open-vault/vault/core.go

2036 lines
62 KiB
Go
Raw Normal View History

2015-03-09 23:33:27 +00:00
package vault
import (
"context"
"crypto/ecdsa"
"crypto/subtle"
"crypto/x509"
Raft Storage Backend (#6888) * Work on raft backend * Add logstore locally * Add encryptor and unsealable interfaces * Add clustering support to raft * Remove client and handler * Bootstrap raft on init * Cleanup raft logic a bit * More raft work * Work on TLS config * More work on bootstrapping * Fix build * More work on bootstrapping * More bootstrapping work * fix build * Remove consul dep * Fix build * merged oss/master into raft-storage * Work on bootstrapping * Get bootstrapping to work * Clean up FMS and node-id * Update local node ID logic * Cleanup node-id change * Work on snapshotting * Raft: Add remove peer API (#906) * Add remove peer API * Add some comments * Fix existing snapshotting (#909) * Raft get peers API (#912) * Read raft configuration * address review feedback * Use the Leadership Transfer API to step-down the active node (#918) * Raft join and unseal using Shamir keys (#917) * Raft join using shamir * Store AEAD instead of master key * Split the raft join process to answer the challenge after a successful unseal * get the follower to standby state * Make unseal work * minor changes * Some input checks * reuse the shamir seal access instead of new default seal access * refactor joinRaftSendAnswer function * Synchronously send answer in auto-unseal case * Address review feedback * Raft snapshots (#910) * Fix existing snapshotting * implement the noop snapshotting * Add comments and switch log libraries * add some snapshot tests * add snapshot test file * add TODO * More work on raft snapshotting * progress on the ConfigStore strategy * Don't use two buckets * Update the snapshot store logic to hide the file logic * Add more backend tests * Cleanup code a bit * [WIP] Raft recovery (#938) * Add recovery functionality * remove fmt.Printfs * Fix a few fsm bugs * Add max size value for raft backend (#942) * Add max size value for raft backend * Include physical.ErrValueTooLarge in the message * Raft snapshot Take/Restore API (#926) * Inital work on raft snapshot APIs * Always redirect snapshot install/download requests * More work on the snapshot APIs * Cleanup code a bit * On restore handle special cases * Use the seal to encrypt the sha sum file * Add sealer mechanism and fix some bugs * Call restore while state lock is held * Send restore cb trigger through raft log * Make error messages nicer * Add test helpers * Add snapshot test * Add shamir unseal test * Add more raft snapshot API tests * Fix locking * Change working to initalize * Add underlying raw object to test cluster core * Move leaderUUID to core * Add raft TLS rotation logic (#950) * Add TLS rotation logic * Cleanup logic a bit * Add/Remove from follower state on add/remove peer * add comments * Update more comments * Update request_forwarding_service.proto * Make sure we populate all nodes in the followerstate obj * Update times * Apply review feedback * Add more raft config setting (#947) * Add performance config setting * Add more config options and fix tests * Test Raft Recovery (#944) * Test raft recovery * Leave out a node during recovery * remove unused struct * Update physical/raft/snapshot_test.go * Update physical/raft/snapshot_test.go * fix vendoring * Switch to new raft interface * Remove unused files * Switch a gogo -> proto instance * Remove unneeded vault dep in go.sum * Update helper/testhelpers/testhelpers.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * Update vault/cluster/cluster.go * track active key within the keyring itself (#6915) * track active key within the keyring itself * lookup and store using the active key ID * update docstring * minor refactor * Small text fixes (#6912) * Update physical/raft/raft.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * review feedback * Move raft logical system into separate file * Update help text a bit * Enforce cluster addr is set and use it for raft bootstrapping * Fix tests * fix http test panic * Pull in latest raft-snapshot library * Add comment
2019-06-20 19:14:58 +00:00
"encoding/base64"
2015-03-10 00:45:34 +00:00
"errors"
2015-03-09 23:33:27 +00:00
"fmt"
"net"
"net/http"
"net/url"
2017-04-04 00:52:29 +00:00
"path/filepath"
2015-03-10 00:45:34 +00:00
"sync"
2017-10-23 20:06:27 +00:00
"sync/atomic"
2015-04-08 23:43:17 +00:00
"time"
2015-03-09 23:33:27 +00:00
"github.com/armon/go-metrics"
"github.com/hashicorp/errwrap"
log "github.com/hashicorp/go-hclog"
"github.com/hashicorp/go-multierror"
"github.com/hashicorp/go-uuid"
"github.com/hashicorp/vault/api"
"github.com/hashicorp/vault/audit"
"github.com/hashicorp/vault/command/server"
"github.com/hashicorp/vault/helper/metricsutil"
"github.com/hashicorp/vault/helper/namespace"
"github.com/hashicorp/vault/helper/reload"
"github.com/hashicorp/vault/physical/raft"
"github.com/hashicorp/vault/sdk/helper/certutil"
"github.com/hashicorp/vault/sdk/helper/consts"
"github.com/hashicorp/vault/sdk/helper/jsonutil"
"github.com/hashicorp/vault/sdk/helper/logging"
"github.com/hashicorp/vault/sdk/helper/mlock"
Raft Storage Backend (#6888) * Work on raft backend * Add logstore locally * Add encryptor and unsealable interfaces * Add clustering support to raft * Remove client and handler * Bootstrap raft on init * Cleanup raft logic a bit * More raft work * Work on TLS config * More work on bootstrapping * Fix build * More work on bootstrapping * More bootstrapping work * fix build * Remove consul dep * Fix build * merged oss/master into raft-storage * Work on bootstrapping * Get bootstrapping to work * Clean up FMS and node-id * Update local node ID logic * Cleanup node-id change * Work on snapshotting * Raft: Add remove peer API (#906) * Add remove peer API * Add some comments * Fix existing snapshotting (#909) * Raft get peers API (#912) * Read raft configuration * address review feedback * Use the Leadership Transfer API to step-down the active node (#918) * Raft join and unseal using Shamir keys (#917) * Raft join using shamir * Store AEAD instead of master key * Split the raft join process to answer the challenge after a successful unseal * get the follower to standby state * Make unseal work * minor changes * Some input checks * reuse the shamir seal access instead of new default seal access * refactor joinRaftSendAnswer function * Synchronously send answer in auto-unseal case * Address review feedback * Raft snapshots (#910) * Fix existing snapshotting * implement the noop snapshotting * Add comments and switch log libraries * add some snapshot tests * add snapshot test file * add TODO * More work on raft snapshotting * progress on the ConfigStore strategy * Don't use two buckets * Update the snapshot store logic to hide the file logic * Add more backend tests * Cleanup code a bit * [WIP] Raft recovery (#938) * Add recovery functionality * remove fmt.Printfs * Fix a few fsm bugs * Add max size value for raft backend (#942) * Add max size value for raft backend * Include physical.ErrValueTooLarge in the message * Raft snapshot Take/Restore API (#926) * Inital work on raft snapshot APIs * Always redirect snapshot install/download requests * More work on the snapshot APIs * Cleanup code a bit * On restore handle special cases * Use the seal to encrypt the sha sum file * Add sealer mechanism and fix some bugs * Call restore while state lock is held * Send restore cb trigger through raft log * Make error messages nicer * Add test helpers * Add snapshot test * Add shamir unseal test * Add more raft snapshot API tests * Fix locking * Change working to initalize * Add underlying raw object to test cluster core * Move leaderUUID to core * Add raft TLS rotation logic (#950) * Add TLS rotation logic * Cleanup logic a bit * Add/Remove from follower state on add/remove peer * add comments * Update more comments * Update request_forwarding_service.proto * Make sure we populate all nodes in the followerstate obj * Update times * Apply review feedback * Add more raft config setting (#947) * Add performance config setting * Add more config options and fix tests * Test Raft Recovery (#944) * Test raft recovery * Leave out a node during recovery * remove unused struct * Update physical/raft/snapshot_test.go * Update physical/raft/snapshot_test.go * fix vendoring * Switch to new raft interface * Remove unused files * Switch a gogo -> proto instance * Remove unneeded vault dep in go.sum * Update helper/testhelpers/testhelpers.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * Update vault/cluster/cluster.go * track active key within the keyring itself (#6915) * track active key within the keyring itself * lookup and store using the active key ID * update docstring * minor refactor * Small text fixes (#6912) * Update physical/raft/raft.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * review feedback * Move raft logical system into separate file * Update help text a bit * Enforce cluster addr is set and use it for raft bootstrapping * Fix tests * fix http test panic * Pull in latest raft-snapshot library * Add comment
2019-06-20 19:14:58 +00:00
"github.com/hashicorp/vault/sdk/helper/strutil"
2019-04-12 22:26:54 +00:00
"github.com/hashicorp/vault/sdk/helper/tlsutil"
"github.com/hashicorp/vault/sdk/logical"
"github.com/hashicorp/vault/sdk/physical"
2015-03-11 18:34:08 +00:00
"github.com/hashicorp/vault/shamir"
"github.com/hashicorp/vault/vault/cluster"
"github.com/hashicorp/vault/vault/seal"
Raft Storage Backend (#6888) * Work on raft backend * Add logstore locally * Add encryptor and unsealable interfaces * Add clustering support to raft * Remove client and handler * Bootstrap raft on init * Cleanup raft logic a bit * More raft work * Work on TLS config * More work on bootstrapping * Fix build * More work on bootstrapping * More bootstrapping work * fix build * Remove consul dep * Fix build * merged oss/master into raft-storage * Work on bootstrapping * Get bootstrapping to work * Clean up FMS and node-id * Update local node ID logic * Cleanup node-id change * Work on snapshotting * Raft: Add remove peer API (#906) * Add remove peer API * Add some comments * Fix existing snapshotting (#909) * Raft get peers API (#912) * Read raft configuration * address review feedback * Use the Leadership Transfer API to step-down the active node (#918) * Raft join and unseal using Shamir keys (#917) * Raft join using shamir * Store AEAD instead of master key * Split the raft join process to answer the challenge after a successful unseal * get the follower to standby state * Make unseal work * minor changes * Some input checks * reuse the shamir seal access instead of new default seal access * refactor joinRaftSendAnswer function * Synchronously send answer in auto-unseal case * Address review feedback * Raft snapshots (#910) * Fix existing snapshotting * implement the noop snapshotting * Add comments and switch log libraries * add some snapshot tests * add snapshot test file * add TODO * More work on raft snapshotting * progress on the ConfigStore strategy * Don't use two buckets * Update the snapshot store logic to hide the file logic * Add more backend tests * Cleanup code a bit * [WIP] Raft recovery (#938) * Add recovery functionality * remove fmt.Printfs * Fix a few fsm bugs * Add max size value for raft backend (#942) * Add max size value for raft backend * Include physical.ErrValueTooLarge in the message * Raft snapshot Take/Restore API (#926) * Inital work on raft snapshot APIs * Always redirect snapshot install/download requests * More work on the snapshot APIs * Cleanup code a bit * On restore handle special cases * Use the seal to encrypt the sha sum file * Add sealer mechanism and fix some bugs * Call restore while state lock is held * Send restore cb trigger through raft log * Make error messages nicer * Add test helpers * Add snapshot test * Add shamir unseal test * Add more raft snapshot API tests * Fix locking * Change working to initalize * Add underlying raw object to test cluster core * Move leaderUUID to core * Add raft TLS rotation logic (#950) * Add TLS rotation logic * Cleanup logic a bit * Add/Remove from follower state on add/remove peer * add comments * Update more comments * Update request_forwarding_service.proto * Make sure we populate all nodes in the followerstate obj * Update times * Apply review feedback * Add more raft config setting (#947) * Add performance config setting * Add more config options and fix tests * Test Raft Recovery (#944) * Test raft recovery * Leave out a node during recovery * remove unused struct * Update physical/raft/snapshot_test.go * Update physical/raft/snapshot_test.go * fix vendoring * Switch to new raft interface * Remove unused files * Switch a gogo -> proto instance * Remove unneeded vault dep in go.sum * Update helper/testhelpers/testhelpers.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * Update vault/cluster/cluster.go * track active key within the keyring itself (#6915) * track active key within the keyring itself * lookup and store using the active key ID * update docstring * minor refactor * Small text fixes (#6912) * Update physical/raft/raft.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * review feedback * Move raft logical system into separate file * Update help text a bit * Enforce cluster addr is set and use it for raft bootstrapping * Fix tests * fix http test panic * Pull in latest raft-snapshot library * Add comment
2019-06-20 19:14:58 +00:00
shamirseal "github.com/hashicorp/vault/vault/seal/shamir"
"github.com/patrickmn/go-cache"
"google.golang.org/grpc"
2015-03-09 23:33:27 +00:00
)
2015-03-10 00:45:34 +00:00
const (
// CoreLockPath is the path used to acquire a coordinating lock
2015-04-14 21:06:15 +00:00
// for a highly-available deploy.
CoreLockPath = "core/lock"
2015-04-14 21:06:15 +00:00
2017-03-03 20:00:46 +00:00
// The poison pill is used as a check during certain scenarios to indicate
// to standby nodes that they should seal
poisonPillPath = "core/poison-pill"
2015-04-14 23:44:48 +00:00
// coreLeaderPrefix is the prefix used for the UUID that contains
// the currently elected leader.
coreLeaderPrefix = "core/leader/"
2017-10-23 20:03:36 +00:00
// knownPrimaryAddrsPrefix is used to store last-known cluster address
// information for primaries
knownPrimaryAddrsPrefix = "core/primary-addrs/"
// coreKeyringCanaryPath is used as a canary to indicate to replicated
// clusters that they need to perform a rekey operation synchronously; this
// isn't keyring-canary to avoid ignoring it when ignoring core/keyring
coreKeyringCanaryPath = "core/canary-keyring"
2015-03-10 00:45:34 +00:00
)
var (
// ErrAlreadyInit is returned if the core is already
// initialized. This prevents a re-initialization.
