminates the xDS stream, which CDP will eventually re-establish and restart everything. For Consul client usages, the `agent/proxycfg` machinery is ahead-of-time launched at service registration time (called "local" in some of the proxycfg machinery) so when the xDS stream comes in the data is already ready to go. If the watching goroutines terminate it should terminate the xDS stream, but there's no mechanism to re-spawn the watching goroutines. If the xDS stream reconnects it will see no `ConfigSnapshot` and will not get one again until the client agent is restarted, or the service is re-registered with something changed in it. This PR fixes a few things in the machinery: - there was an inadvertent deadlock in fetching snapshot from the proxycfg machinery by xDS, such that when the watching goroutine terminated the snapshots would never be fetched. This caused some of the xDS machinery to get indefinitely paused and not finish the teardown properly. - Every 30s we now attempt to re-insert all locally registered services into the proxycfg machinery. - When services are re-inserted into the proxycfg machinery we special case "dead" ones such that we unilaterally replace them rather that doing that conditionally.
2023-03-03 20:27:53 +00:00
"github.com/hashicorp/consul/testrpc"
)
func TestAgent_local_proxycfg(t *testing.T) {
a := NewTestAgent(t, TestACLConfig())
defer a.Shutdown()
testrpc.WaitForLeader(t, a.RPC, "dc1")
token := generateUUID()
svc := &structs.NodeService{
ID: "db",
Service: "db",
Port: 5000,
EnterpriseMeta: *structs.DefaultEnterpriseMetaInDefaultPartition(),
}
require.NoError(t, a.State.AddServiceWithChecks(svc, nil, token, true))
proxy := &structs.NodeService{
Kind: structs.ServiceKindConnectProxy,
ID: "db-sidecar-proxy",
Service: "db-sidecar-proxy",
Port: 5000,
// Set this internal state that we expect sidecar registrations to have.
LocallyRegisteredAsSidecar: true,
Proxy: structs.ConnectProxyConfig{
DestinationServiceName: "db",
2023-03-22 18:56:18 +00:00
Upstreams: structs.TestUpstreams(t, false),
proxycfg: ensure that an irrecoverable error in proxycfg closes the xds session and triggers a replacement proxycfg watcher (#16497) Receiving an "acl not found" error from an RPC in the agent cache and the streaming/event components will cause any request loops to cease under the assumption that they will never work again if the token was destroyed. This prevents log spam (#14144, #9738). Unfortunately due to things like: - authz requests going to stale servers that may not have witnessed the token creation yet - authz requests in a secondary datacenter happening before the tokens get replicated to that datacenter - authz requests from a primary TO a secondary datacenter happening before the tokens get replicated to that datacenter The caller will get an "acl not found" *before* the token exists, rather than just after. The machinery added above in the linked PRs will kick in and prevent the request loop from looping around again once the tokens actually exist. For `consul-dataplane` usages, where xDS is served by the Consul servers rather than the clients ultimately this is not a problem because in that scenario the `agent/proxycfg` machinery is on-demand and launched by a new xDS stream needing data for a specific service in the catalog. If the watching goroutines are terminated it ripples down and terminates the xDS stream, which CDP will eventually re-establish and restart everything. For Consul client usages, the `agent/proxycfg` machinery is ahead-of-time launched at service registration time (called "local" in some of the proxycfg machinery) so when the xDS stream comes in the data is already ready to go. If the watching goroutines terminate it should terminate the xDS stream, but there's no mechanism to re-spawn the watching goroutines. If the xDS stream reconnects it will see no `ConfigSnapshot` and will not get one again until the client agent is restarted, or the service is re-registered with something changed in it. This PR fixes a few things in the machinery: - there was an inadvertent deadlock in fetching snapshot from the proxycfg machinery by xDS, such that when the watching goroutine terminated the snapshots would never be fetched. This caused some of the xDS machinery to get indefinitely paused and not finish the teardown properly. - Every 30s we now attempt to re-insert all locally registered services into the proxycfg machinery. - When services are re-inserted into the proxycfg machinery we special case "dead" ones such that we unilaterally replace them rather that doing that conditionally.
