depot/packages/networking/ipfs-cluster/cluster_config.go

578 lines
18 KiB
Go
Raw Normal View History

2022-10-19 23:23:11 +03:00
package ipfscluster
import (
"crypto/rand"
"encoding/hex"
"encoding/json"
"errors"
"fmt"
"os"
"path/filepath"
"reflect"
"time"
"github.com/ipfs-cluster/ipfs-cluster/config"
ipfsconfig "github.com/ipfs/go-ipfs-config"
pnet "github.com/libp2p/go-libp2p-core/pnet"
ma "github.com/multiformats/go-multiaddr"
"github.com/kelseyhightower/envconfig"
)
const configKey = "cluster"
// DefaultListenAddrs contains TCP and QUIC listen addresses.
var DefaultListenAddrs = []string{
"/ip4/0.0.0.0/tcp/9096",
"/ip4/0.0.0.0/udp/9096/quic",
}
// Configuration defaults
const (
DefaultEnableRelayHop = true
DefaultStateSyncInterval = 5 * time.Minute
DefaultPinRecoverInterval = 12 * time.Minute
DefaultMonitorPingInterval = 15 * time.Second
DefaultPeerWatchInterval = 5 * time.Second
DefaultReplicationFactor = -1
DefaultLeaveOnShutdown = false
DefaultPinOnlyOnTrustedPeers = false
DefaultDisableRepinning = true
DefaultPeerstoreFile = "peerstore"
DefaultConnMgrHighWater = 400
DefaultConnMgrLowWater = 100
DefaultConnMgrGracePeriod = 2 * time.Minute
DefaultDialPeerTimeout = 3 * time.Second
DefaultFollowerMode = false
DefaultMDNSInterval = 10 * time.Second
)
// ConnMgrConfig configures the libp2p host connection manager.
type ConnMgrConfig struct {
HighWater int
LowWater int
GracePeriod time.Duration
}
// Config is the configuration object containing customizable variables to
// initialize the main ipfs-cluster component. It implements the
// config.ComponentConfig interface.
type Config struct {
config.Saver
// User-defined peername for use as human-readable identifier.
Peername string
// Cluster secret for private network. Peers will be in the same cluster if and
// only if they have the same ClusterSecret. The cluster secret must be exactly
// 64 characters and contain only hexadecimal characters (`[0-9a-f]`).
Secret pnet.PSK
// RPCPolicy defines access control to RPC endpoints.
RPCPolicy map[string]RPCEndpointType
// Leave Cluster on shutdown. Politely informs other peers
// of the departure and removes itself from the consensus
// peer set. The Cluster size will be reduced by one.
LeaveOnShutdown bool
// Listen parameters for the Cluster libp2p Host. Used by
// the RPC and Consensus components.
ListenAddr []ma.Multiaddr
// Enables HOP relay for the node. If this is enabled, the node will act as
// an intermediate (Hop Relay) node in relay circuits for connected peers.
EnableRelayHop bool
// ConnMgr holds configuration values for the connection manager for
// the libp2p host.
// FIXME: This only applies to ipfs-cluster-service.
ConnMgr ConnMgrConfig
// Sets the default dial timeout for libp2p connections to other
// peers.
DialPeerTimeout time.Duration
// Time between syncs of the consensus state to the
// tracker state. Normally states are synced anyway, but this helps
// when new nodes are joining the cluster. Reduce for faster
// consistency, increase with larger states.
StateSyncInterval time.Duration
// Time between automatic runs of the "recover" operation
// which will retry to pin/unpin items in error state.
PinRecoverInterval time.Duration
// ReplicationFactorMax indicates the target number of nodes
// that should pin content. For exampe, a replication_factor of
// 3 will have cluster allocate each pinned hash to 3 peers if
// possible.
// See also ReplicationFactorMin. A ReplicationFactorMax of -1
// will allocate to every available node.
ReplicationFactorMax int
// ReplicationFactorMin indicates the minimum number of healthy
// nodes pinning content. If the number of nodes available to pin
// is less than this threshold, an error will be returned.
// In the case of peer health issues, content pinned will be
// re-allocated if the threshold is crossed.
// For exampe, a ReplicationFactorMin of 2 will allocate at least
// two peer to hold content, and return an error if this is not
// possible.