ErrAlreadyInit = errors.New("Vault is already initialized")
// ErrNotInit is returned if a non-initialized barrier
// is attempted to be unsealed.
ErrNotInit = errors.New("Vault is not initialized")
// ErrInternalError is returned when we don't want to leak
// any information about an internal error
ErrInternalError = errors.New("internal error")
2015-04-14 23:53:40 +00:00
// ErrHANotEnabled is returned if the operation only makes sense
// in an HA setting
ErrHANotEnabled = errors.New("Vault is not configured for highly-available mode")
// manualStepDownSleepPeriod is how long to sleep after a user-initiated
// step down of the active node, to prevent instantly regrabbing the lock.
// It's var not const so that tests can manipulate it.
manualStepDownSleepPeriod = 10 * time.Second
// Functions only in the Enterprise version
enterprisePostUnseal = enterprisePostUnsealImpl
enterprisePreSeal = enterprisePreSealImpl
startReplication = startReplicationImpl
stopReplication = stopReplicationImpl
LastWAL = lastWALImpl
LastPerformanceWAL = lastPerformanceWALImpl
PerformanceMerkleRoot = merkleRootImpl
DRMerkleRoot = merkleRootImpl
LastRemoteWAL = lastRemoteWALImpl
WaitUntilWALShipped = waitUntilWALShippedImpl
2015-03-10 00:45:34 +00:00
)
2016-04-04 14:44:22 +00:00
// NonFatalError is an error that can be returned during NewCore that should be
// displayed but not cause a program exit
type NonFatalError struct {
Err error
2015-03-09 23:33:27 +00:00
}
2016-04-04 14:44:22 +00:00
func (e *NonFatalError) WrappedErrors() []error {
return []error{e.Err}
2015-03-10 00:45:34 +00:00
}
2016-04-04 14:44:22 +00:00
func (e *NonFatalError) Error() string {
return e.Err.Error()
2015-03-10 00:45:34 +00:00
}
// NewNonFatalError returns a new non-fatal error.
func NewNonFatalError(err error) *NonFatalError {
return &NonFatalError{Err: err}
}
// IsFatalError returns true if the given error is a fatal error.
func IsFatalError(err error) bool {
return !errwrap.ContainsType(err, new(NonFatalError))
}
// ErrInvalidKey is returned if there is a user-based error with a provided
// unseal key. This will be shown to the user, so should not contain
// information that is sensitive.
2015-03-12 18:20:27 +00:00
type ErrInvalidKey struct {
Reason string
}
func (e *ErrInvalidKey) Error() string {
return fmt.Sprintf("invalid key: %v", e.Reason)
}
2018-09-18 03:03:00 +00:00
type RegisterAuthFunc func(context.Context, time.Duration, string, *logical.Auth) error
type activeAdvertisement struct {
RedirectAddr string `json:"redirect_addr"`
ClusterAddr string `json:"cluster_addr,omitempty"`
ClusterCert []byte `json:"cluster_cert,omitempty"`
ClusterKeyParams *certutil.ClusterKeyParams `json:"cluster_key_params,omitempty"`
}
type unlockInformation struct {
Parts [][]byte
Nonce string
}
type raftInformation struct {
challenge *physical.EncryptedBlobInfo
leaderClient *api.Client
leaderBarrierConfig *SealConfig
nonVoter bool
}
2015-03-09 23:33:27 +00:00
// Core is used as the central manager of Vault activity. It is the primary point of
// interface for API handlers and is responsible for managing the logical and physical
// backends, router, security barrier, and audit trails.
type Core struct {
2018-09-18 03:03:00 +00:00
entCore
2018-11-07 01:21:24 +00:00
// The registry of builtin plugins is passed in here as an interface because
// if it's used directly, it results in import cycles.
builtinRegistry BuiltinRegistry
// N.B.: This is used to populate a dev token down replication, as
// otherwise, after replication is started, a dev would have to go through
// the generate-root process simply to talk to the new follower cluster.
devToken string
2015-04-14 21:06:15 +00:00
// HABackend may be available depending on the physical backend
ha physical.HABackend
// storageType is the the storage type set in the storage configuration
storageType string
// redirectAddr is the address we advertise as leader if held
redirectAddr string
// clusterAddr is the address we use for clustering
clusterAddr *atomic.Value
2015-04-14 23:44:48 +00:00
2015-03-09 23:33:27 +00:00
// physical backend is the un-trusted backend with durable data
physical physical.Backend
Raft Storage Backend (#6888) * Work on raft backend * Add logstore locally * Add encryptor and unsealable interfaces * Add clustering support to raft * Remove client and handler * Bootstrap raft on init * Cleanup raft logic a bit * More raft work * Work on TLS config * More work on bootstrapping * Fix build * More work on bootstrapping * More bootstrapping work * fix build * Remove consul dep * Fix build * merged oss/master into raft-storage * Work on bootstrapping * Get bootstrapping to work * Clean up FMS and node-id * Update local node ID logic * Cleanup node-id change * Work on snapshotting * Raft: Add remove peer API (#906) * Add remove peer API * Add some comments * Fix existing snapshotting (#909) * Raft get peers API (#912) * Read raft configuration * address review feedback * Use the Leadership Transfer API to step-down the active node (#918) * Raft join and unseal using Shamir keys (#917) * Raft join using shamir * Store AEAD instead of master key * Split the raft join process to answer the challenge after a successful unseal * get the follower to standby state * Make unseal work * minor changes * Some input checks * reuse the shamir seal access instead of new default seal access * refactor joinRaftSendAnswer function * Synchronously send answer in auto-unseal case * Address review feedback * Raft snapshots (#910) * Fix existing snapshotting * implement the noop snapshotting * Add comments and switch log libraries * add some snapshot tests * add snapshot test file * add TODO * More work on raft snapshotting * progress on the ConfigStore strategy * Don't use two buckets * Update the snapshot store logic to hide the file logic * Add more backend tests * Cleanup code a bit * [WIP] Raft recovery (#938) * Add recovery functionality * remove fmt.Printfs * Fix a few fsm bugs * Add max size value for raft backend (#942) * Add max size value for raft backend * Include physical.ErrValueTooLarge in the message * Raft snapshot Take/Restore API (#926) * Inital work on raft snapshot APIs * Always redirect snapshot install/download requests * More work on the snapshot APIs * Cleanup code a bit * On restore handle special cases * Use the seal to encrypt the sha sum file * Add sealer mechanism and fix some bugs * Call restore while state lock is held * Send restore cb trigger through raft log * Make error messages nicer * Add test helpers * Add snapshot test * Add shamir unseal test * Add more raft snapshot API tests * Fix locking * Change working to initalize * Add underlying raw object to test cluster core * Move leaderUUID to core * Add raft TLS rotation logic (#950) * Add TLS rotation logic * Cleanup logic a bit * Add/Remove from follower state on add/remove peer * add comments * Update more comments * Update request_forwarding_service.proto * Make sure we populate all nodes in the followerstate obj * Update times * Apply review feedback * Add more raft config setting (#947) * Add performance config setting * Add more config options and fix tests * Test Raft Recovery (#944) * Test raft recovery * Leave out a node during recovery * remove unused struct * Update physical/raft/snapshot_test.go * Update physical/raft/snapshot_test.go * fix vendoring * Switch to new raft interface * Remove unused files * Switch a gogo -> proto instance * Remove unneeded vault dep in go.sum * Update helper/testhelpers/testhelpers.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * Update vault/cluster/cluster.go * track active key within the keyring itself (#6915) * track active key within the keyring itself * lookup and store using the active key ID * update docstring * minor refactor * Small text fixes (#6912) * Update physical/raft/raft.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * review feedback * Move raft logical system into separate file * Update help text a bit * Enforce cluster addr is set and use it for raft bootstrapping * Fix tests * fix http test panic * Pull in latest raft-snapshot library * Add comment
2019-06-20 19:14:58 +00:00
// underlyingPhysical will always point to the underlying backend
// implementation. This is an un-trusted backend with durable data
underlyingPhysical physical.Backend
2018-10-23 06:34:02 +00:00
// seal is our seal, for seal configuration information
2016-04-04 14:44:22 +00:00
seal Seal
// raftInfo will contain information required for this node to join as a
// peer to an existing raft cluster
raftInfo *raftInformation
Raft Storage Backend (#6888) * Work on raft backend * Add logstore locally * Add encryptor and unsealable interfaces * Add clustering support to raft * Remove client and handler * Bootstrap raft on init * Cleanup raft logic a bit * More raft work * Work on TLS config * More work on bootstrapping * Fix build * More work on bootstrapping * More bootstrapping work * fix build * Remove consul dep * Fix build * merged oss/master into raft-storage * Work on bootstrapping * Get bootstrapping to work * Clean up FMS and node-id * Update local node ID logic * Cleanup node-id change * Work on snapshotting * Raft: Add remove peer API (#906) * Add remove peer API * Add some comments * Fix existing snapshotting (#909) * Raft get peers API (#912) * Read raft configuration * address review feedback * Use the Leadership Transfer API to step-down the active node (#918) * Raft join and unseal using Shamir keys (#917) * Raft join using shamir * Store AEAD instead of master key * Split the raft join process to answer the challenge after a successful unseal * get the follower to standby state * Make unseal work * minor changes * Some input checks * reuse the shamir seal access instead of new default seal access * refactor joinRaftSendAnswer function * Synchronously send answer in auto-unseal case * Address review feedback * Raft snapshots (#910) * Fix existing snapshotting * implement the noop snapshotting * Add comments and switch log libraries * add some snapshot tests * add snapshot test file * add TODO * More work on raft snapshotting * progress on the ConfigStore strategy * Don't use two buckets * Update the snapshot store logic to hide the file logic * Add more backend tests * Cleanup code a bit * [WIP] Raft recovery (#938) * Add recovery functionality * remove fmt.Printfs * Fix a few fsm bugs * Add max size value for raft backend (#942) * Add max size value for raft backend * Include physical.ErrValueTooLarge in the message * Raft snapshot Take/Restore API (#926) * Inital work on raft snapshot APIs * Always redirect snapshot install/download requests * More work on the snapshot APIs * Cleanup code a bit * On restore handle special cases * Use the seal to encrypt the sha sum file * Add sealer mechanism and fix some bugs * Call restore while state lock is held * Send restore cb trigger through raft log * Make error messages nicer * Add test helpers * Add snapshot test * Add shamir unseal test * Add more raft snapshot API tests * Fix locking * Change working to initalize * Add underlying raw object to test cluster core * Move leaderUUID to core * Add raft TLS rotation logic (#950) * Add TLS rotation logic * Cleanup logic a bit * Add/Remove from follower state on add/remove peer * add comments * Update more comments * Update request_forwarding_service.proto * Make sure we populate all nodes in the followerstate obj * Update times * Apply review feedback * Add more raft config setting (#947) * Add performance config setting * Add more config options and fix tests * Test Raft Recovery (#944) * Test raft recovery * Leave out a node during recovery * remove unused struct * Update physical/raft/snapshot_test.go * Update physical/raft/snapshot_test.go * fix vendoring * Switch to new raft interface * Remove unused files * Switch a gogo -> proto instance * Remove unneeded vault dep in go.sum * Update helper/testhelpers/testhelpers.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * Update vault/cluster/cluster.go * track active key within the keyring itself (#6915) * track active key within the keyring itself * lookup and store using the active key ID * update docstring * minor refactor * Small text fixes (#6912) * Update physical/raft/raft.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * review feedback * Move raft logical system into separate file * Update help text a bit * Enforce cluster addr is set and use it for raft bootstrapping * Fix tests * fix http test panic * Pull in latest raft-snapshot library * Add comment
2019-06-20 19:14:58 +00:00
2018-10-23 06:34:02 +00:00
// migrationSeal is the seal to use during a migration operation. It is the
// seal we're migrating *from*.
migrationSeal Seal
// unwrapSeal is the seal to use on Enterprise to unwrap values wrapped
// with the previous seal.
unwrapSeal Seal
2015-03-09 23:33:27 +00:00
// barrier is the security barrier wrapping the physical backend
barrier SecurityBarrier
// router is responsible for managing the mount points for logical backends.
router *Router
2015-03-10 00:45:34 +00:00
// logicalBackends is the mapping of backends to use for this core
logicalBackends map[string]logical.Factory
// credentialBackends is the mapping of backends to use for this core
credentialBackends map[string]logical.Factory
2015-03-15 23:25:38 +00:00
// auditBackends is the mapping of backends to use for this core
auditBackends map[string]audit.Factory
2015-03-10 00:45:34 +00:00
// stateLock protects mutable state
stateLock sync.RWMutex
sealed *uint32
2015-03-10 00:45:34 +00:00
standby bool
perfStandby bool
standbyDoneCh chan struct{}
standbyStopCh chan struct{}
manualStepDownCh chan struct{}
keepHALockOnStepDown *uint32
heldHALock physical.Lock
2015-04-14 21:06:15 +00:00
// unlockInfo has the keys provided to Unseal until the threshold number of parts is available, as well as the operation nonce
unlockInfo *unlockInformation
2015-03-10 00:45:34 +00:00
2016-01-15 15:55:35 +00:00
// generateRootProgress holds the shares until we reach enough
// to verify the master key
2016-01-15 15:55:35 +00:00
generateRootConfig *GenerateRootConfig
generateRootProgress [][]byte
generateRootLock sync.Mutex
2016-04-04 14:44:22 +00:00
// These variables holds the config and shares we have until we reach
// enough to verify the appropriate master key. Note that the same lock is
// used; this isn't time-critical so this shouldn't be a problem.