2023-03-03 20:27:53 +00:00
},
EnterpriseMeta: *structs.DefaultEnterpriseMetaInDefaultPartition(),
}
require.NoError(t, a.State.AddServiceWithChecks(proxy, nil, token, true))
// This is a little gross, but this gives us the layered pair of
// local/catalog sources for now.
cfg := a.xdsServer.ProxyWatcher
proxycfg: ensure that an irrecoverable error in proxycfg closes the xds session and triggers a replacement proxycfg watcher (#16497) Receiving an "acl not found" error from an RPC in the agent cache and the streaming/event components will cause any request loops to cease under the assumption that they will never work again if the token was destroyed. This prevents log spam (#14144, #9738). Unfortunately due to things like: - authz requests going to stale servers that may not have witnessed the token creation yet - authz requests in a secondary datacenter happening before the tokens get replicated to that datacenter - authz requests from a primary TO a secondary datacenter happening before the tokens get replicated to that datacenter The caller will get an "acl not found" *before* the token exists, rather than just after. The machinery added above in the linked PRs will kick in and prevent the request loop from looping around again once the tokens actually exist. For `consul-dataplane` usages, where xDS is served by the Consul servers rather than the clients ultimately this is not a problem because in that scenario the `agent/proxycfg` machinery is on-demand and launched by a new xDS stream needing data for a specific service in the catalog. If the watching goroutines are terminated it ripples down and terminates the xDS stream, which CDP will eventually re-establish and restart everything. For Consul client usages, the `agent/proxycfg` machinery is ahead-of-time launched at service registration time (called "local" in some of the proxycfg machinery) so when the xDS stream comes in the data is already ready to go. If the watching goroutines terminate it should terminate the xDS stream, but there's no mechanism to re-spawn the watching goroutines. If the xDS stream reconnects it will see no `ConfigSnapshot` and will not get one again until the client agent is restarted, or the service is re-registered with something changed in it. This PR fixes a few things in the machinery: - there was an inadvertent deadlock in fetching snapshot from the proxycfg machinery by xDS, such that when the watching goroutine terminated the snapshots would never be fetched. This caused some of the xDS machinery to get indefinitely paused and not finish the teardown properly. - Every 30s we now attempt to re-insert all locally registered services into the proxycfg machinery. - When services are re-inserted into the proxycfg machinery we special case "dead" ones such that we unilaterally replace them rather that doing that conditionally.
2023-03-03 20:27:53 +00:00
var (
timer = time.After(100 * time.Millisecond)
timerFired = false
finalTimer <-chan time.Time
)
var (
firstTime = true
ch <-chan proxysnapshot.ProxySnapshot
proxycfg: ensure that an irrecoverable error in proxycfg closes the xds session and triggers a replacement proxycfg watcher (#16497) Receiving an "acl not found" error from an RPC in the agent cache and the streaming/event components will cause any request loops to cease under the assumption that they will never work again if the token was destroyed. This prevents log spam (#14144, #9738). Unfortunately due to things like: - authz requests going to stale servers that may not have witnessed the token creation yet - authz requests in a secondary datacenter happening before the tokens get replicated to that datacenter - authz requests from a primary TO a secondary datacenter happening before the tokens get replicated to that datacenter The caller will get an "acl not found" *before* the token exists, rather than just after. The machinery added above in the linked PRs will kick in and prevent the request loop from looping around again once the tokens actually exist. For `consul-dataplane` usages, where xDS is served by the Consul servers rather than the clients ultimately this is not a problem because in that scenario the `agent/proxycfg` machinery is on-demand and launched by a new xDS stream needing data for a specific service in the catalog. If the watching goroutines are terminated it ripples down and terminates the xDS stream, which CDP will eventually re-establish and restart everything. For Consul client usages, the `agent/proxycfg` machinery is ahead-of-time launched at service registration time (called "local" in some of the proxycfg machinery) so when the xDS stream comes in the data is already ready to go. If the watching goroutines terminate it should terminate the xDS stream, but there's no mechanism to re-spawn the watching goroutines. If the xDS stream reconnects it will see no `ConfigSnapshot` and will not get one again until the client agent is restarted, or the service is re-registered with something changed in it. This PR fixes a few things in the machinery: - there was an inadvertent deadlock in fetching snapshot from the proxycfg machinery by xDS, such that when the watching goroutine terminated the snapshots would never be fetched. This caused some of the xDS machinery to get indefinitely paused and not finish the teardown properly. - Every 30s we now attempt to re-insert all locally registered services into the proxycfg machinery. - When services are re-inserted into the proxycfg machinery we special case "dead" ones such that we unilaterally replace them rather that doing that conditionally.