ReplicationFactorMin int
// MonitorPingInterval is the frequency with which a cluster peer
// sends a "ping" metric. The metric has a TTL set to the double of
// this value. This metric sends information about this peer to other
// peers.
MonitorPingInterval time.Duration
// PeerWatchInterval is the frequency that we use to watch for changes
// in the consensus peerset and save new peers to the configuration
// file. This also affects how soon we realize that we have
// been removed from a cluster.
PeerWatchInterval time.Duration
// MDNSInterval controls the time between mDNS broadcasts to the
// network announcing the peer addresses. Set to 0 to disable
// mDNS.
MDNSInterval time.Duration
// PinOnlyOnTrustedPeers limits allocations to trusted peers only.
PinOnlyOnTrustedPeers bool
// If true, DisableRepinning, ensures that no repinning happens
// when a node goes down.
// This is useful when doing certain types of maintenance, or simply
// when not wanting to rely on the monitoring system which needs a revamp.
DisableRepinning bool
// FollowerMode disables broadcast requests from this peer
// (sync, recover, status) and disallows pinset management
// operations (Pin/Unpin).
FollowerMode bool
// Peerstore file specifies the file on which we persist the
// libp2p host peerstore addresses. This file is regularly saved.
PeerstoreFile string
// PeerAddresses stores additional addresses for peers that may or may
// not be in the peerstore file. These are considered high priority
// when bootstrapping the initial cluster connections.
PeerAddresses []ma.Multiaddr
// Tracing flag used to skip tracing specific paths when not enabled.
Tracing bool
}
// configJSON represents a Cluster configuration as it will look when it is
// saved using JSON. Most configuration keys are converted into simple types
// like strings, and key names aim to be self-explanatory for the user.
type configJSON struct {
ID string `json:"id,omitempty"`
Peername string `json:"peername"`
PrivateKey string `json:"private_key,omitempty" hidden:"true"`
Secret string `json:"secret" hidden:"true"`
LeaveOnShutdown bool `json:"leave_on_shutdown"`
ListenMultiaddress ipfsconfig.Strings `json:"listen_multiaddress"`
EnableRelayHop bool `json:"enable_relay_hop"`
ConnectionManager *connMgrConfigJSON `json:"connection_manager"`
DialPeerTimeout string `json:"dial_peer_timeout"`
StateSyncInterval string `json:"state_sync_interval"`
PinRecoverInterval string `json:"pin_recover_interval"`
ReplicationFactorMin int `json:"replication_factor_min"`
ReplicationFactorMax int `json:"replication_factor_max"`
MonitorPingInterval string `json:"monitor_ping_interval"`
PeerWatchInterval string `json:"peer_watch_interval"`
MDNSInterval string `json:"mdns_interval"`
PinOnlyOnTrustedPeers bool `json:"pin_only_on_trusted_peers"`
DisableRepinning bool `json:"disable_repinning"`
FollowerMode bool `json:"follower_mode,omitempty"`
PeerstoreFile string `json:"peerstore_file,omitempty"`
PeerAddresses []string `json:"peer_addresses"`
}
// connMgrConfigJSON configures the libp2p host connection manager.
type connMgrConfigJSON struct {
HighWater int `json:"high_water"`
LowWater int `json:"low_water"`
GracePeriod string `json:"grace_period"`
}
// ConfigKey returns a human-readable string to identify
// a cluster Config.
func (cfg *Config) ConfigKey() string {
return configKey
}
// Default fills in all the Config fields with
// default working values. This means, it will
// generate a Secret.
func (cfg *Config) Default() error {
cfg.setDefaults()
clusterSecret := make([]byte, 32)
n, err := rand.Read(clusterSecret)
if err != nil {
return err
}
if n != 32 {
return errors.New("did not generate 32-byte secret")
}
cfg.Secret = clusterSecret
return nil
}
// ApplyEnvVars fills in any Config fields found
// as environment variables.
func (cfg *Config) ApplyEnvVars() error {
jcfg, err := cfg.toConfigJSON()
if err != nil {
return err
}
err = envconfig.Process(cfg.ConfigKey(), jcfg)
if err != nil {
return err
}
return cfg.applyConfigJSON(jcfg)
}
// Validate will check that the values of this config
// seem to be working ones.