2018-05-21 21:46:32 +00:00
barrierRekeyConfig *SealConfig
recoveryRekeyConfig *SealConfig
rekeyLock sync.RWMutex
2015-05-28 18:40:01 +00:00
2015-03-11 22:19:41 +00:00
// mounts is loaded after unseal since it is a protected
// configuration
mounts *MountTable
2015-03-11 22:19:41 +00:00
// mountsLock is used to ensure that the mounts table does not
// change underneath a calling function
mountsLock sync.RWMutex
// auth is loaded after unseal since it is a protected
// configuration
auth *MountTable
// authLock is used to ensure that the auth table does not
// change underneath a calling function
authLock sync.RWMutex
// audit is loaded after unseal since it is a protected
// configuration
audit *MountTable
// auditLock is used to ensure that the audit table does not
// change underneath a calling function
auditLock sync.RWMutex
// auditBroker is used to ingest the audit events and fan
// out into the configured audit backends
auditBroker *AuditBroker
// auditedHeaders is used to configure which http headers
// can be output in the audit logs
auditedHeaders *AuditedHeadersConfig
// systemBackend is the backend which is used to manage internal operations
systemBackend *SystemBackend
2018-09-18 03:03:00 +00:00
// cubbyholeBackend is the backend which manages the per-token storage
cubbyholeBackend *CubbyholeBackend
// systemBarrierView is the barrier view for the system backend
systemBarrierView *BarrierView
2015-03-12 19:41:12 +00:00
// expiration manager is used for managing LeaseIDs,
// renewal, expiration and revocation
expiration *ExpirationManager
// rollback manager is used to run rollbacks periodically
rollback *RollbackManager
// policy store is used to manage named ACL policies
policyStore *PolicyStore
2015-03-23 20:41:05 +00:00
// token store is used to manage authentication tokens
tokenStore *TokenStore
// identityStore is used to manage client entities
identityStore *IdentityStore
2015-04-08 23:43:17 +00:00
// metricsCh is used to stop the metrics streaming
metricsCh chan struct{}
// metricsMutex is used to prevent a race condition between
// metrics emission and sealing leading to a nil pointer
metricsMutex sync.Mutex
defaultLeaseTTL time.Duration
maxLeaseTTL time.Duration
2018-08-23 19:04:18 +00:00
// baseLogger is used to avoid ResetNamed as it strips useful prefixes in
// e.g. testing
baseLogger log.Logger
logger log.Logger
// cachingDisabled indicates whether caches are disabled
cachingDisabled bool
// Cache stores the actual cache; we always have this but may bypass it if
// disabled
physicalCache physical.ToggleablePurgemonster
2016-09-30 04:06:40 +00:00
// reloadFuncs is a map containing reload functions
reloadFuncs map[string][]reload.ReloadFunc
2016-09-30 04:06:40 +00:00
// reloadFuncsLock controls access to the funcs
2016-09-30 04:06:40 +00:00
reloadFuncsLock sync.RWMutex
2017-01-04 21:44:03 +00:00
// wrappingJWTKey is the key used for generating JWTs containing response
// wrapping information
wrappingJWTKey *ecdsa.PrivateKey
//
// Cluster information
//
// Name
clusterName string
// Specific cipher suites to use for clustering, if any
clusterCipherSuites []uint16
// Used to modify cluster parameters
clusterParamsLock sync.RWMutex
// The private key stored in the barrier used for establishing
// mutually-authenticated connections between Vault cluster members
localClusterPrivateKey *atomic.Value
// The local cluster cert
localClusterCert *atomic.Value
// The parsed form of the local cluster cert
localClusterParsedCert *atomic.Value
2016-08-19 15:03:53 +00:00
// The TCP addresses we should use for clustering
clusterListenerAddrs []*net.TCPAddr
// The handler to use for request forwarding
clusterHandler http.Handler
// Write lock used to ensure that we don't have multiple connections adjust
// this value at the same time
requestForwardingConnectionLock sync.RWMutex
// Lock for the leader values, ensuring we don't run the parts of Leader()
// that change things concurrently
leaderParamsLock sync.RWMutex
// Current cluster leader values
clusterLeaderParams *atomic.Value
2017-05-25 00:51:53 +00:00
// Info on cluster members
2017-05-25 01:10:32 +00:00
clusterPeerClusterAddrsCache *cache.Cache
// The context for the client
rpcClientConnContext context.Context
2016-08-19 15:03:53 +00:00
// The function for canceling the client connection
rpcClientConnCancelFunc context.CancelFunc
// The grpc ClientConn for RPC calls
rpcClientConn *grpc.ClientConn
// The grpc forwarding client
rpcForwardingClient *forwardingClient
Raft Storage Backend (#6888) * Work on raft backend * Add logstore locally * Add encryptor and unsealable interfaces * Add clustering support to raft * Remove client and handler * Bootstrap raft on init * Cleanup raft logic a bit * More raft work * Work on TLS config * More work on bootstrapping * Fix build * More work on bootstrapping * More bootstrapping work * fix build * Remove consul dep * Fix build * merged oss/master into raft-storage * Work on bootstrapping * Get bootstrapping to work * Clean up FMS and node-id * Update local node ID logic * Cleanup node-id change * Work on snapshotting * Raft: Add remove peer API (#906) * Add remove peer API * Add some comments * Fix existing snapshotting (#909) * Raft get peers API (#912) * Read raft configuration * address review feedback * Use the Leadership Transfer API to step-down the active node (#918) * Raft join and unseal using Shamir keys (#917) * Raft join using shamir * Store AEAD instead of master key * Split the raft join process to answer the challenge after a successful unseal * get the follower to standby state * Make unseal work * minor changes * Some input checks * reuse the shamir seal access instead of new default seal access * refactor joinRaftSendAnswer function * Synchronously send answer in auto-unseal case * Address review feedback * Raft snapshots (#910) * Fix existing snapshotting * implement the noop snapshotting * Add comments and switch log libraries * add some snapshot tests * add snapshot test file * add TODO * More work on raft snapshotting * progress on the ConfigStore strategy * Don't use two buckets * Update the snapshot store logic to hide the file logic * Add more backend tests * Cleanup code a bit * [WIP] Raft recovery (#938) * Add recovery functionality * remove fmt.Printfs * Fix a few fsm bugs * Add max size value for raft backend (#942) * Add max size value for raft backend * Include physical.ErrValueTooLarge in the message * Raft snapshot Take/Restore API (#926) * Inital work on raft snapshot APIs * Always redirect snapshot install/download requests * More work on the snapshot APIs * Cleanup code a bit * On restore handle special cases * Use the seal to encrypt the sha sum file * Add sealer mechanism and fix some bugs * Call restore while state lock is held * Send restore cb trigger through raft log * Make error messages nicer * Add test helpers * Add snapshot test * Add shamir unseal test * Add more raft snapshot API tests * Fix locking * Change working to initalize * Add underlying raw object to test cluster core * Move leaderUUID to core * Add raft TLS rotation logic (#950) * Add TLS rotation logic * Cleanup logic a bit * Add/Remove from follower state on add/remove peer * add comments * Update more comments * Update request_forwarding_service.proto * Make sure we populate all nodes in the followerstate obj * Update times * Apply review feedback * Add more raft config setting (#947) * Add performance config setting * Add more config options and fix tests * Test Raft Recovery (#944) * Test raft recovery * Leave out a node during recovery * remove unused struct * Update physical/raft/snapshot_test.go * Update physical/raft/snapshot_test.go * fix vendoring * Switch to new raft interface * Remove unused files * Switch a gogo -> proto instance * Remove unneeded vault dep in go.sum * Update helper/testhelpers/testhelpers.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * Update vault/cluster/cluster.go * track active key within the keyring itself (#6915) * track active key within the keyring itself * lookup and store using the active key ID * update docstring * minor refactor * Small text fixes (#6912) * Update physical/raft/raft.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * review feedback * Move raft logical system into separate file * Update help text a bit * Enforce cluster addr is set and use it for raft bootstrapping * Fix tests * fix http test panic * Pull in latest raft-snapshot library * Add comment
2019-06-20 19:14:58 +00:00
// The UUID used to hold the leader lock. Only set on active node
leaderUUID string
2017-06-17 04:04:55 +00:00
// CORS Information
corsConfig *CORSConfig
2017-10-23 20:03:36 +00:00
// The active set of upstream cluster addresses; stored via the Echo
// mechanism, loaded by the balancer
atomicPrimaryClusterAddrs *atomic.Value
atomicPrimaryFailoverAddrs *atomic.Value
2018-09-18 03:03:00 +00:00
// replicationState keeps the current replication state cached for quick
// lookup; activeNodeReplicationState stores the active value on standbys
replicationState *uint32
activeNodeReplicationState *uint32
2017-02-24 15:45:29 +00:00
// uiConfig contains UI configuration
uiConfig *UIConfig
2017-04-04 00:52:29 +00:00
// rawEnabled indicates whether the Raw endpoint is enabled
rawEnabled bool
2017-04-11 00:12:52 +00:00
// pluginDirectory is the location vault will look for plugin binaries
2017-04-04 00:52:29 +00:00
pluginDirectory string
// pluginCatalog is used to manage plugin configurations
pluginCatalog *PluginCatalog
2017-04-11 00:12:52 +00:00
2017-04-24 19:21:49 +00:00
enableMlock bool
// This can be used to trigger operations to stop running when Vault is
// going to be shut down, stepped down, or sealed
activeContext context.Context
activeContextCancelFunc *atomic.Value
// Stores the sealunwrapper for downgrade needs
sealUnwrapper physical.Backend
// unsealwithStoredKeysLock is a mutex that prevents multiple processes from
// unsealing with stored keys are the same time.
unsealWithStoredKeysLock sync.Mutex
// Stores any funcs that should be run on successful postUnseal
postUnsealFuncs []func()
Recovery Mode (#7559) * Initial work * rework * s/dr/recovery * Add sys/raw support to recovery mode (#7577) * Factor the raw paths out so they can be run with a SystemBackend. # Conflicts: # vault/logical_system.go * Add handleLogicalRecovery which is like handleLogical but is only sufficient for use with the sys-raw endpoint in recovery mode. No authentication is done yet. * Integrate with recovery-mode. We now handle unauthenticated sys/raw requests, albeit on path v1/raw instead v1/sys/raw. * Use sys/raw instead raw during recovery. * Don't bother persisting the recovery token. Authenticate sys/raw requests with it. * RecoveryMode: Support generate-root for autounseals (#7591) * Recovery: Abstract config creation and log settings * Recovery mode integration test. (#7600) * Recovery: Touch up (#7607) * Recovery: Touch up * revert the raw backend creation changes * Added recovery operation token prefix * Move RawBackend to its own file * Update API path and hit it using CLI flag on generate-root * Fix a panic triggered when handling a request that yields a nil response. (#7618) * Improve integ test to actually make changes while in recovery mode and verify they're still there after coming back in regular mode. * Refuse to allow a second recovery token to be generated. * Resize raft cluster to size 1 and start as leader (#7626) * RecoveryMode: Setup raft cluster post unseal (#7635) * Setup raft cluster post unseal in recovery mode * Remove marking as unsealed as its not needed * Address review comments * Accept only one seal config in recovery mode as there is no scope for migration
2019-10-15 04:55:31 +00:00
// Stores any funcs that should be run on successful barrier unseal in
// recovery mode
postRecoveryUnsealFuncs []func() error
2018-09-18 03:03:00 +00:00
// replicationFailure is used to mark when replication has entered an
// unrecoverable failure.
replicationFailure *uint32
// disablePerfStanby is used to tell a standby not to attempt to become a
// perf standby
disablePerfStandby bool
licensingStopCh chan struct{}
// Stores loggers so we can reset the level
allLoggers []log.Logger
allLoggersLock sync.RWMutex
// Can be toggled atomically to cause the core to never try to become
// active, or give up active as soon as it gets it
neverBecomeActive *uint32
// loadCaseSensitiveIdentityStore enforces the loading of identity store
// artifacts in a case sensitive manner. To be used only in testing.