2023-03-03 20:27:53 +00:00
stc limiter.SessionTerminatedChan
cancel proxysnapshot.CancelFunc
proxycfg: ensure that an irrecoverable error in proxycfg closes the xds session and triggers a replacement proxycfg watcher (#16497) Receiving an "acl not found" error from an RPC in the agent cache and the streaming/event components will cause any request loops to cease under the assumption that they will never work again if the token was destroyed. This prevents log spam (#14144, #9738). Unfortunately due to things like: - authz requests going to stale servers that may not have witnessed the token creation yet - authz requests in a secondary datacenter happening before the tokens get replicated to that datacenter - authz requests from a primary TO a secondary datacenter happening before the tokens get replicated to that datacenter The caller will get an "acl not found" *before* the token exists, rather than just after. The machinery added above in the linked PRs will kick in and prevent the request loop from looping around again once the tokens actually exist. For `consul-dataplane` usages, where xDS is served by the Consul servers rather than the clients ultimately this is not a problem because in that scenario the `agent/proxycfg` machinery is on-demand and launched by a new xDS stream needing data for a specific service in the catalog. If the watching goroutines are terminated it ripples down and terminates the xDS stream, which CDP will eventually re-establish and restart everything. For Consul client usages, the `agent/proxycfg` machinery is ahead-of-time launched at service registration time (called "local" in some of the proxycfg machinery) so when the xDS stream comes in the data is already ready to go. If the watching goroutines terminate it should terminate the xDS stream, but there's no mechanism to re-spawn the watching goroutines. If the xDS stream reconnects it will see no `ConfigSnapshot` and will not get one again until the client agent is restarted, or the service is re-registered with something changed in it. This PR fixes a few things in the machinery: - there was an inadvertent deadlock in fetching snapshot from the proxycfg machinery by xDS, such that when the watching goroutine terminated the snapshots would never be fetched. This caused some of the xDS machinery to get indefinitely paused and not finish the teardown properly. - Every 30s we now attempt to re-insert all locally registered services into the proxycfg machinery. - When services are re-inserted into the proxycfg machinery we special case "dead" ones such that we unilaterally replace them rather that doing that conditionally.
2023-03-03 20:27:53 +00:00
)
defer func() {
if cancel != nil {
cancel()
}
}()
for {
if ch == nil {
// Sign up for a stream of config snapshots, in the same manner as the xds server.
sid := proxy.CompoundServiceID()
if firstTime {
firstTime = false
} else {
t.Logf("re-creating watch")
}
// Prior to fixes in https://github.com/hashicorp/consul/pull/16497
// this call to Watch() would deadlock.
var err error
Fix xDS deadlock due to syncLoop termination. (#20867) * Fix xDS deadlock due to syncLoop termination. This fixes an issue where agentless xDS streams can deadlock permanently until a server is restarted. When this issue occurs, no new proxies are able to successfully connect to the server. Effectively, the trigger for this deadlock stems from the following return statement: https://github.com/hashicorp/consul/blob/v1.18.0/agent/proxycfg-sources/catalog/config_source.go#L199-L202 When this happens, the entire `syncLoop()` terminates and stops consuming from the following channel: https://github.com/hashicorp/consul/blob/v1.18.0/agent/proxycfg-sources/catalog/config_source.go#L182-L192 Which results in the `ConfigSource.cleanup()` function never receiving a response and holding a mutex indefinitely: https://github.com/hashicorp/consul/blob/v1.18.0/agent/proxycfg-sources/catalog/config_source.go#L241-L247 Because this mutex is shared, it effectively deadlocks the server's ability to process new xDS streams. ---- The fix to this issue involves removing the `chan chan struct{}` used like an RPC-over-channels pattern and replacing it with two distinct channels: + `stopSyncLoopCh` - indicates that the `syncLoop()` should terminate soon. + `syncLoopDoneCh` - indicates that the `syncLoop()` has terminated. Splitting these two concepts out and deferring a `close(syncLoopDoneCh)` in the `syncLoop()` function ensures that the deadlock above should no longer occur. We also now evict xDS connections of all proxies for the corresponding `syncLoop()` whenever it encounters an irrecoverable error. This is done by hoisting the new `syncLoopDoneCh` upwards so that it's visible to the xDS delta processing. Prior to this fix, the behavior was to simply orphan them so they would never receive catalog-registration or service-defaults updates. * Add changelog.
