consul/testutil/server.go

398 lines
12 KiB
Go
Raw Normal View History

package testutil
// TestServer is a test helper. It uses a fork/exec model to create
// a test Consul server instance in the background and initialize it
// with some data and/or services. The test server can then be used
// to run a unit test, and offers an easy API to tear itself down
// when the test has completed. The only prerequisite is to have a consul
// binary available on the $PATH.
//
// This package does not use Consul's official API client. This is
// because we use TestServer to test the API client, which would
// otherwise cause an import cycle.
import (
"context"
"encoding/json"
"fmt"
2015-03-11 01:08:14 +00:00
"io"
"io/ioutil"
"log"
"net"
"net/http"
"os"
"os/exec"
"path/filepath"
"strconv"
"strings"
"testing"
"time"
"github.com/hashicorp/consul/lib/freeport"
"github.com/hashicorp/consul/testutil/retry"
2015-10-22 18:14:22 +00:00
"github.com/hashicorp/go-cleanhttp"
"github.com/hashicorp/go-uuid"
"github.com/pkg/errors"
)
// TestPerformanceConfig configures the performance parameters.
type TestPerformanceConfig struct {
RaftMultiplier uint `json:"raft_multiplier,omitempty"`
}
2015-03-11 04:53:51 +00:00
// TestPortConfig configures the various ports used for services
// provided by the Consul server.
type TestPortConfig struct {
2018-06-11 20:25:13 +00:00
DNS int `json:"dns,omitempty"`
HTTP int `json:"http,omitempty"`
HTTPS int `json:"https,omitempty"`
SerfLan int `json:"serf_lan,omitempty"`
SerfWan int `json:"serf_wan,omitempty"`
Server int `json:"server,omitempty"`
ProxyMinPort int `json:"proxy_min_port,omitempty"`
ProxyMaxPort int `json:"proxy_max_port,omitempty"`
}
2015-03-11 04:53:51 +00:00
// TestAddressConfig contains the bind addresses for various
// components of the Consul server.
type TestAddressConfig struct {
HTTP string `json:"http,omitempty"`
}
// TestNetworkSegment contains the configuration for a network segment.
type TestNetworkSegment struct {
Name string `json:"name"`
Bind string `json:"bind"`
Port int `json:"port"`
Advertise string `json:"advertise"`
}
2015-03-11 04:53:51 +00:00
// TestServerConfig is the main server configuration struct.
type TestServerConfig struct {
NodeName string `json:"node_name"`
NodeID string `json:"node_id"`
NodeMeta map[string]string `json:"node_meta,omitempty"`
Performance *TestPerformanceConfig `json:"performance,omitempty"`
Bootstrap bool `json:"bootstrap,omitempty"`
Server bool `json:"server,omitempty"`
DataDir string `json:"data_dir,omitempty"`
Datacenter string `json:"datacenter,omitempty"`
Segments []TestNetworkSegment `json:"segments"`
DisableCheckpoint bool `json:"disable_update_check"`
LogLevel string `json:"log_level,omitempty"`
Bind string `json:"bind_addr,omitempty"`
Addresses *TestAddressConfig `json:"addresses,omitempty"`
Ports *TestPortConfig `json:"ports,omitempty"`
RaftProtocol int `json:"raft_protocol,omitempty"`
ACLMasterToken string `json:"acl_master_token,omitempty"`
ACLDatacenter string `json:"acl_datacenter,omitempty"`
ACLDefaultPolicy string `json:"acl_default_policy,omitempty"`
ACLEnforceVersion8 bool `json:"acl_enforce_version_8"`
Encrypt string `json:"encrypt,omitempty"`
CAFile string `json:"ca_file,omitempty"`
CertFile string `json:"cert_file,omitempty"`
KeyFile string `json:"key_file,omitempty"`
VerifyIncoming bool `json:"verify_incoming,omitempty"`
VerifyIncomingRPC bool `json:"verify_incoming_rpc,omitempty"`
VerifyIncomingHTTPS bool `json:"verify_incoming_https,omitempty"`
VerifyOutgoing bool `json:"verify_outgoing,omitempty"`
EnableScriptChecks bool `json:"enable_script_checks,omitempty"`
Connect map[string]interface{} `json:"connect,omitempty"`
ReadyTimeout time.Duration `json:"-"`
Stdout, Stderr io.Writer `json:"-"`
Args []string `json:"-"`
}
2015-03-11 04:53:51 +00:00
// ServerConfigCallback is a function interface which can be
// passed to NewTestServerConfig to modify the server config.