loadCaseSensitiveIdentityStore bool
// clusterListener starts up and manages connections on the cluster ports
clusterListener *atomic.Value
2019-02-20 20:12:21 +00:00
// Telemetry objects
2019-02-20 20:12:21 +00:00
metricsHelper *metricsutil.MetricsHelper
// Stores request counters
counters counters
Raft Storage Backend (#6888) * Work on raft backend * Add logstore locally * Add encryptor and unsealable interfaces * Add clustering support to raft * Remove client and handler * Bootstrap raft on init * Cleanup raft logic a bit * More raft work * Work on TLS config * More work on bootstrapping * Fix build * More work on bootstrapping * More bootstrapping work * fix build * Remove consul dep * Fix build * merged oss/master into raft-storage * Work on bootstrapping * Get bootstrapping to work * Clean up FMS and node-id * Update local node ID logic * Cleanup node-id change * Work on snapshotting * Raft: Add remove peer API (#906) * Add remove peer API * Add some comments * Fix existing snapshotting (#909) * Raft get peers API (#912) * Read raft configuration * address review feedback * Use the Leadership Transfer API to step-down the active node (#918) * Raft join and unseal using Shamir keys (#917) * Raft join using shamir * Store AEAD instead of master key * Split the raft join process to answer the challenge after a successful unseal * get the follower to standby state * Make unseal work * minor changes * Some input checks * reuse the shamir seal access instead of new default seal access * refactor joinRaftSendAnswer function * Synchronously send answer in auto-unseal case * Address review feedback * Raft snapshots (#910) * Fix existing snapshotting * implement the noop snapshotting * Add comments and switch log libraries * add some snapshot tests * add snapshot test file * add TODO * More work on raft snapshotting * progress on the ConfigStore strategy * Don't use two buckets * Update the snapshot store logic to hide the file logic * Add more backend tests * Cleanup code a bit * [WIP] Raft recovery (#938) * Add recovery functionality * remove fmt.Printfs * Fix a few fsm bugs * Add max size value for raft backend (#942) * Add max size value for raft backend * Include physical.ErrValueTooLarge in the message * Raft snapshot Take/Restore API (#926) * Inital work on raft snapshot APIs * Always redirect snapshot install/download requests * More work on the snapshot APIs * Cleanup code a bit * On restore handle special cases * Use the seal to encrypt the sha sum file * Add sealer mechanism and fix some bugs * Call restore while state lock is held * Send restore cb trigger through raft log * Make error messages nicer * Add test helpers * Add snapshot test * Add shamir unseal test * Add more raft snapshot API tests * Fix locking * Change working to initalize * Add underlying raw object to test cluster core * Move leaderUUID to core * Add raft TLS rotation logic (#950) * Add TLS rotation logic * Cleanup logic a bit * Add/Remove from follower state on add/remove peer * add comments * Update more comments * Update request_forwarding_service.proto * Make sure we populate all nodes in the followerstate obj * Update times * Apply review feedback * Add more raft config setting (#947) * Add performance config setting * Add more config options and fix tests * Test Raft Recovery (#944) * Test raft recovery * Leave out a node during recovery * remove unused struct * Update physical/raft/snapshot_test.go * Update physical/raft/snapshot_test.go * fix vendoring * Switch to new raft interface * Remove unused files * Switch a gogo -> proto instance * Remove unneeded vault dep in go.sum * Update helper/testhelpers/testhelpers.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * Update vault/cluster/cluster.go * track active key within the keyring itself (#6915) * track active key within the keyring itself * lookup and store using the active key ID * update docstring * minor refactor * Small text fixes (#6912) * Update physical/raft/raft.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * review feedback * Move raft logical system into separate file * Update help text a bit * Enforce cluster addr is set and use it for raft bootstrapping * Fix tests * fix http test panic * Pull in latest raft-snapshot library * Add comment
2019-06-20 19:14:58 +00:00
// Stores the raft applied index for standby nodes
raftFollowerStates *raftFollowerStates
// Stop channel for raft TLS rotations
raftTLSRotationStopCh chan struct{}
// Stores the pending peers we are waiting to give answers
pendingRaftPeers map[string][]byte
Raft Storage Backend (#6888) * Work on raft backend * Add logstore locally * Add encryptor and unsealable interfaces * Add clustering support to raft * Remove client and handler * Bootstrap raft on init * Cleanup raft logic a bit * More raft work * Work on TLS config * More work on bootstrapping * Fix build * More work on bootstrapping * More bootstrapping work * fix build * Remove consul dep * Fix build * merged oss/master into raft-storage * Work on bootstrapping * Get bootstrapping to work * Clean up FMS and node-id * Update local node ID logic * Cleanup node-id change * Work on snapshotting * Raft: Add remove peer API (#906) * Add remove peer API * Add some comments * Fix existing snapshotting (#909) * Raft get peers API (#912) * Read raft configuration * address review feedback * Use the Leadership Transfer API to step-down the active node (#918) * Raft join and unseal using Shamir keys (#917) * Raft join using shamir * Store AEAD instead of master key * Split the raft join process to answer the challenge after a successful unseal * get the follower to standby state * Make unseal work * minor changes * Some input checks * reuse the shamir seal access instead of new default seal access * refactor joinRaftSendAnswer function * Synchronously send answer in auto-unseal case * Address review feedback * Raft snapshots (#910) * Fix existing snapshotting * implement the noop snapshotting * Add comments and switch log libraries * add some snapshot tests * add snapshot test file * add TODO * More work on raft snapshotting * progress on the ConfigStore strategy * Don't use two buckets * Update the snapshot store logic to hide the file logic * Add more backend tests * Cleanup code a bit * [WIP] Raft recovery (#938) * Add recovery functionality * remove fmt.Printfs * Fix a few fsm bugs * Add max size value for raft backend (#942) * Add max size value for raft backend * Include physical.ErrValueTooLarge in the message * Raft snapshot Take/Restore API (#926) * Inital work on raft snapshot APIs * Always redirect snapshot install/download requests * More work on the snapshot APIs * Cleanup code a bit * On restore handle special cases * Use the seal to encrypt the sha sum file * Add sealer mechanism and fix some bugs * Call restore while state lock is held * Send restore cb trigger through raft log * Make error messages nicer * Add test helpers * Add snapshot test * Add shamir unseal test * Add more raft snapshot API tests * Fix locking * Change working to initalize * Add underlying raw object to test cluster core * Move leaderUUID to core * Add raft TLS rotation logic (#950) * Add TLS rotation logic * Cleanup logic a bit * Add/Remove from follower state on add/remove peer * add comments * Update more comments * Update request_forwarding_service.proto * Make sure we populate all nodes in the followerstate obj * Update times * Apply review feedback * Add more raft config setting (#947) * Add performance config setting * Add more config options and fix tests * Test Raft Recovery (#944) * Test raft recovery * Leave out a node during recovery * remove unused struct * Update physical/raft/snapshot_test.go * Update physical/raft/snapshot_test.go * fix vendoring * Switch to new raft interface * Remove unused files * Switch a gogo -> proto instance * Remove unneeded vault dep in go.sum * Update helper/testhelpers/testhelpers.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * Update vault/cluster/cluster.go * track active key within the keyring itself (#6915) * track active key within the keyring itself * lookup and store using the active key ID * update docstring * minor refactor * Small text fixes (#6912) * Update physical/raft/raft.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * review feedback * Move raft logical system into separate file * Update help text a bit * Enforce cluster addr is set and use it for raft bootstrapping * Fix tests * fix http test panic * Pull in latest raft-snapshot library * Add comment
2019-06-20 19:14:58 +00:00
// rawConfig stores the config as-is from the provided server configuration.
rawConfig *server.Config
coreNumber int
Recovery Mode (#7559) * Initial work * rework * s/dr/recovery * Add sys/raw support to recovery mode (#7577) * Factor the raw paths out so they can be run with a SystemBackend. # Conflicts: # vault/logical_system.go * Add handleLogicalRecovery which is like handleLogical but is only sufficient for use with the sys-raw endpoint in recovery mode. No authentication is done yet. * Integrate with recovery-mode. We now handle unauthenticated sys/raw requests, albeit on path v1/raw instead v1/sys/raw. * Use sys/raw instead raw during recovery. * Don't bother persisting the recovery token. Authenticate sys/raw requests with it. * RecoveryMode: Support generate-root for autounseals (#7591) * Recovery: Abstract config creation and log settings * Recovery mode integration test. (#7600) * Recovery: Touch up (#7607) * Recovery: Touch up * revert the raw backend creation changes * Added recovery operation token prefix * Move RawBackend to its own file * Update API path and hit it using CLI flag on generate-root * Fix a panic triggered when handling a request that yields a nil response. (#7618) * Improve integ test to actually make changes while in recovery mode and verify they're still there after coming back in regular mode. * Refuse to allow a second recovery token to be generated. * Resize raft cluster to size 1 and start as leader (#7626) * RecoveryMode: Setup raft cluster post unseal (#7635) * Setup raft cluster post unseal in recovery mode * Remove marking as unsealed as its not needed * Address review comments * Accept only one seal config in recovery mode as there is no scope for migration
2019-10-15 04:55:31 +00:00
recoveryMode bool
2015-03-09 23:33:27 +00:00
}
2015-03-11 18:52:01 +00:00
// CoreConfig is used to parameterize a core
type CoreConfig struct {
DevToken string `json:"dev_token" structs:"dev_token" mapstructure:"dev_token"`
2018-11-07 01:21:24 +00:00
BuiltinRegistry BuiltinRegistry `json:"builtin_registry" structs:"builtin_registry" mapstructure:"builtin_registry"`
LogicalBackends map[string]logical.Factory `json:"logical_backends" structs:"logical_backends" mapstructure:"logical_backends"`
CredentialBackends map[string]logical.Factory `json:"credential_backends" structs:"credential_backends" mapstructure:"credential_backends"`
AuditBackends map[string]audit.Factory `json:"audit_backends" structs:"audit_backends" mapstructure:"audit_backends"`
Physical physical.Backend `json:"physical" structs:"physical" mapstructure:"physical"`
StorageType string `json:"storage_type" structs:"storage_type" mapstructure:"storage_type"`
// May be nil, which disables HA operations
HAPhysical physical.HABackend `json:"ha_physical" structs:"ha_physical" mapstructure:"ha_physical"`
Seal Seal `json:"seal" structs:"seal" mapstructure:"seal"`
2016-08-19 20:45:17 +00:00
Logger log.Logger `json:"logger" structs:"logger" mapstructure:"logger"`
// Disables the LRU cache on the physical backend
DisableCache bool `json:"disable_cache" structs:"disable_cache" mapstructure:"disable_cache"`
// Disables mlock syscall
DisableMlock bool `json:"disable_mlock" structs:"disable_mlock" mapstructure:"disable_mlock"`
// Custom cache size for the LRU cache on the physical backend, or zero for default
CacheSize int `json:"cache_size" structs:"cache_size" mapstructure:"cache_size"`
// Set as the leader address for HA
RedirectAddr string `json:"redirect_addr" structs:"redirect_addr" mapstructure:"redirect_addr"`
// Set as the cluster address for HA
ClusterAddr string `json:"cluster_addr" structs:"cluster_addr" mapstructure:"cluster_addr"`
DefaultLeaseTTL time.Duration `json:"default_lease_ttl" structs:"default_lease_ttl" mapstructure:"default_lease_ttl"`
MaxLeaseTTL time.Duration `json:"max_lease_ttl" structs:"max_lease_ttl" mapstructure:"max_lease_ttl"`
ClusterName string `json:"cluster_name" structs:"cluster_name" mapstructure:"cluster_name"`
2016-09-30 04:06:40 +00:00
ClusterCipherSuites string `json:"cluster_cipher_suites" structs:"cluster_cipher_suites" mapstructure:"cluster_cipher_suites"`
2017-02-24 15:45:29 +00:00
EnableUI bool `json:"ui" structs:"ui" mapstructure:"ui"`
// Enable the raw endpoint
EnableRaw bool `json:"enable_raw" structs:"enable_raw" mapstructure:"enable_raw"`
PluginDirectory string `json:"plugin_directory" structs:"plugin_directory" mapstructure:"plugin_directory"`