func (cfg *Config) Validate() error {
if cfg.ListenAddr == nil {
return errors.New("cluster.listen_multiaddress is undefined")
}
if len(cfg.ListenAddr) == 0 {
return errors.New("cluster.listen_multiaddress is empty")
}
if cfg.ConnMgr.LowWater <= 0 {
return errors.New("cluster.connection_manager.low_water is invalid")
}
if cfg.ConnMgr.HighWater <= 0 {
return errors.New("cluster.connection_manager.high_water is invalid")
}
if cfg.ConnMgr.LowWater > cfg.ConnMgr.HighWater {
return errors.New("cluster.connection_manager.low_water is greater than high_water")
}
if cfg.ConnMgr.GracePeriod == 0 {
return errors.New("cluster.connection_manager.grace_period is invalid")
}
if cfg.DialPeerTimeout <= 0 {
return errors.New("cluster.dial_peer_timeout is invalid")
}
if cfg.StateSyncInterval <= 0 {
return errors.New("cluster.state_sync_interval is invalid")
}
if cfg.PinRecoverInterval <= 0 {
return errors.New("cluster.pin_recover_interval is invalid")
}
if cfg.MonitorPingInterval <= 0 {
return errors.New("cluster.monitoring_interval is invalid")
}
if cfg.PeerWatchInterval <= 0 {
return errors.New("cluster.peer_watch_interval is invalid")
}
rfMax := cfg.ReplicationFactorMax
rfMin := cfg.ReplicationFactorMin
if err := isReplicationFactorValid(rfMin, rfMax); err != nil {
return err
}
return isRPCPolicyValid(cfg.RPCPolicy)
}
func isReplicationFactorValid(rplMin, rplMax int) error {
// check Max and Min are correct
if rplMin == 0 || rplMax == 0 {
return errors.New("cluster.replication_factor_min and max must be set")
}
if rplMin > rplMax {
return errors.New("cluster.replication_factor_min is larger than max")
}
if rplMin < -1 {
return errors.New("cluster.replication_factor_min is wrong")
}
if rplMax < -1 {
return errors.New("cluster.replication_factor_max is wrong")
}
if (rplMin == -1 && rplMax != -1) || (rplMin != -1 && rplMax == -1) {
return errors.New("cluster.replication_factor_min and max must be -1 when one of them is")
}
return nil
}
func isRPCPolicyValid(p map[string]RPCEndpointType) error {
rpcComponents := []interface{}{
&ClusterRPCAPI{},
&PinTrackerRPCAPI{},
&IPFSConnectorRPCAPI{},
&ConsensusRPCAPI{},
&PeerMonitorRPCAPI{},
}
total := 0
for _, c := range rpcComponents {
t := reflect.TypeOf(c)
for i := 0; i < t.NumMethod(); i++ {
total++
method := t.Method(i)
name := fmt.Sprintf("%s.%s", RPCServiceID(c), method.Name)
_, ok := p[name]
if !ok {
return fmt.Errorf("RPCPolicy is missing the %s method", name)
}
}
}
if len(p) != total {
logger.Warn("defined RPC policy has more entries than needed")
}
return nil
}
// this just sets non-generated defaults
func (cfg *Config) setDefaults() {
hostname, err := os.Hostname()
if err != nil {
hostname = ""
}
cfg.Peername = hostname
listenAddrs := []ma.Multiaddr{}
for _, m := range DefaultListenAddrs {
addr, _ := ma.NewMultiaddr(m)
listenAddrs = append(listenAddrs, addr)
}
cfg.ListenAddr = listenAddrs
cfg.EnableRelayHop = DefaultEnableRelayHop
cfg.ConnMgr = ConnMgrConfig{
HighWater: DefaultConnMgrHighWater,
LowWater: DefaultConnMgrLowWater,
GracePeriod: DefaultConnMgrGracePeriod,
}
cfg.DialPeerTimeout = DefaultDialPeerTimeout
cfg.LeaveOnShutdown = DefaultLeaveOnShutdown
cfg.StateSyncInterval = DefaultStateSyncInterval
cfg.PinRecoverInterval = DefaultPinRecoverInterval
cfg.ReplicationFactorMin = DefaultReplicationFactor
cfg.ReplicationFactorMax = DefaultReplicationFactor
cfg.MonitorPingInterval = DefaultMonitorPingInterval
cfg.PeerWatchInterval = DefaultPeerWatchInterval
cfg.MDNSInterval = DefaultMDNSInterval
cfg.PinOnlyOnTrustedPeers = DefaultPinOnlyOnTrustedPeers
cfg.DisableRepinning = DefaultDisableRepinning
cfg.FollowerMode = DefaultFollowerMode
cfg.PeerstoreFile = "" // empty so it gets omitted.