type ServerConfigCallback func(c *TestServerConfig)
2015-03-11 04:53:51 +00:00
// defaultServerConfig returns a new TestServerConfig struct
// with all of the listen ports incremented by one.
func defaultServerConfig() *TestServerConfig {
nodeID, err := uuid.GenerateUUID()
if err != nil {
panic(err)
}
ports := freeport.Get(6)
return &TestServerConfig{
New config parser, HCL support, multiple bind addrs (#3480) * new config parser for agent This patch implements a new config parser for the consul agent which makes the following changes to the previous implementation: * add HCL support * all configuration fragments in tests and for default config are expressed as HCL fragments * HCL fragments can be provided on the command line so that they can eventually replace the command line flags. * HCL/JSON fragments are parsed into a temporary Config structure which can be merged using reflection (all values are pointers). The existing merge logic of overwrite for values and append for slices has been preserved. * A single builder process generates a typed runtime configuration for the agent. The new implementation is more strict and fails in the builder process if no valid runtime configuration can be generated. Therefore, additional validations in other parts of the code should be removed. The builder also pre-computes all required network addresses so that no address/port magic should be required where the configuration is used and should therefore be removed. * Upgrade github.com/hashicorp/hcl to support int64 * improve error messages * fix directory permission test * Fix rtt test * Fix ForceLeave test * Skip performance test for now until we know what to do * Update github.com/hashicorp/memberlist to update log prefix * Make memberlist use the default logger * improve config error handling * do not fail on non-existing data-dir * experiment with non-uniform timeouts to get a handle on stalled leader elections * Run tests for packages separately to eliminate the spurious port conflicts * refactor private address detection and unify approach for ipv4 and ipv6. Fixes #2825 * do not allow unix sockets for DNS * improve bind and advertise addr error handling * go through builder using test coverage * minimal update to the docs * more coverage tests fixed * more tests * fix makefile * cleanup * fix port conflicts with external port server 'porter' * stop test server on error * do not run api test that change global ENV concurrently with the other tests * Run remaining api tests concurrently * no need for retry with the port number service * monkey patch race condition in go-sockaddr until we understand why that fails * monkey patch hcl decoder race condidtion until we understand why that fails * monkey patch spurious errors in strings.EqualFold from here * add test for hcl decoder race condition. Run with go test -parallel 128 * Increase timeout again * cleanup * don't log port allocations by default * use base command arg parsing to format help output properly * handle -dc deprecation case in Build * switch autopilot.max_trailing_logs to int * remove duplicate test case * remove unused methods * remove comments about flag/config value inconsistencies * switch got and want around since the error message was misleading. * Removes a stray debug log. * Removes a stray newline in imports. * Fixes TestACL_Version8. * Runs go fmt. * Adds a default case for unknown address types. * Reoders and reformats some imports. * Adds some comments and fixes typos. * Reorders imports. * add unix socket support for dns later * drop all deprecated flags and arguments * fix wrong field name * remove stray node-id file * drop unnecessary patch section in test * drop duplicate test * add test for LeaveOnTerm and SkipLeaveOnInt in client mode * drop "bla" and add clarifying comment for the test * split up tests to support enterprise/non-enterprise tests * drop raft multiplier and derive values during build phase * sanitize runtime config reflectively and add test * detect invalid config fields * fix tests with invalid config fields * use different values for wan sanitiziation test * drop recursor in favor of recursors * allow dns_config.udp_answer_limit to be zero * make sure tests run on machines with multiple ips * Fix failing tests in a few more places by providing a bind address in the test * Gets rid of skipped TestAgent_CheckPerformanceSettings and adds case for builder. * Add porter to server_test.go to make tests there less flaky * go fmt
2017-09-25 18:40:42 +00:00
NodeName: "node-" + nodeID,
NodeID: nodeID,
DisableCheckpoint: true,
Performance: &TestPerformanceConfig{