2017-04-04 00:52:29 +00:00
2018-09-18 03:03:00 +00:00
DisableSealWrap bool `json:"disable_sealwrap" structs:"disable_sealwrap" mapstructure:"disable_sealwrap"`
RawConfig *server.Config
ReloadFuncs *map[string][]reload.ReloadFunc
2016-09-30 04:06:40 +00:00
ReloadFuncsLock *sync.RWMutex
2018-09-18 03:03:00 +00:00
// Licensing
LicensingConfig *LicensingConfig
// Don't set this unless in dev mode, ideally only when using inmem
DevLicenseDuration time.Duration
DisablePerformanceStandby bool
2018-10-23 19:03:17 +00:00
DisableIndexing bool
DisableKeyEncodingChecks bool
2018-09-18 03:03:00 +00:00
AllLoggers []log.Logger
// Telemetry objects
MetricsHelper *metricsutil.MetricsHelper
CounterSyncInterval time.Duration
Recovery Mode (#7559) * Initial work * rework * s/dr/recovery * Add sys/raw support to recovery mode (#7577) * Factor the raw paths out so they can be run with a SystemBackend. # Conflicts: # vault/logical_system.go * Add handleLogicalRecovery which is like handleLogical but is only sufficient for use with the sys-raw endpoint in recovery mode. No authentication is done yet. * Integrate with recovery-mode. We now handle unauthenticated sys/raw requests, albeit on path v1/raw instead v1/sys/raw. * Use sys/raw instead raw during recovery. * Don't bother persisting the recovery token. Authenticate sys/raw requests with it. * RecoveryMode: Support generate-root for autounseals (#7591) * Recovery: Abstract config creation and log settings * Recovery mode integration test. (#7600) * Recovery: Touch up (#7607) * Recovery: Touch up * revert the raw backend creation changes * Added recovery operation token prefix * Move RawBackend to its own file * Update API path and hit it using CLI flag on generate-root * Fix a panic triggered when handling a request that yields a nil response. (#7618) * Improve integ test to actually make changes while in recovery mode and verify they're still there after coming back in regular mode. * Refuse to allow a second recovery token to be generated. * Resize raft cluster to size 1 and start as leader (#7626) * RecoveryMode: Setup raft cluster post unseal (#7635) * Setup raft cluster post unseal in recovery mode * Remove marking as unsealed as its not needed * Address review comments * Accept only one seal config in recovery mode as there is no scope for migration
2019-10-15 04:55:31 +00:00
RecoveryMode bool
2015-03-11 18:52:01 +00:00
}
2018-10-23 06:34:02 +00:00
func (c *CoreConfig) Clone() *CoreConfig {
return &CoreConfig{
DevToken: c.DevToken,
LogicalBackends: c.LogicalBackends,
CredentialBackends: c.CredentialBackends,
AuditBackends: c.AuditBackends,
Physical: c.Physical,
HAPhysical: c.HAPhysical,
Seal: c.Seal,
Logger: c.Logger,
DisableCache: c.DisableCache,
DisableMlock: c.DisableMlock,
CacheSize: c.CacheSize,
StorageType: c.StorageType,
2018-10-23 06:34:02 +00:00
RedirectAddr: c.RedirectAddr,
ClusterAddr: c.ClusterAddr,
DefaultLeaseTTL: c.DefaultLeaseTTL,
MaxLeaseTTL: c.MaxLeaseTTL,
ClusterName: c.ClusterName,
ClusterCipherSuites: c.ClusterCipherSuites,
EnableUI: c.EnableUI,
EnableRaw: c.EnableRaw,
PluginDirectory: c.PluginDirectory,
DisableSealWrap: c.DisableSealWrap,
ReloadFuncs: c.ReloadFuncs,
ReloadFuncsLock: c.ReloadFuncsLock,
LicensingConfig: c.LicensingConfig,
DevLicenseDuration: c.DevLicenseDuration,
DisablePerformanceStandby: c.DisablePerformanceStandby,
2018-10-23 19:03:17 +00:00
DisableIndexing: c.DisableIndexing,
2018-10-23 06:34:02 +00:00
AllLoggers: c.AllLoggers,
CounterSyncInterval: c.CounterSyncInterval,
2018-10-23 06:34:02 +00:00
}
}
// NewCore is used to construct a new core
2015-03-11 18:52:01 +00:00
func NewCore(conf *CoreConfig) (*Core, error) {
if conf.HAPhysical != nil && conf.HAPhysical.HAEnabled() {
if conf.RedirectAddr == "" {
return nil, fmt.Errorf("missing API address, please set in configuration or via environment")
}
2015-04-14 23:44:48 +00:00
}
2015-04-14 21:06:15 +00:00
if conf.DefaultLeaseTTL == 0 {
conf.DefaultLeaseTTL = defaultLeaseTTL
}
if conf.MaxLeaseTTL == 0 {
conf.MaxLeaseTTL = maxLeaseTTL
}
if conf.DefaultLeaseTTL > conf.MaxLeaseTTL {
return nil, fmt.Errorf("cannot have DefaultLeaseTTL larger than MaxLeaseTTL")
}
// Validate the advertise addr if its given to us
if conf.RedirectAddr != "" {
u, err := url.Parse(conf.RedirectAddr)
if err != nil {
return nil, errwrap.Wrapf("redirect address is not valid url: {{err}}", err)
}
if u.Scheme == "" {
2016-08-19 14:52:14 +00:00
return nil, fmt.Errorf("redirect address must include scheme (ex. 'http')")
}
}
// Make a default logger if not provided
if conf.Logger == nil {
conf.Logger = logging.NewVaultLogger(log.Trace)
}
// Instantiate a non-nil raw config if none is provided
if conf.RawConfig == nil {
conf.RawConfig = new(server.Config)
}
syncInterval := conf.CounterSyncInterval
if syncInterval.Nanoseconds() == 0 {
syncInterval = 30 * time.Second
}
2015-03-09 23:33:27 +00:00
// Setup the core
c := &Core{
entCore: entCore{},
devToken: conf.DevToken,
physical: conf.Physical,
Raft Storage Backend (#6888) * Work on raft backend * Add logstore locally * Add encryptor and unsealable interfaces * Add clustering support to raft * Remove client and handler * Bootstrap raft on init * Cleanup raft logic a bit * More raft work * Work on TLS config * More work on bootstrapping * Fix build * More work on bootstrapping * More bootstrapping work * fix build * Remove consul dep * Fix build * merged oss/master into raft-storage * Work on bootstrapping * Get bootstrapping to work * Clean up FMS and node-id * Update local node ID logic * Cleanup node-id change * Work on snapshotting * Raft: Add remove peer API (#906) * Add remove peer API * Add some comments * Fix existing snapshotting (#909) * Raft get peers API (#912) * Read raft configuration * address review feedback * Use the Leadership Transfer API to step-down the active node (#918) * Raft join and unseal using Shamir keys (#917) * Raft join using shamir * Store AEAD instead of master key * Split the raft join process to answer the challenge after a successful unseal * get the follower to standby state * Make unseal work * minor changes * Some input checks * reuse the shamir seal access instead of new default seal access * refactor joinRaftSendAnswer function * Synchronously send answer in auto-unseal case * Address review feedback * Raft snapshots (#910) * Fix existing snapshotting * implement the noop snapshotting * Add comments and switch log libraries * add some snapshot tests * add snapshot test file * add TODO * More work on raft snapshotting * progress on the ConfigStore strategy * Don't use two buckets * Update the snapshot store logic to hide the file logic * Add more backend tests * Cleanup code a bit * [WIP] Raft recovery (#938) * Add recovery functionality * remove fmt.Printfs * Fix a few fsm bugs * Add max size value for raft backend (#942) * Add max size value for raft backend * Include physical.ErrValueTooLarge in the message * Raft snapshot Take/Restore API (#926) * Inital work on raft snapshot APIs * Always redirect snapshot install/download requests * More work on the snapshot APIs * Cleanup code a bit * On restore handle special cases * Use the seal to encrypt the sha sum file * Add sealer mechanism and fix some bugs * Call restore while state lock is held * Send restore cb trigger through raft log * Make error messages nicer * Add test helpers * Add snapshot test * Add shamir unseal test * Add more raft snapshot API tests * Fix locking * Change working to initalize * Add underlying raw object to test cluster core * Move leaderUUID to core * Add raft TLS rotation logic (#950) * Add TLS rotation logic * Cleanup logic a bit * Add/Remove from follower state on add/remove peer * add comments * Update more comments * Update request_forwarding_service.proto * Make sure we populate all nodes in the followerstate obj * Update times * Apply review feedback * Add more raft config setting (#947) * Add performance config setting * Add more config options and fix tests * Test Raft Recovery (#944) * Test raft recovery * Leave out a node during recovery * remove unused struct * Update physical/raft/snapshot_test.go * Update physical/raft/snapshot_test.go * fix vendoring * Switch to new raft interface * Remove unused files * Switch a gogo -> proto instance * Remove unneeded vault dep in go.sum * Update helper/testhelpers/testhelpers.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * Update vault/cluster/cluster.go * track active key within the keyring itself (#6915) * track active key within the keyring itself * lookup and store using the active key ID * update docstring * minor refactor * Small text fixes (#6912) * Update physical/raft/raft.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * review feedback * Move raft logical system into separate file * Update help text a bit * Enforce cluster addr is set and use it for raft bootstrapping * Fix tests * fix http test panic * Pull in latest raft-snapshot library * Add comment
2019-06-20 19:14:58 +00:00
underlyingPhysical: conf.Physical,
storageType: conf.StorageType,
redirectAddr: conf.RedirectAddr,
clusterAddr: new(atomic.Value),
clusterListener: new(atomic.Value),
seal: conf.Seal,
router: NewRouter(),
sealed: new(uint32),
standby: true,
baseLogger: conf.Logger,
logger: conf.Logger.Named("core"),
defaultLeaseTTL: conf.DefaultLeaseTTL,
maxLeaseTTL: conf.MaxLeaseTTL,
cachingDisabled: conf.DisableCache,
clusterName: conf.ClusterName,
clusterPeerClusterAddrsCache: cache.New(3*cluster.HeartbeatInterval, time.Second),
enableMlock: !conf.DisableMlock,
rawEnabled: conf.EnableRaw,
replicationState: new(uint32),
atomicPrimaryClusterAddrs: new(atomic.Value),
atomicPrimaryFailoverAddrs: new(atomic.Value),
localClusterPrivateKey: new(atomic.Value),
localClusterCert: new(atomic.Value),
localClusterParsedCert: new(atomic.Value),
activeNodeReplicationState: new(uint32),
keepHALockOnStepDown: new(uint32),
replicationFailure: new(uint32),
disablePerfStandby: true,
activeContextCancelFunc: new(atomic.Value),
allLoggers: conf.AllLoggers,
builtinRegistry: conf.BuiltinRegistry,
neverBecomeActive: new(uint32),
clusterLeaderParams: new(atomic.Value),
2019-02-20 20:12:21 +00:00
metricsHelper: conf.MetricsHelper,
rawConfig: conf.RawConfig,
counters: counters{
requests: new(uint64),
syncInterval: syncInterval,
},
Recovery Mode (#7559) * Initial work * rework * s/dr/recovery * Add sys/raw support to recovery mode (#7577) * Factor the raw paths out so they can be run with a SystemBackend. # Conflicts: # vault/logical_system.go * Add handleLogicalRecovery which is like handleLogical but is only sufficient for use with the sys-raw endpoint in recovery mode. No authentication is done yet. * Integrate with recovery-mode. We now handle unauthenticated sys/raw requests, albeit on path v1/raw instead v1/sys/raw. * Use sys/raw instead raw during recovery. * Don't bother persisting the recovery token. Authenticate sys/raw requests with it. * RecoveryMode: Support generate-root for autounseals (#7591) * Recovery: Abstract config creation and log settings * Recovery mode integration test. (#7600) * Recovery: Touch up (#7607) * Recovery: Touch up * revert the raw backend creation changes * Added recovery operation token prefix * Move RawBackend to its own file * Update API path and hit it using CLI flag on generate-root * Fix a panic triggered when handling a request that yields a nil response. (#7618) * Improve integ test to actually make changes while in recovery mode and verify they're still there after coming back in regular mode. * Refuse to allow a second recovery token to be generated. * Resize raft cluster to size 1 and start as leader (#7626) * RecoveryMode: Setup raft cluster post unseal (#7635) * Setup raft cluster post unseal in recovery mode * Remove marking as unsealed as its not needed * Address review comments * Accept only one seal config in recovery mode as there is no scope for migration
2019-10-15 04:55:31 +00:00
recoveryMode: conf.RecoveryMode,
2015-03-09 23:33:27 +00:00
}
2015-03-15 23:25:38 +00:00
atomic.StoreUint32(c.sealed, 1)
c.allLoggers = append(c.allLoggers, c.logger)
c.router.logger = c.logger.Named("router")
c.allLoggers = append(c.allLoggers, c.router.logger)
2018-01-23 02:44:38 +00:00
atomic.StoreUint32(c.replicationState, uint32(consts.ReplicationDRDisabled|consts.ReplicationPerformanceDisabled))
c.localClusterCert.Store(([]byte)(nil))
c.localClusterParsedCert.Store((*x509.Certificate)(nil))
c.localClusterPrivateKey.Store((*ecdsa.PrivateKey)(nil))
2018-01-23 02:44:38 +00:00
c.clusterLeaderParams.Store((*ClusterLeaderParams)(nil))
c.clusterAddr.Store(conf.ClusterAddr)
c.activeContextCancelFunc.Store((context.CancelFunc)(nil))
if conf.ClusterCipherSuites != "" {
suites, err := tlsutil.ParseCiphers(conf.ClusterCipherSuites)
if err != nil {
return nil, errwrap.Wrapf("error parsing cluster cipher suites: {{err}}", err)
}
c.clusterCipherSuites = suites
}
// Load CORS config and provide a value for the core field.