cfg.PeerAddresses = []ma.Multiaddr{}
cfg.RPCPolicy = DefaultRPCPolicy
}
// LoadJSON receives a raw json-formatted configuration and
// sets the Config fields from it. Note that it should be JSON
// as generated by ToJSON().
func (cfg *Config) LoadJSON(raw []byte) error {
jcfg := &configJSON{}
err := json.Unmarshal(raw, jcfg)
if err != nil {
logger.Error("Error unmarshaling cluster config")
return err
}
cfg.setDefaults()
return cfg.applyConfigJSON(jcfg)
}
func (cfg *Config) applyConfigJSON(jcfg *configJSON) error {
config.SetIfNotDefault(jcfg.PeerstoreFile, &cfg.PeerstoreFile)
config.SetIfNotDefault(jcfg.Peername, &cfg.Peername)
clusterSecret, err := DecodeClusterSecret(jcfg.Secret)
if err != nil {
err = fmt.Errorf("error loading cluster secret from config: %s", err)
return err
}
cfg.Secret = clusterSecret
var listenAddrs []ma.Multiaddr
for _, addr := range jcfg.ListenMultiaddress {
listenAddr, err := ma.NewMultiaddr(addr)
if err != nil {
err = fmt.Errorf("error parsing a listen_multiaddress: %s", err)
return err
}
listenAddrs = append(listenAddrs, listenAddr)
}
cfg.ListenAddr = listenAddrs
cfg.EnableRelayHop = jcfg.EnableRelayHop
if conman := jcfg.ConnectionManager; conman != nil {
cfg.ConnMgr = ConnMgrConfig{
HighWater: jcfg.ConnectionManager.HighWater,
LowWater: jcfg.ConnectionManager.LowWater,
}
err = config.ParseDurations("cluster",
&config.DurationOpt{Duration: jcfg.ConnectionManager.GracePeriod, Dst: &cfg.ConnMgr.GracePeriod, Name: "connection_manager.grace_period"},
)
if err != nil {
return err
}
}
rplMin := jcfg.ReplicationFactorMin
rplMax := jcfg.ReplicationFactorMax
config.SetIfNotDefault(rplMin, &cfg.ReplicationFactorMin)
config.SetIfNotDefault(rplMax, &cfg.ReplicationFactorMax)
err = config.ParseDurations("cluster",
&config.DurationOpt{Duration: jcfg.DialPeerTimeout, Dst: &cfg.DialPeerTimeout, Name: "dial_peer_timeout"},
&config.DurationOpt{Duration: jcfg.StateSyncInterval, Dst: &cfg.StateSyncInterval, Name: "state_sync_interval"},
&config.DurationOpt{Duration: jcfg.PinRecoverInterval, Dst: &cfg.PinRecoverInterval, Name: "pin_recover_interval"},
&config.DurationOpt{Duration: jcfg.MonitorPingInterval, Dst: &cfg.MonitorPingInterval, Name: "monitor_ping_interval"},
&config.DurationOpt{Duration: jcfg.PeerWatchInterval, Dst: &cfg.PeerWatchInterval, Name: "peer_watch_interval"},
&config.DurationOpt{Duration: jcfg.MDNSInterval, Dst: &cfg.MDNSInterval, Name: "mdns_interval"},
)
if err != nil {
return err
}
// PeerAddresses
peerAddrs := []ma.Multiaddr{}
for _, addr := range jcfg.PeerAddresses {
peerAddr, err := ma.NewMultiaddr(addr)
if err != nil {
err = fmt.Errorf("error parsing peer_addresses: %s", err)
return err
}
peerAddrs = append(peerAddrs, peerAddr)
}
cfg.PeerAddresses = peerAddrs
cfg.LeaveOnShutdown = jcfg.LeaveOnShutdown
cfg.PinOnlyOnTrustedPeers = jcfg.PinOnlyOnTrustedPeers
cfg.DisableRepinning = jcfg.DisableRepinning
cfg.FollowerMode = jcfg.FollowerMode
return cfg.Validate()
}
// ToJSON generates a human-friendly version of Config.