RaftMultiplier: 1,
},
Bootstrap: true,
Server: true,
LogLevel: "debug",
Bind: "127.0.0.1",
Addresses: &TestAddressConfig{},
Ports: &TestPortConfig{
New config parser, HCL support, multiple bind addrs (#3480) * new config parser for agent This patch implements a new config parser for the consul agent which makes the following changes to the previous implementation: * add HCL support * all configuration fragments in tests and for default config are expressed as HCL fragments * HCL fragments can be provided on the command line so that they can eventually replace the command line flags. * HCL/JSON fragments are parsed into a temporary Config structure which can be merged using reflection (all values are pointers). The existing merge logic of overwrite for values and append for slices has been preserved. * A single builder process generates a typed runtime configuration for the agent. The new implementation is more strict and fails in the builder process if no valid runtime configuration can be generated. Therefore, additional validations in other parts of the code should be removed. The builder also pre-computes all required network addresses so that no address/port magic should be required where the configuration is used and should therefore be removed. * Upgrade github.com/hashicorp/hcl to support int64 * improve error messages * fix directory permission test * Fix rtt test * Fix ForceLeave test * Skip performance test for now until we know what to do * Update github.com/hashicorp/memberlist to update log prefix * Make memberlist use the default logger * improve config error handling * do not fail on non-existing data-dir * experiment with non-uniform timeouts to get a handle on stalled leader elections * Run tests for packages separately to eliminate the spurious port conflicts * refactor private address detection and unify approach for ipv4 and ipv6. Fixes #2825 * do not allow unix sockets for DNS * improve bind and advertise addr error handling * go through builder using test coverage * minimal update to the docs * more coverage tests fixed * more tests * fix makefile * cleanup * fix port conflicts with external port server 'porter' * stop test server on error * do not run api test that change global ENV concurrently with the other tests * Run remaining api tests concurrently * no need for retry with the port number service * monkey patch race condition in go-sockaddr until we understand why that fails * monkey patch hcl decoder race condidtion until we understand why that fails * monkey patch spurious errors in strings.EqualFold from here * add test for hcl decoder race condition. Run with go test -parallel 128 * Increase timeout again * cleanup * don't log port allocations by default * use base command arg parsing to format help output properly * handle -dc deprecation case in Build * switch autopilot.max_trailing_logs to int * remove duplicate test case * remove unused methods * remove comments about flag/config value inconsistencies * switch got and want around since the error message was misleading. * Removes a stray debug log. * Removes a stray newline in imports. * Fixes TestACL_Version8. * Runs go fmt. * Adds a default case for unknown address types. * Reoders and reformats some imports. * Adds some comments and fixes typos. * Reorders imports. * add unix socket support for dns later * drop all deprecated flags and arguments * fix wrong field name * remove stray node-id file * drop unnecessary patch section in test * drop duplicate test * add test for LeaveOnTerm and SkipLeaveOnInt in client mode * drop "bla" and add clarifying comment for the test * split up tests to support enterprise/non-enterprise tests * drop raft multiplier and derive values during build phase * sanitize runtime config reflectively and add test * detect invalid config fields * fix tests with invalid config fields * use different values for wan sanitiziation test * drop recursor in favor of recursors * allow dns_config.udp_answer_limit to be zero * make sure tests run on machines with multiple ips * Fix failing tests in a few more places by providing a bind address in the test * Gets rid of skipped TestAgent_CheckPerformanceSettings and adds case for builder. * Add porter to server_test.go to make tests there less flaky * go fmt
2017-09-25 18:40:42 +00:00
DNS: ports[0],
HTTP: ports[1],
HTTPS: ports[2],
SerfLan: ports[3],
SerfWan: ports[4],
Server: ports[5],
},
2017-05-12 19:04:34 +00:00
ReadyTimeout: 10 * time.Second,
Connect: map[string]interface{}{
"enabled": true,
"ca_config": map[string]interface{}{
// const TestClusterID causes import cycle so hard code it here.