2018-06-09 20:57:57 +00:00
c.corsConfig = &CORSConfig{
core: c,
Enabled: new(uint32),
}
2017-06-17 05:26:25 +00:00
2017-10-23 20:03:36 +00:00
if c.seal == nil {
Raft Storage Backend (#6888) * Work on raft backend * Add logstore locally * Add encryptor and unsealable interfaces * Add clustering support to raft * Remove client and handler * Bootstrap raft on init * Cleanup raft logic a bit * More raft work * Work on TLS config * More work on bootstrapping * Fix build * More work on bootstrapping * More bootstrapping work * fix build * Remove consul dep * Fix build * merged oss/master into raft-storage * Work on bootstrapping * Get bootstrapping to work * Clean up FMS and node-id * Update local node ID logic * Cleanup node-id change * Work on snapshotting * Raft: Add remove peer API (#906) * Add remove peer API * Add some comments * Fix existing snapshotting (#909) * Raft get peers API (#912) * Read raft configuration * address review feedback * Use the Leadership Transfer API to step-down the active node (#918) * Raft join and unseal using Shamir keys (#917) * Raft join using shamir * Store AEAD instead of master key * Split the raft join process to answer the challenge after a successful unseal * get the follower to standby state * Make unseal work * minor changes * Some input checks * reuse the shamir seal access instead of new default seal access * refactor joinRaftSendAnswer function * Synchronously send answer in auto-unseal case * Address review feedback * Raft snapshots (#910) * Fix existing snapshotting * implement the noop snapshotting * Add comments and switch log libraries * add some snapshot tests * add snapshot test file * add TODO * More work on raft snapshotting * progress on the ConfigStore strategy * Don't use two buckets * Update the snapshot store logic to hide the file logic * Add more backend tests * Cleanup code a bit * [WIP] Raft recovery (#938) * Add recovery functionality * remove fmt.Printfs * Fix a few fsm bugs * Add max size value for raft backend (#942) * Add max size value for raft backend * Include physical.ErrValueTooLarge in the message * Raft snapshot Take/Restore API (#926) * Inital work on raft snapshot APIs * Always redirect snapshot install/download requests * More work on the snapshot APIs * Cleanup code a bit * On restore handle special cases * Use the seal to encrypt the sha sum file * Add sealer mechanism and fix some bugs * Call restore while state lock is held * Send restore cb trigger through raft log * Make error messages nicer * Add test helpers * Add snapshot test * Add shamir unseal test * Add more raft snapshot API tests * Fix locking * Change working to initalize * Add underlying raw object to test cluster core * Move leaderUUID to core * Add raft TLS rotation logic (#950) * Add TLS rotation logic * Cleanup logic a bit * Add/Remove from follower state on add/remove peer * add comments * Update more comments * Update request_forwarding_service.proto * Make sure we populate all nodes in the followerstate obj * Update times * Apply review feedback * Add more raft config setting (#947) * Add performance config setting * Add more config options and fix tests * Test Raft Recovery (#944) * Test raft recovery * Leave out a node during recovery * remove unused struct * Update physical/raft/snapshot_test.go * Update physical/raft/snapshot_test.go * fix vendoring * Switch to new raft interface * Remove unused files * Switch a gogo -> proto instance * Remove unneeded vault dep in go.sum * Update helper/testhelpers/testhelpers.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * Update vault/cluster/cluster.go * track active key within the keyring itself (#6915) * track active key within the keyring itself * lookup and store using the active key ID * update docstring * minor refactor * Small text fixes (#6912) * Update physical/raft/raft.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * review feedback * Move raft logical system into separate file * Update help text a bit * Enforce cluster addr is set and use it for raft bootstrapping * Fix tests * fix http test panic * Pull in latest raft-snapshot library * Add comment
2019-06-20 19:14:58 +00:00
c.seal = NewDefaultSeal(shamirseal.NewSeal(c.logger.Named("shamir")))
2017-10-23 20:03:36 +00:00
}
c.seal.SetCore(c)
2018-09-18 03:03:00 +00:00
if err := coreInit(c, conf); err != nil {
return nil, err
2017-01-06 20:42:18 +00:00
}
if !conf.DisableMlock {
// Ensure our memory usage is locked into physical RAM
if err := mlock.LockMemory(); err != nil {
return nil, fmt.Errorf(
"Failed to lock memory: %v\n\n"+
"This usually means that the mlock syscall is not available.\n"+
"Vault uses mlock to prevent memory from being swapped to\n"+
"disk. This requires root privileges as well as a machine\n"+
"that supports mlock. Please enable mlock on your system or\n"+
"disable Vault from using it. To disable Vault from using it,\n"+
"set the `disable_mlock` configuration option in your configuration\n"+
"file.",
err)
}
}
var err error
2018-09-18 03:03:00 +00:00
2017-04-04 00:52:29 +00:00
// Construct a new AES-GCM barrier
c.barrier, err = NewAESGCMBarrier(c.physical)
if err != nil {
return nil, errwrap.Wrapf("barrier setup failed: {{err}}", err)
}
2016-09-30 04:06:40 +00:00
// We create the funcs here, then populate the given config with it so that
// the caller can share state
conf.ReloadFuncsLock = &c.reloadFuncsLock
c.reloadFuncsLock.Lock()
c.reloadFuncs = make(map[string][]reload.ReloadFunc)
2016-09-30 04:06:40 +00:00
c.reloadFuncsLock.Unlock()
conf.ReloadFuncs = &c.reloadFuncs
Recovery Mode (#7559) * Initial work * rework * s/dr/recovery * Add sys/raw support to recovery mode (#7577) * Factor the raw paths out so they can be run with a SystemBackend. # Conflicts: # vault/logical_system.go * Add handleLogicalRecovery which is like handleLogical but is only sufficient for use with the sys-raw endpoint in recovery mode. No authentication is done yet. * Integrate with recovery-mode. We now handle unauthenticated sys/raw requests, albeit on path v1/raw instead v1/sys/raw. * Use sys/raw instead raw during recovery. * Don't bother persisting the recovery token. Authenticate sys/raw requests with it. * RecoveryMode: Support generate-root for autounseals (#7591) * Recovery: Abstract config creation and log settings * Recovery mode integration test. (#7600) * Recovery: Touch up (#7607) * Recovery: Touch up * revert the raw backend creation changes * Added recovery operation token prefix * Move RawBackend to its own file * Update API path and hit it using CLI flag on generate-root * Fix a panic triggered when handling a request that yields a nil response. (#7618) * Improve integ test to actually make changes while in recovery mode and verify they're still there after coming back in regular mode. * Refuse to allow a second recovery token to be generated. * Resize raft cluster to size 1 and start as leader (#7626) * RecoveryMode: Setup raft cluster post unseal (#7635) * Setup raft cluster post unseal in recovery mode * Remove marking as unsealed as its not needed * Address review comments * Accept only one seal config in recovery mode as there is no scope for migration
2019-10-15 04:55:31 +00:00
// All the things happening below this are not required in
// recovery mode
if c.recoveryMode {
return c, nil
}
if conf.PluginDirectory != "" {
c.pluginDirectory, err = filepath.Abs(conf.PluginDirectory)
if err != nil {
return nil, errwrap.Wrapf("core setup failed, could not verify plugin directory: {{err}}", err)
}
}
createSecondaries(c, conf)
if conf.HAPhysical != nil && conf.HAPhysical.HAEnabled() {
c.ha = conf.HAPhysical
}
logicalBackends := make(map[string]logical.Factory)
for k, f := range conf.LogicalBackends {
logicalBackends[k] = f
2015-03-15 23:25:38 +00:00
}
2018-11-07 01:21:24 +00:00
_, ok := logicalBackends["kv"]
if !ok {
logicalBackends["kv"] = PassthroughBackendFactory
}
2018-11-07 01:21:24 +00:00
logicalBackends["cubbyhole"] = CubbyholeBackendFactory
2018-09-18 03:03:00 +00:00
logicalBackends[systemMountType] = func(ctx context.Context, config *logical.BackendConfig) (logical.Backend, error) {
sysBackendLogger := conf.Logger.Named("system")
c.AddLogger(sysBackendLogger)
b := NewSystemBackend(c, sysBackendLogger)
if err := b.Setup(ctx, config); err != nil {
Backend plugin system (#2874) * Add backend plugin changes * Fix totp backend plugin tests * Fix logical/plugin InvalidateKey test * Fix plugin catalog CRUD test, fix NoopBackend * Clean up commented code block * Fix system backend mount test * Set plugin_name to omitempty, fix handleMountTable config parsing * Clean up comments, keep shim connections alive until cleanup * Include pluginClient, disallow LookupPlugin call from within a plugin * Add wrapper around backendPluginClient for proper cleanup * Add logger shim tests * Add logger, storage, and system shim tests * Use pointer receivers for system view shim * Use plugin name if no path is provided on mount * Enable plugins for auth backends * Add backend type attribute, move builtin/plugin/package * Fix merge conflict * Fix missing plugin name in mount config * Add integration tests on enabling auth backend plugins * Remove dependency cycle on mock-plugin * Add passthrough backend plugin, use logical.BackendType to determine lease generation * Remove vault package dependency on passthrough package * Add basic impl test for passthrough plugin * Incorporate feedback; set b.backend after shims creation on backendPluginServer * Fix totp plugin test * Add plugin backends docs * Fix tests * Fix builtin/plugin tests * Remove flatten from PluginRunner fields * Move mock plugin to logical/plugin, remove totp and passthrough plugins * Move pluginMap into newPluginClient * Do not create storage RPC connection on HandleRequest and HandleExistenceCheck * Change shim logger's Fatal to no-op * Change BackendType to uint32, match UX backend types * Change framework.Backend Setup signature * Add Setup func to logical.Backend interface * Move OptionallyEnableMlock call into plugin.Serve, update docs and comments * Remove commented var in plugin package * RegisterLicense on logical.Backend interface (#3017) * Add RegisterLicense to logical.Backend interface * Update RegisterLicense to use callback func on framework.Backend * Refactor framework.Backend.RegisterLicense * plugin: Prevent plugin.SystemViewClient.ResponseWrapData from getting JWTs * plugin: Revert BackendType to remove TypePassthrough and related references * Fix typo in plugin backends docs
2017-07-20 17:28:40 +00:00
return nil, err
}
return b, nil
2015-03-15 23:25:38 +00:00
}
logicalBackends["identity"] = func(ctx context.Context, config *logical.BackendConfig) (logical.Backend, error) {
identityLogger := conf.Logger.Named("identity")
c.AddLogger(identityLogger)
return NewIdentityStore(ctx, c, config, identityLogger)
}
2018-09-18 03:03:00 +00:00
addExtraLogicalBackends(c, logicalBackends)
c.logicalBackends = logicalBackends
2015-03-15 23:25:38 +00:00
credentialBackends := make(map[string]logical.Factory)
for k, f := range conf.CredentialBackends {
credentialBackends[k] = f
}
credentialBackends["token"] = func(ctx context.Context, config *logical.BackendConfig) (logical.Backend, error) {
tsLogger := conf.Logger.Named("token")
c.AddLogger(tsLogger)
return NewTokenStore(ctx, tsLogger, c, config)
}
2018-09-18 03:03:00 +00:00
addExtraCredentialBackends(c, credentialBackends)
c.credentialBackends = credentialBackends
auditBackends := make(map[string]audit.Factory)
for k, f := range conf.AuditBackends {
auditBackends[k] = f
}
c.auditBackends = auditBackends
2016-04-04 14:44:22 +00:00
uiStoragePrefix := systemBarrierPrefix + "ui"
c.uiConfig = NewUIConfig(conf.EnableUI, physical.NewView(c.physical, uiStoragePrefix), NewBarrierView(c.barrier, uiStoragePrefix))
c.clusterListener.Store((*cluster.Listener)(nil))
2017-10-23 20:03:36 +00:00
return c, nil
2015-03-09 23:33:27 +00:00
}
2015-06-18 01:23:59 +00:00
// Shutdown is invoked when the Vault instance is about to be terminated. It
// should not be accessible as part of an API call as it will cause an availability
// problem. It is only used to gracefully quit in the case of HA so that failover
// happens as quickly as possible.
func (c *Core) Shutdown() error {
c.logger.Debug("shutdown called")
2018-07-25 03:26:28 +00:00
return c.sealInternal()
2015-06-18 01:23:59 +00:00
}
2017-06-17 04:04:55 +00:00
// CORSConfig returns the current CORS configuration
func (c *Core) CORSConfig() *CORSConfig {
return c.corsConfig
}
func (c *Core) GetContext() (context.Context, context.CancelFunc) {
c.stateLock.RLock()
defer c.stateLock.RUnlock()
2018-09-18 03:03:00 +00:00
return context.WithCancel(namespace.RootContext(c.activeContext))
}
2015-03-09 23:33:27 +00:00
// Sealed checks if the Vault is current sealed
func (c *Core) Sealed() bool {
return atomic.LoadUint32(c.sealed) == 1
2015-03-09 23:33:27 +00:00
}
2015-03-11 18:52:01 +00:00
// SecretProgress returns the number of keys provided so far
func (c *Core) SecretProgress() (int, string) {
2015-03-11 18:52:01 +00:00
c.stateLock.RLock()
defer c.stateLock.RUnlock()
switch c.unlockInfo {
case nil:
return 0, ""
default:
return len(c.unlockInfo.Parts), c.unlockInfo.Nonce
}
2015-03-11 18:52:01 +00:00
}
// ResetUnsealProcess removes the current unlock parts from memory, to reset
// the unsealing process
func (c *Core) ResetUnsealProcess() {
c.stateLock.Lock()
defer c.stateLock.Unlock()
c.unlockInfo = nil
}
// Unseal is used to provide one of the key parts to unseal the Vault.
//
// They key given as a parameter will automatically be zerod after
// this method is done with it. If you want to keep the key around, a copy
// should be made.
func (c *Core) Unseal(key []byte) (bool, error) {
2018-10-23 06:34:02 +00:00
return c.unseal(key, false)
}
func (c *Core) UnsealWithRecoveryKeys(key []byte) (bool, error) {
return c.unseal(key, true)
}
func (c *Core) unseal(key []byte, useRecoveryKeys bool) (bool, error) {
2015-04-08 23:43:17 +00:00
defer metrics.MeasureSince([]string{"core", "unseal"}, time.Now())
c.stateLock.Lock()
defer c.stateLock.Unlock()
c.logger.Debug("unseal key supplied")
ctx := context.Background()
// Explicitly check for init status. This also checks if the seal
// configuration is valid (i.e. non-nil).