2024-03-15 18:57:11 +00:00
ch, stc, _, cancel, err = cfg.Watch(rtest.Resource(pbmesh.ProxyConfigurationType, sid.ID).ID(), a.config.NodeName, token)
proxycfg: ensure that an irrecoverable error in proxycfg closes the xds session and triggers a replacement proxycfg watcher (#16497) Receiving an "acl not found" error from an RPC in the agent cache and the streaming/event components will cause any request loops to cease under the assumption that they will never work again if the token was destroyed. This prevents log spam (#14144, #9738). Unfortunately due to things like: - authz requests going to stale servers that may not have witnessed the token creation yet - authz requests in a secondary datacenter happening before the tokens get replicated to that datacenter - authz requests from a primary TO a secondary datacenter happening before the tokens get replicated to that datacenter The caller will get an "acl not found" *before* the token exists, rather than just after. The machinery added above in the linked PRs will kick in and prevent the request loop from looping around again once the tokens actually exist. For `consul-dataplane` usages, where xDS is served by the Consul servers rather than the clients ultimately this is not a problem because in that scenario the `agent/proxycfg` machinery is on-demand and launched by a new xDS stream needing data for a specific service in the catalog. If the watching goroutines are terminated it ripples down and terminates the xDS stream, which CDP will eventually re-establish and restart everything. For Consul client usages, the `agent/proxycfg` machinery is ahead-of-time launched at service registration time (called "local" in some of the proxycfg machinery) so when the xDS stream comes in the data is already ready to go. If the watching goroutines terminate it should terminate the xDS stream, but there's no mechanism to re-spawn the watching goroutines. If the xDS stream reconnects it will see no `ConfigSnapshot` and will not get one again until the client agent is restarted, or the service is re-registered with something changed in it. This PR fixes a few things in the machinery: - there was an inadvertent deadlock in fetching snapshot from the proxycfg machinery by xDS, such that when the watching goroutine terminated the snapshots would never be fetched. This caused some of the xDS machinery to get indefinitely paused and not finish the teardown properly. - Every 30s we now attempt to re-insert all locally registered services into the proxycfg machinery. - When services are re-inserted into the proxycfg machinery we special case "dead" ones such that we unilaterally replace them rather that doing that conditionally.
2023-03-03 20:27:53 +00:00
require.NoError(t, err)
}
select {
case <-stc:
t.Fatal("session unexpectedly terminated")
case snap, ok := <-ch:
if !ok {
t.Logf("channel is closed")
cancel()
ch, stc, cancel = nil, nil, nil
continue
}
require.NotNil(t, snap)
if !timerFired {
t.Fatal("should not have gotten snapshot until after we manifested the token")
}
return
case <-timer:
timerFired = true
finalTimer = time.After(1 * time.Second)
// This simulates the eventual consistency of a token
// showing up on a server after it's creation by
// pre-creating the UUID and later using that as the
// initial SecretID for a real token.
gotToken := testWriteToken(t, a, &api.ACLToken{
AccessorID: generateUUID(),
SecretID: token,
Description: "my token",
ServiceIdentities: []*api.ACLServiceIdentity{{
ServiceName: "db",
}},
})
require.Equal(t, token, gotToken)
case <-finalTimer:
t.Fatal("did not receive a snapshot after the token manifested")
}
}
}
func testWriteToken(t *testing.T, a *TestAgent, tok *api.ACLToken) string {
req, _ := http.NewRequest("PUT", "/v1/acl/token", jsonReader(tok))
req.Header.Add("X-Consul-Token", "root")
resp := httptest.NewRecorder()
a.srv.h.ServeHTTP(resp, req)
require.Equal(t, http.StatusOK, resp.Code)
dec := json.NewDecoder(resp.Body)
aclResp := &structs.ACLToken{}
require.NoError(t, dec.Decode(aclResp))
return aclResp.SecretID
}