"cluster_id": "11111111-2222-3333-4444-555555555555",
},
2018-06-14 13:07:33 +00:00
"proxy": map[string]interface{}{
"allow_managed_api_registration": true,
},
},
}
}
2015-03-11 04:53:51 +00:00
// TestService is used to serialize a service definition.
2015-03-11 01:47:45 +00:00
type TestService struct {
ID string `json:",omitempty"`
Name string `json:",omitempty"`
Tags []string `json:",omitempty"`
Address string `json:",omitempty"`
Port int `json:",omitempty"`
}
2015-03-11 04:53:51 +00:00
// TestCheck is used to serialize a check definition.
2015-03-11 01:47:45 +00:00
type TestCheck struct {
ID string `json:",omitempty"`
Name string `json:",omitempty"`
ServiceID string `json:",omitempty"`
TTL string `json:",omitempty"`
}
2015-03-11 23:10:07 +00:00
// TestKVResponse is what we use to decode KV data.
type TestKVResponse struct {
Value string
}
2015-03-11 04:53:51 +00:00
// TestServer is the main server wrapper struct.
type TestServer struct {
cmd *exec.Cmd
Config *TestServerConfig
2017-04-14 20:37:29 +00:00
HTTPAddr string
HTTPSAddr string
LANAddr string
WANAddr string
2017-04-21 00:02:42 +00:00
HTTPClient *http.Client
tmpdir string
}
2015-03-11 04:53:51 +00:00
// NewTestServer is an easy helper method to create a new Consul
// test server with the most basic configuration.
func NewTestServer() (*TestServer, error) {
return NewTestServerConfigT(nil, nil)
}
func NewTestServerConfig(cb ServerConfigCallback) (*TestServer, error) {
return NewTestServerConfigT(nil, cb)
}
// NewTestServerConfig creates a new TestServer, and makes a call to an optional
// callback function to modify the configuration. If there is an error
// configuring or starting the server, the server will NOT be running when the
// function returns (thus you do not need to stop it).
func NewTestServerConfigT(t *testing.T, cb ServerConfigCallback) (*TestServer, error) {
New config parser, HCL support, multiple bind addrs (#3480) * new config parser for agent This patch implements a new config parser for the consul agent which makes the following changes to the previous implementation: * add HCL support * all configuration fragments in tests and for default config are expressed as HCL fragments * HCL fragments can be provided on the command line so that they can eventually replace the command line flags. * HCL/JSON fragments are parsed into a temporary Config structure which can be merged using reflection (all values are pointers). The existing merge logic of overwrite for values and append for slices has been preserved. * A single builder process generates a typed runtime configuration for the agent. The new implementation is more strict and fails in the builder process if no valid runtime configuration can be generated. Therefore, additional validations in other parts of the code should be removed. The builder also pre-computes all required network addresses so that no address/port magic should be required where the configuration is used and should therefore be removed. * Upgrade github.com/hashicorp/hcl to support int64 * improve error messages * fix directory permission test * Fix rtt test * Fix ForceLeave test * Skip performance test for now until we know what to do * Update github.com/hashicorp/memberlist to update log prefix * Make memberlist use the default logger * improve config error handling * do not fail on non-existing data-dir * experiment with non-uniform timeouts to get a handle on stalled leader elections * Run tests for packages separately to eliminate the spurious port conflicts * refactor private address detection and unify approach for ipv4 and ipv6. Fixes #2825 * do not allow unix sockets for DNS * improve bind and advertise addr error handling * go through builder using test coverage * minimal update to the docs * more coverage tests fixed * more tests * fix makefile * cleanup * fix port conflicts with external port server 'porter' * stop test server on error * do not run api test that change global ENV concurrently with the other tests * Run remaining api tests concurrently * no need for retry with the port number service * monkey patch race condition in go-sockaddr until we understand why that fails * monkey patch hcl decoder race