init, err := c.Initialized(ctx)
if err != nil {
return false, err
}
Raft Storage Backend (#6888) * Work on raft backend * Add logstore locally * Add encryptor and unsealable interfaces * Add clustering support to raft * Remove client and handler * Bootstrap raft on init * Cleanup raft logic a bit * More raft work * Work on TLS config * More work on bootstrapping * Fix build * More work on bootstrapping * More bootstrapping work * fix build * Remove consul dep * Fix build * merged oss/master into raft-storage * Work on bootstrapping * Get bootstrapping to work * Clean up FMS and node-id * Update local node ID logic * Cleanup node-id change * Work on snapshotting * Raft: Add remove peer API (#906) * Add remove peer API * Add some comments * Fix existing snapshotting (#909) * Raft get peers API (#912) * Read raft configuration * address review feedback * Use the Leadership Transfer API to step-down the active node (#918) * Raft join and unseal using Shamir keys (#917) * Raft join using shamir * Store AEAD instead of master key * Split the raft join process to answer the challenge after a successful unseal * get the follower to standby state * Make unseal work * minor changes * Some input checks * reuse the shamir seal access instead of new default seal access * refactor joinRaftSendAnswer function * Synchronously send answer in auto-unseal case * Address review feedback * Raft snapshots (#910) * Fix existing snapshotting * implement the noop snapshotting * Add comments and switch log libraries * add some snapshot tests * add snapshot test file * add TODO * More work on raft snapshotting * progress on the ConfigStore strategy * Don't use two buckets * Update the snapshot store logic to hide the file logic * Add more backend tests * Cleanup code a bit * [WIP] Raft recovery (#938) * Add recovery functionality * remove fmt.Printfs * Fix a few fsm bugs * Add max size value for raft backend (#942) * Add max size value for raft backend * Include physical.ErrValueTooLarge in the message * Raft snapshot Take/Restore API (#926) * Inital work on raft snapshot APIs * Always redirect snapshot install/download requests * More work on the snapshot APIs * Cleanup code a bit * On restore handle special cases * Use the seal to encrypt the sha sum file * Add sealer mechanism and fix some bugs * Call restore while state lock is held * Send restore cb trigger through raft log * Make error messages nicer * Add test helpers * Add snapshot test * Add shamir unseal test * Add more raft snapshot API tests * Fix locking * Change working to initalize * Add underlying raw object to test cluster core * Move leaderUUID to core * Add raft TLS rotation logic (#950) * Add TLS rotation logic * Cleanup logic a bit * Add/Remove from follower state on add/remove peer * add comments * Update more comments * Update request_forwarding_service.proto * Make sure we populate all nodes in the followerstate obj * Update times * Apply review feedback * Add more raft config setting (#947) * Add performance config setting * Add more config options and fix tests * Test Raft Recovery (#944) * Test raft recovery * Leave out a node during recovery * remove unused struct * Update physical/raft/snapshot_test.go * Update physical/raft/snapshot_test.go * fix vendoring * Switch to new raft interface * Remove unused files * Switch a gogo -> proto instance * Remove unneeded vault dep in go.sum * Update helper/testhelpers/testhelpers.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * Update vault/cluster/cluster.go * track active key within the keyring itself (#6915) * track active key within the keyring itself * lookup and store using the active key ID * update docstring * minor refactor * Small text fixes (#6912) * Update physical/raft/raft.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * review feedback * Move raft logical system into separate file * Update help text a bit * Enforce cluster addr is set and use it for raft bootstrapping * Fix tests * fix http test panic * Pull in latest raft-snapshot library * Add comment
2019-06-20 19:14:58 +00:00
if !init && !c.isRaftUnseal() {
return false, ErrNotInit
}
2015-03-12 18:20:27 +00:00
// Verify the key length
min, max := c.barrier.KeyLength()
max += shamir.ShareOverhead
if len(key) < min {
return false, &ErrInvalidKey{fmt.Sprintf("key is shorter than minimum %d bytes", min)}
}
if len(key) > max {
return false, &ErrInvalidKey{fmt.Sprintf("key is longer than maximum %d bytes", max)}
}
// Check if already unsealed
if !c.Sealed() {
return true, nil
}
2018-10-23 06:34:02 +00:00
sealToUse := c.seal
if c.migrationSeal != nil {
sealToUse = c.migrationSeal
}
2018-10-23 06:34:02 +00:00
masterKey, err := c.unsealPart(ctx, sealToUse, key, useRecoveryKeys)
if err != nil {
return false, err
}
Raft Storage Backend (#6888) * Work on raft backend * Add logstore locally * Add encryptor and unsealable interfaces * Add clustering support to raft * Remove client and handler * Bootstrap raft on init * Cleanup raft logic a bit * More raft work * Work on TLS config * More work on bootstrapping * Fix build * More work on bootstrapping * More bootstrapping work * fix build * Remove consul dep * Fix build * merged oss/master into raft-storage * Work on bootstrapping * Get bootstrapping to work * Clean up FMS and node-id * Update local node ID logic * Cleanup node-id change * Work on snapshotting * Raft: Add remove peer API (#906) * Add remove peer API * Add some comments * Fix existing snapshotting (#909) * Raft get peers API (#912) * Read raft configuration * address review feedback * Use the Leadership Transfer API to step-down the active node (#918) * Raft join and unseal using Shamir keys (#917) * Raft join using shamir * Store AEAD instead of master key * Split the raft join process to answer the challenge after a successful unseal * get the follower to standby state * Make unseal work * minor changes * Some input checks * reuse the shamir seal access instead of new default seal access * refactor joinRaftSendAnswer function * Synchronously send answer in auto-unseal case * Address review feedback * Raft snapshots (#910) * Fix existing snapshotting * implement the noop snapshotting * Add comments and switch log libraries * add some snapshot tests * add snapshot test file * add TODO * More work on raft snapshotting * progress on the ConfigStore strategy * Don't use two buckets * Update the snapshot store logic to hide the file logic * Add more backend tests * Cleanup code a bit * [WIP] Raft recovery (#938) * Add recovery functionality * remove fmt.Printfs * Fix a few fsm bugs * Add max size value for raft backend (#942) * Add max size value for raft backend * Include physical.ErrValueTooLarge in the message * Raft snapshot Take/Restore API (#926) * Inital work on raft snapshot APIs * Always redirect snapshot install/download requests * More work on the snapshot APIs * Cleanup code a bit * On restore handle special cases * Use the seal to encrypt the sha sum file * Add sealer mechanism and fix some bugs * Call restore while state lock is held * Send restore cb trigger through raft log * Make error messages nicer * Add test helpers * Add snapshot test * Add shamir unseal test * Add more raft snapshot API tests * Fix locking * Change working to initalize * Add underlying raw object to test cluster core * Move leaderUUID to core * Add raft TLS rotation logic (#950) * Add TLS rotation logic * Cleanup logic a bit * Add/Remove from follower state on add/remove peer * add comments * Update more comments * Update request_forwarding_service.proto * Make sure we populate all nodes in the followerstate obj * Update times * Apply review feedback * Add more raft config setting (#947) * Add performance config setting * Add more config options and fix tests * Test Raft Recovery (#944) * Test raft recovery * Leave out a node during recovery * remove unused struct * Update physical/raft/snapshot_test.go * Update physical/raft/snapshot_test.go * fix vendoring * Switch to new raft interface * Remove unused files * Switch a gogo -> proto instance * Remove unneeded vault dep in go.sum * Update helper/testhelpers/testhelpers.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * Update vault/cluster/cluster.go * track active key within the keyring itself (#6915) * track active key within the keyring itself * lookup and store using the active key ID * update docstring * minor refactor * Small text fixes (#6912) * Update physical/raft/raft.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * review feedback * Move raft logical system into separate file * Update help text a bit * Enforce cluster addr is set and use it for raft bootstrapping * Fix tests * fix http test panic * Pull in latest raft-snapshot library * Add comment
2019-06-20 19:14:58 +00:00
if masterKey != nil {
Raft Storage Backend (#6888) * Work on raft backend * Add logstore locally * Add encryptor and unsealable interfaces * Add clustering support to raft * Remove client and handler * Bootstrap raft on init * Cleanup raft logic a bit * More raft work * Work on TLS config * More work on bootstrapping * Fix build * More work on bootstrapping * More bootstrapping work * fix build * Remove consul dep * Fix build * merged oss/master into raft-storage * Work on bootstrapping * Get bootstrapping to work * Clean up FMS and node-id * Update local node ID logic * Cleanup node-id change * Work on snapshotting * Raft: Add remove peer API (#906) * Add remove peer API * Add some comments * Fix existing snapshotting (#909) * Raft get peers API (#912) * Read raft configuration * address review feedback * Use the Leadership Transfer API to step-down the active node (#918) * Raft join and unseal using Shamir keys (#917) * Raft join using shamir * Store AEAD instead of master key * Split the raft join process to answer the challenge after a successful unseal * get the follower to standby state * Make unseal work * minor changes * Some input checks * reuse the shamir seal access instead of new default seal access * refactor joinRaftSendAnswer function * Synchronously send answer in auto-unseal case * Address review feedback * Raft snapshots (#910) * Fix existing snapshotting * implement the noop snapshotting * Add comments and switch log libraries * add some snapshot tests * add snapshot test file * add TODO * More work on raft snapshotting * progress on the ConfigStore strategy * Don't use two buckets * Update the snapshot store logic to hide the file logic * Add more backend tests * Cleanup code a bit * [WIP] Raft recovery (#938) * Add recovery functionality * remove fmt.Printfs * Fix a few fsm bugs * Add max size value for raft backend (#942) * Add max size value for raft backend * Include physical.ErrValueTooLarge in the message * Raft snapshot Take/Restore API (#926) * Inital work on raft snapshot APIs * Always redirect snapshot install/download requests * More work on the snapshot APIs * Cleanup code a bit * On restore handle special cases * Use the seal to encrypt the sha sum file * Add sealer mechanism and fix some bugs * Call restore while state lock is held * Send restore cb trigger through raft log * Make error messages nicer * Add test helpers * Add snapshot test * Add shamir unseal test * Add more raft snapshot API tests * Fix locking * Change working to initalize * Add underlying raw object to test cluster core * Move leaderUUID to core * Add raft TLS rotation logic (#950) * Add TLS rotation logic * Cleanup logic a bit * Add/Remove from follower state on add/remove peer * add comments * Update more comments * Update request_forwarding_service.proto * Make sure we populate all nodes in the followerstate obj * Update times * Apply review feedback * Add more raft config setting (#947) * Add performance config setting * Add more config options and fix tests * Test Raft Recovery (#944) * Test raft recovery * Leave out a node during recovery * remove unused struct * Update physical/raft/snapshot_test.go * Update physical/raft/snapshot_test.go * fix vendoring * Switch to new raft interface * Remove unused files * Switch a gogo -> proto instance * Remove unneeded vault dep in go.sum * Update helper/testhelpers/testhelpers.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * Update vault/cluster/cluster.go * track active key within the keyring itself (#6915) * track active key within the keyring itself * lookup and store using the active key ID * update docstring * minor refactor * Small text fixes (#6912) * Update physical/raft/raft.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * review feedback * Move raft logical system into separate file * Update help text a bit * Enforce cluster addr is set and use it for raft bootstrapping * Fix tests * fix http test panic * Pull in latest raft-snapshot library * Add comment
2019-06-20 19:14:58 +00:00
if c.seal.BarrierType() == seal.Shamir {
_, err := c.seal.GetAccess().(*shamirseal.ShamirSeal).SetConfig(map[string]string{
"key": base64.StdEncoding.EncodeToString(masterKey),
})
if err != nil {
return false, err
}
}
if !c.isRaftUnseal() {
return c.unsealInternal(ctx, masterKey)
}
// If we are in the middle of a raft join send the answer and wait for
// data to start streaming in.
if err := c.joinRaftSendAnswer(ctx, c.seal.GetAccess(), c.raftInfo); err != nil {
Raft Storage Backend (#6888) * Work on raft backend * Add logstore locally * Add encryptor and unsealable interfaces * Add clustering support to raft * Remove client and handler * Bootstrap raft on init * Cleanup raft logic a bit * More raft work * Work on TLS config * More work on bootstrapping * Fix build * More work on bootstrapping * More bootstrapping work * fix build * Remove consul dep * Fix build * merged oss/master into raft-storage * Work on bootstrapping * Get bootstrapping to work * Clean up FMS and node-id * Update local node ID logic * Cleanup node-id change * Work on snapshotting * Raft: Add remove peer API (#906) * Add remove peer API * Add some comments * Fix existing snapshotting (#909) * Raft get peers API (#912) * Read raft configuration * address review feedback * Use the Leadership Transfer API to step-down the active node (#918) * Raft join and unseal using Shamir keys (#917) * Raft join using shamir * Store AEAD instead of master key * Split the raft join process to answer the challenge after a successful unseal * get the follower to standby state * Make unseal work * minor changes * Some input checks * reuse the shamir seal access instead of new default seal access * refactor joinRaftSendAnswer function * Synchronously send answer in auto-unseal case * Address review feedback * Raft snapshots (#910) * Fix existing snapshotting * implement the noop snapshotting * Add comments and switch log libraries * add some snapshot tests * add snapshot test file * add TODO * More work on raft snapshotting * progress on the ConfigStore strategy * Don't use two buckets * Update the snapshot store logic to hide the file logic * Add more backend tests * Cleanup code a bit * [WIP] Raft recovery (#938) * Add recovery functionality * remove fmt.Printfs * Fix a few fsm bugs * Add max size value for raft backend (#942) * Add max size value for raft backend * Include physical.ErrValueTooLarge in the message * Raft snapshot Take/Restore API (#926) * Inital work on raft snapshot APIs * Always redirect snapshot install/download requests * More work on the snapshot APIs * Cleanup code a bit * On restore handle special cases * Use the seal to encrypt the sha sum file * Add sealer mechanism and fix some bugs * Call restore while state lock is held * Send restore cb trigger through raft log * Make error messages nicer * Add test helpers * Add snapshot test * Add shamir unseal test * Add more raft snapshot API tests * Fix locking * Change working to initalize * Add underlying raw object to test cluster core * Move leaderUUID to core * Add raft TLS rotation logic (#950) * Add TLS rotation logic * Cleanup logic a bit * Add/Remove from follower state on add/remove peer * add comments * Update more comments * Update request_forwarding_service.proto * Make sure we populate all nodes in the followerstate obj * Update times * Apply review feedback * Add more raft config setting (#947) * Add performance config setting * Add more config options and fix tests * Test Raft Recovery (#944) * Test raft recovery * Leave out a node during recovery * remove unused struct * Update physical/raft/snapshot_test.go * Update physical/raft/snapshot_test.go * fix vendoring * Switch to new raft interface * Remove unused files * Switch a gogo -> proto instance * Remove unneeded vault dep in go.sum * Update helper/testhelpers/testhelpers.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * Update vault/cluster/cluster.go * track active key within the keyring itself (#6915) * track active key within the keyring itself * lookup and store using the active key ID * update docstring * minor refactor * Small text fixes (#6912) * Update physical/raft/raft.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * review feedback * Move raft logical system into separate file * Update help text a bit * Enforce cluster addr is set and use it for raft bootstrapping * Fix tests * fix http test panic * Pull in latest raft-snapshot library * Add comment
2019-06-20 19:14:58 +00:00
return false, err
}
// Reset the state
c.raftInfo = nil
Raft Storage Backend (#6888) * Work on raft backend * Add logstore locally * Add encryptor and unsealable interfaces * Add clustering support to raft * Remove client and handler * Bootstrap raft on init * Cleanup raft logic a bit * More raft work * Work on TLS config * More work on bootstrapping * Fix build * More work on bootstrapping * More bootstrapping work * fix build * Remove consul dep * Fix build * merged oss/master into raft-storage * Work on bootstrapping * Get bootstrapping to work * Clean up FMS and node-id * Update local node ID logic * Cleanup node-id change * Work on snapshotting * Raft: Add remove peer API (#906) * Add remove peer API * Add some comments * Fix existing snapshotting (#909) * Raft get peers API (#912) * Read raft configuration * address review feedback * Use the Leadership Transfer API to step-down the active node (#918) * Raft join and unseal using Shamir keys (#917) * Raft join using shamir * Store AEAD instead of master key * Split the raft join process to answer the challenge after a successful unseal * get the follower to standby state * Make unseal work * minor changes * Some input checks * reuse the shamir seal access instead of new default seal access * refactor joinRaftSendAnswer function * Synchronously send answer in auto-unseal case * Address review feedback * Raft snapshots (#910) * Fix existing snapshotting * implement the noop snapshotting * Add comments and switch log libraries * add some snapshot tests * add snapshot test file * add TODO * More work on raft snapshotting * progress on the ConfigStore strategy * Don't use two buckets * Update the snapshot store logic to hide the file logic * Add more backend tests * Cleanup code a bit * [WIP] Raft recovery (#938) * Add recovery functionality * remove fmt.Printfs * Fix a few fsm bugs * Add max size value for raft backend (#942) * Add max size value for raft backend * Include physical.ErrValueTooLarge in the message * Raft snapshot Take/Restore API (#926) * Inital work on raft snapshot APIs * Always redirect snapshot install/download requests * More work on the snapshot APIs * Cleanup code a bit * On restore handle special cases * Use the seal to encrypt the sha sum file * Add sealer mechanism and fix some bugs * Call restore while state lock is held * Send restore cb trigger through raft log * Make error messages nicer * Add test helpers * Add snapshot test * Add shamir unseal test * Add more raft snapshot API tests * Fix locking * Change working to initalize * Add underlying raw object to test cluster core * Move leaderUUID to core * Add raft TLS rotation logic (#950) * Add TLS rotation logic * Cleanup logic a bit * Add/Remove from follower state on add/remove peer * add comments * Update more comments * Update request_forwarding_service.proto * Make sure we populate all nodes in the followerstate obj * Update times * Apply review feedback * Add more raft config setting (#947) * Add performance config setting * Add more config options and fix tests * Test Raft Recovery (#944) * Test raft recovery * Leave out a node during recovery * remove unused struct * Update physical/raft/snapshot_test.go * Update physical/raft/snapshot_test.go * fix vendoring * Switch to new raft interface * Remove unused files * Switch a gogo -> proto instance * Remove unneeded vault dep in go.sum * Update helper/testhelpers/testhelpers.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * Update vault/cluster/cluster.go * track active key within the keyring itself (#6915) * track active key within the keyring itself * lookup and store using the active key ID * update docstring * minor refactor * Small text fixes (#6912) * Update physical/raft/raft.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * review feedback * Move raft logical system into separate file * Update help text a bit * Enforce cluster addr is set and use it for raft bootstrapping * Fix tests * fix http test panic * Pull in latest raft-snapshot library * Add comment
2019-06-20 19:14:58 +00:00
go func() {
keyringFound := false
defer func() {
if keyringFound {
_, err := c.unsealInternal(ctx, masterKey)
if err != nil {
c.logger.Error("failed to unseal", "error", err)
}
}
}()
// Wait until we at least have the keyring before we attempt to
// unseal the node.