func (cfg *Config) ToJSON() (raw []byte, err error) {
jcfg, err := cfg.toConfigJSON()
if err != nil {
return
}
raw, err = json.MarshalIndent(jcfg, "", " ")
return
}
func (cfg *Config) toConfigJSON() (jcfg *configJSON, err error) {
// Multiaddress String() may panic
defer func() {
if r := recover(); r != nil {
err = fmt.Errorf("%s", r)
}
}()
jcfg = &configJSON{}
// Set all configuration fields
jcfg.Peername = cfg.Peername
jcfg.Secret = EncodeProtectorKey(cfg.Secret)
jcfg.ReplicationFactorMin = cfg.ReplicationFactorMin
jcfg.ReplicationFactorMax = cfg.ReplicationFactorMax
jcfg.LeaveOnShutdown = cfg.LeaveOnShutdown
var listenAddrs ipfsconfig.Strings
for _, addr := range cfg.ListenAddr {
listenAddrs = append(listenAddrs, addr.String())
}
jcfg.ListenMultiaddress = ipfsconfig.Strings(listenAddrs)
jcfg.EnableRelayHop = cfg.EnableRelayHop
jcfg.ConnectionManager = &connMgrConfigJSON{
HighWater: cfg.ConnMgr.HighWater,
LowWater: cfg.ConnMgr.LowWater,
GracePeriod: cfg.ConnMgr.GracePeriod.String(),
}
jcfg.DialPeerTimeout = cfg.DialPeerTimeout.String()
jcfg.StateSyncInterval = cfg.StateSyncInterval.String()
jcfg.PinRecoverInterval = cfg.PinRecoverInterval.String()
jcfg.MonitorPingInterval = cfg.MonitorPingInterval.String()
jcfg.PeerWatchInterval = cfg.PeerWatchInterval.String()
jcfg.MDNSInterval = cfg.MDNSInterval.String()
jcfg.PinOnlyOnTrustedPeers = cfg.PinOnlyOnTrustedPeers
jcfg.DisableRepinning = cfg.DisableRepinning
jcfg.PeerstoreFile = cfg.PeerstoreFile
jcfg.PeerAddresses = []string{}
for _, addr := range cfg.PeerAddresses {
jcfg.PeerAddresses = append(jcfg.PeerAddresses, addr.String())
}
jcfg.FollowerMode = cfg.FollowerMode
return
}
// GetPeerstorePath returns the full path of the
// PeerstoreFile, obtained by concatenating that value
// with BaseDir of the configuration, if set.
// An empty string is returned when BaseDir is not set.
func (cfg *Config) GetPeerstorePath() string {
if cfg.BaseDir == "" {
return ""
}
filename := DefaultPeerstoreFile
if cfg.PeerstoreFile != "" {
filename = cfg.PeerstoreFile
}
return filepath.Join(cfg.BaseDir, filename)
}
// ToDisplayJSON returns JSON config as a string.
func (cfg *Config) ToDisplayJSON() ([]byte, error) {
jcfg, err := cfg.toConfigJSON()
if err != nil {
return nil, err
}
return config.DisplayJSON(jcfg)
}
// DecodeClusterSecret parses a hex-encoded string, checks that it is exactly
// 32 bytes long and returns its value as a byte-slice.x
func DecodeClusterSecret(hexSecret string) ([]byte, error) {
secret, err := hex.DecodeString(hexSecret)
if err != nil {
return nil, err
}
switch secretLen := len(secret); secretLen {
case 0:
logger.Warn("Cluster secret is empty, cluster will start on unprotected network.")
return nil, nil
case 32:
return secret, nil
default:
return nil, fmt.Errorf("input secret is %d bytes, cluster secret should be 32", secretLen)
}
}