condidtion until we understand why that fails * monkey patch spurious errors in strings.EqualFold from here * add test for hcl decoder race condition. Run with go test -parallel 128 * Increase timeout again * cleanup * don't log port allocations by default * use base command arg parsing to format help output properly * handle -dc deprecation case in Build * switch autopilot.max_trailing_logs to int * remove duplicate test case * remove unused methods * remove comments about flag/config value inconsistencies * switch got and want around since the error message was misleading. * Removes a stray debug log. * Removes a stray newline in imports. * Fixes TestACL_Version8. * Runs go fmt. * Adds a default case for unknown address types. * Reoders and reformats some imports. * Adds some comments and fixes typos. * Reorders imports. * add unix socket support for dns later * drop all deprecated flags and arguments * fix wrong field name * remove stray node-id file * drop unnecessary patch section in test * drop duplicate test * add test for LeaveOnTerm and SkipLeaveOnInt in client mode * drop "bla" and add clarifying comment for the test * split up tests to support enterprise/non-enterprise tests * drop raft multiplier and derive values during build phase * sanitize runtime config reflectively and add test * detect invalid config fields * fix tests with invalid config fields * use different values for wan sanitiziation test * drop recursor in favor of recursors * allow dns_config.udp_answer_limit to be zero * make sure tests run on machines with multiple ips * Fix failing tests in a few more places by providing a bind address in the test * Gets rid of skipped TestAgent_CheckPerformanceSettings and adds case for builder. * Add porter to server_test.go to make tests there less flaky * go fmt
2017-09-25 18:40:42 +00:00
return newTestServerConfigT(t, cb)
}
// newTestServerConfigT is the internal helper for NewTestServerConfigT.
func newTestServerConfigT(t *testing.T, cb ServerConfigCallback) (*TestServer, error) {
path, err := exec.LookPath("consul")
if err != nil || path == "" {
return nil, fmt.Errorf("consul not found on $PATH - download and install " +
"consul or skip this test")
}
tmpdir := TempDir(t, "consul")
cfg := defaultServerConfig()
cfg.DataDir = filepath.Join(tmpdir, "data")
if cb != nil {
cb(cfg)
}
b, err := json.Marshal(cfg)
if err != nil {
return nil, errors.Wrap(err, "failed marshaling json")
}
log.Printf("CONFIG JSON: %s", string(b))
configFile := filepath.Join(tmpdir, "config.json")
if err := ioutil.WriteFile(configFile, b, 0644); err != nil {
defer os.RemoveAll(tmpdir)
return nil, errors.Wrap(err, "failed writing config content")
}
stdout := io.Writer(os.Stdout)
if cfg.Stdout != nil {
stdout = cfg.Stdout
}
stderr := io.Writer(os.Stderr)
if cfg.Stderr != nil {
stderr = cfg.Stderr
}
// Start the server
args := []string{"agent", "-config-file", configFile}
args = append(args, cfg.Args...)
cmd := exec.Command("consul", args...)
cmd.Stdout = stdout
cmd.Stderr = stderr
if err := cmd.Start(); err != nil {
return nil, errors.Wrap(err, "failed starting command")
}
httpAddr := fmt.Sprintf("127.0.0.1:%d", cfg.Ports.HTTP)
client := cleanhttp.DefaultClient()
if strings.HasPrefix(cfg.Addresses.HTTP, "unix://") {
httpAddr = cfg.Addresses.HTTP
tr := cleanhttp.DefaultTransport()
tr.DialContext = func(_ context.Context, _, _ string) (net.Conn, error) {
return net.Dial("unix", httpAddr[len("unix://"):])
}
client = &http.Client{Transport: tr}
}
server := &TestServer{
Config: cfg,
cmd: cmd,
2017-04-14 20:37:29 +00:00
HTTPAddr: httpAddr,
HTTPSAddr: fmt.Sprintf("127.0.0.1:%d", cfg.Ports.HTTPS),
LANAddr: fmt.Sprintf("127.0.0.1:%d", cfg.Ports.SerfLan),
WANAddr: fmt.Sprintf("127.0.0.1:%d", cfg.Ports.SerfWan),
2017-04-21 00:02:42 +00:00
HTTPClient: client,
tmpdir: tmpdir,
}
// Wait for the server to be ready
if cfg.Bootstrap {
err = server.waitForLeader()
} else {
err = server.waitForAPI()
}
if err != nil {
defer server.Stop()
return nil, errors.Wrap(err, "failed waiting for server to start")
}
return server, nil
}
2015-03-11 04:53:51 +00:00
// Stop stops the test Consul server, and removes the Consul data
// directory once we are done.
func (s *TestServer) Stop() error {
defer os.RemoveAll(s.tmpdir)
// There was no process
if s.cmd == nil {
return nil
}
if s.cmd.Process != nil {
2017-05-19 08:49:46 +00:00
if err := s.cmd.Process.Signal(os.Interrupt); err != nil {
return errors.Wrap(err, "failed to kill consul server")
}
}
// wait for the process to exit to be sure that the data dir can be
// deleted on all platforms.
return s.cmd.Wait()
}
type failer struct {
failed bool
}
func (f *failer) Log(args ...interface{}) { fmt.Println(args) }
func (f *failer) FailNow() { f.failed = true }
// waitForAPI waits for only the agent HTTP endpoint to start
// responding. This is an indication that the agent has started,
// but will likely return before a leader is elected.
func (s *TestServer) waitForAPI() error {
f := &failer{}
retry.Run(f, func(r *retry.R) {
2017-04-21 00:02:42 +00:00
resp, err := s.HTTPClient.Get(s.url("/v1/agent/self"))
if err != nil {
r.Fatal(err)
}
defer resp.Body.Close()
if err := s.requireOK(resp); err != nil {
r.Fatal("failed OK response", err)
}
})
if f.failed {
return errors.New("failed waiting for API")
}
return nil
}
2015-03-11 04:53:51 +00:00
// waitForLeader waits for the Consul server's HTTP API to become
// available, and then waits for a known leader and an index of
// 1 or more to be observed to confirm leader election is done.
// It then waits to ensure the anti-entropy sync has completed.
func (s *TestServer) waitForLeader() error {
f := &failer{}
timer := &retry.Timer{
Timeout: s.Config.ReadyTimeout,
Wait: 250 * time.Millisecond,
}
var index int64
retry.RunWith(timer, f, func(r *retry.R) {
// Query the API and check the status code.
url := s.url(fmt.Sprintf("/v1/catalog/nodes?index=%d", index))
2017-04-21 00:02:42 +00:00
resp, err := s.HTTPClient.Get(url)
if err != nil {
r.Fatal("failed http get", err)
}
defer resp.Body.Close()
if err := s.requireOK(resp); err != nil {
r.Fatal("failed OK response", err)
}
// Ensure we have a leader and a node registration.
if leader := resp.Header.Get("X-Consul-KnownLeader"); leader != "true" {
r.Fatalf("Consul leader status: %#v", leader)
}
index, err = strconv.ParseInt(resp.Header.Get("X-Consul-Index"), 10, 64)
if err != nil {
r.Fatal("bad consul index", err)
}
if index == 0 {
r.Fatal("consul index is 0")
}
// Watch for the anti-entropy sync to finish.
var v []map[string]interface{}
dec := json.NewDecoder(resp.Body)
if err := dec.Decode(&v); err != nil {
r.Fatal(err)
}
if len(v) < 1 {
r.Fatal("No nodes")
}
taggedAddresses, ok := v[0]["TaggedAddresses"].(map[string]interface{})
if !ok {
r.Fatal("Missing tagged addresses")
}
if _, ok := taggedAddresses["lan"]; !ok {
r.Fatal("No lan tagged addresses")
}
})
if f.failed {
return errors.New("failed waiting for leader")
2015-03-11 01:08:14 +00:00
}
return nil
2015-03-11 01:08:14 +00:00
}