for {
keys, err := c.underlyingPhysical.List(ctx, keyringPrefix)
if err != nil {
c.logger.Error("failed to list physical keys", "error", err)
return
}
if strutil.StrListContains(keys, "keyring") {
keyringFound = true
return
}
select {
case <-time.After(1 * time.Second):
}
}
}()
// Return Vault as sealed since unsealing happens in background
// which gets delayed until the data from the leader is streamed to
// the follower.
return true, nil
}
return false, nil
}
// unsealPart takes in a key share, and returns the master key if the threshold
// is met. If recovery keys are supported, recovery key shares may be provided.
2018-10-23 06:34:02 +00:00
func (c *Core) unsealPart(ctx context.Context, seal Seal, key []byte, useRecoveryKeys bool) ([]byte, error) {
2015-03-11 18:43:36 +00:00
// Check if we already have this piece
if c.unlockInfo != nil {
for _, existing := range c.unlockInfo.Parts {
if subtle.ConstantTimeCompare(existing, key) == 1 {
return nil, nil
}
}
} else {
uuid, err := uuid.GenerateUUID()
if err != nil {
return nil, err
}
c.unlockInfo = &unlockInformation{
Nonce: uuid,
2015-03-11 18:43:36 +00:00
}
}
// Store this key
c.unlockInfo.Parts = append(c.unlockInfo.Parts, key)
2015-03-11 18:43:36 +00:00
2018-10-23 06:34:02 +00:00
var config *SealConfig
var err error
Raft Storage Backend (#6888) * Work on raft backend * Add logstore locally * Add encryptor and unsealable interfaces * Add clustering support to raft * Remove client and handler * Bootstrap raft on init * Cleanup raft logic a bit * More raft work * Work on TLS config * More work on bootstrapping * Fix build * More work on bootstrapping * More bootstrapping work * fix build * Remove consul dep * Fix build * merged oss/master into raft-storage * Work on bootstrapping * Get bootstrapping to work * Clean up FMS and node-id * Update local node ID logic * Cleanup node-id change * Work on snapshotting * Raft: Add remove peer API (#906) * Add remove peer API * Add some comments * Fix existing snapshotting (#909) * Raft get peers API (#912) * Read raft configuration * address review feedback * Use the Leadership Transfer API to step-down the active node (#918) * Raft join and unseal using Shamir keys (#917) * Raft join using shamir * Store AEAD instead of master key * Split the raft join process to answer the challenge after a successful unseal * get the follower to standby state * Make unseal work * minor changes * Some input checks * reuse the shamir seal access instead of new default seal access * refactor joinRaftSendAnswer function * Synchronously send answer in auto-unseal case * Address review feedback * Raft snapshots (#910) * Fix existing snapshotting * implement the noop snapshotting * Add comments and switch log libraries * add some snapshot tests * add snapshot test file * add TODO * More work on raft snapshotting * progress on the ConfigStore strategy * Don't use two buckets * Update the snapshot store logic to hide the file logic * Add more backend tests * Cleanup code a bit * [WIP] Raft recovery (#938) * Add recovery functionality * remove fmt.Printfs * Fix a few fsm bugs * Add max size value for raft backend (#942) * Add max size value for raft backend * Include physical.ErrValueTooLarge in the message * Raft snapshot Take/Restore API (#926) * Inital work on raft snapshot APIs * Always redirect snapshot install/download requests * More work on the snapshot APIs * Cleanup code a bit * On restore handle special cases * Use the seal to encrypt the sha sum file * Add sealer mechanism and fix some bugs * Call restore while state lock is held * Send restore cb trigger through raft log * Make error messages nicer * Add test helpers * Add snapshot test * Add shamir unseal test * Add more raft snapshot API tests * Fix locking * Change working to initalize * Add underlying raw object to test cluster core * Move leaderUUID to core * Add raft TLS rotation logic (#950) * Add TLS rotation logic * Cleanup logic a bit * Add/Remove from follower state on add/remove peer * add comments * Update more comments * Update request_forwarding_service.proto * Make sure we populate all nodes in the followerstate obj * Update times * Apply review feedback * Add more raft config setting (#947) * Add performance config setting * Add more config options and fix tests * Test Raft Recovery (#944) * Test raft recovery * Leave out a node during recovery * remove unused struct * Update physical/raft/snapshot_test.go * Update physical/raft/snapshot_test.go * fix vendoring * Switch to new raft interface * Remove unused files * Switch a gogo -> proto instance * Remove unneeded vault dep in go.sum * Update helper/testhelpers/testhelpers.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * Update vault/cluster/cluster.go * track active key within the keyring itself (#6915) * track active key within the keyring itself * lookup and store using the active key ID * update docstring * minor refactor * Small text fixes (#6912) * Update physical/raft/raft.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * review feedback * Move raft logical system into separate file * Update help text a bit * Enforce cluster addr is set and use it for raft bootstrapping * Fix tests * fix http test panic * Pull in latest raft-snapshot library * Add comment
2019-06-20 19:14:58 +00:00
switch {
case seal.RecoveryKeySupported() && (useRecoveryKeys || c.migrationSeal != nil):
2018-10-23 06:34:02 +00:00
config, err = seal.RecoveryConfig(ctx)
Raft Storage Backend (#6888) * Work on raft backend * Add logstore locally * Add encryptor and unsealable interfaces * Add clustering support to raft * Remove client and handler * Bootstrap raft on init * Cleanup raft logic a bit * More raft work * Work on TLS config * More work on bootstrapping * Fix build * More work on bootstrapping * More bootstrapping work * fix build * Remove consul dep * Fix build * merged oss/master into raft-storage * Work on bootstrapping * Get bootstrapping to work * Clean up FMS and node-id * Update local node ID logic * Cleanup node-id change * Work on snapshotting * Raft: Add remove peer API (#906) * Add remove peer API * Add some comments * Fix existing snapshotting (#909) * Raft get peers API (#912) * Read raft configuration * address review feedback * Use the Leadership Transfer API to step-down the active node (#918) * Raft join and unseal using Shamir keys (#917) * Raft join using shamir * Store AEAD instead of master key * Split the raft join process to answer the challenge after a successful unseal * get the follower to standby state * Make unseal work * minor changes * Some input checks * reuse the shamir seal access instead of new default seal access * refactor joinRaftSendAnswer function * Synchronously send answer in auto-unseal case * Address review feedback * Raft snapshots (#910) * Fix existing snapshotting * implement the noop snapshotting * Add comments and switch log libraries * add some snapshot tests * add snapshot test file * add TODO * More work on raft snapshotting * progress on the ConfigStore strategy * Don't use two buckets * Update the snapshot store logic to hide the file logic * Add more backend tests * Cleanup code a bit * [WIP] Raft recovery (#938) * Add recovery functionality * remove fmt.Printfs * Fix a few fsm bugs * Add max size value for raft backend (#942) * Add max size value for raft backend * Include physical.ErrValueTooLarge in the message * Raft snapshot Take/Restore API (#926) * Inital work on raft snapshot APIs * Always redirect snapshot install/download requests * More work on the snapshot APIs * Cleanup code a bit * On restore handle special cases * Use the seal to encrypt the sha sum file * Add sealer mechanism and fix some bugs * Call restore while state lock is held * Send restore cb trigger through raft log * Make error messages nicer * Add test helpers * Add snapshot test * Add shamir unseal test * Add more raft snapshot API tests * Fix locking * Change working to initalize * Add underlying raw object to test cluster core * Move leaderUUID to core * Add raft TLS rotation logic (#950) * Add TLS rotation logic * Cleanup logic a bit * Add/Remove from follower state on add/remove peer * add comments * Update more comments * Update request_forwarding_service.proto * Make sure we populate all nodes in the followerstate obj * Update times * Apply review feedback * Add more raft config setting (#947) * Add performance config setting * Add more config options and fix tests * Test Raft Recovery (#944) * Test raft recovery * Leave out a node during recovery * remove unused struct * Update physical/raft/snapshot_test.go * Update physical/raft/snapshot_test.go * fix vendoring * Switch to new raft interface * Remove unused files * Switch a gogo -> proto instance * Remove unneeded vault dep in go.sum * Update helper/testhelpers/testhelpers.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * Update vault/cluster/cluster.go * track active key within the keyring itself (#6915) * track active key within the keyring itself * lookup and store using the active key ID * update docstring * minor refactor * Small text fixes (#6912) * Update physical/raft/raft.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * review feedback * Move raft logical system into separate file * Update help text a bit * Enforce cluster addr is set and use it for raft bootstrapping * Fix tests * fix http test panic * Pull in latest raft-snapshot library * Add comment
2019-06-20 19:14:58 +00:00
case c.isRaftUnseal():
// Ignore follower's seal config and refer to leader's barrier
// configuration.
config = c.raftInfo.leaderBarrierConfig
Raft Storage Backend (#6888) * Work on raft backend * Add logstore locally * Add encryptor and unsealable interfaces * Add clustering support to raft * Remove client and handler * Bootstrap raft on init * Cleanup raft logic a bit * More raft work * Work on TLS config * More work on bootstrapping * Fix build * More work on bootstrapping * More bootstrapping work * fix build * Remove consul dep * Fix build * merged oss/master into raft-storage * Work on bootstrapping * Get bootstrapping to work * Clean up FMS and node-id * Update local node ID logic * Cleanup node-id change * Work on snapshotting * Raft: Add remove peer API (#906) * Add remove peer API * Add some comments * Fix existing snapshotting (#909) * Raft get peers API (#912) * Read raft configuration * address review feedback * Use the Leadership Transfer API to step-down the active node (#918) * Raft join and unseal using Shamir keys (#917) * Raft join using shamir * Store AEAD instead of master key * Split the raft join process to answer the challenge after a successful unseal * get the follower to standby state * Make unseal work * minor changes * Some input checks * reuse the shamir seal access instead of new default seal access * refactor joinRaftSendAnswer function * Synchronously send answer in auto-unseal case * Address review feedback * Raft snapshots (#910) * Fix existing snapshotting * implement the noop snapshotting * Add comments and switch log libraries * add some snapshot tests * add snapshot test file * add TODO * More work on raft snapshotting * progress on the ConfigStore strategy * Don't use two buckets * Update the snapshot store logic to hide the file logic * Add more backend tests * Cleanup code a bit * [WIP] Raft recovery (#938) * Add recovery functionality * remove fmt.Printfs * Fix a few fsm bugs * Add max size value for raft backend (#942) * Add max size value for raft backend * Include physical.ErrValueTooLarge in the message * Raft snapshot Take/Restore API (#926) * Inital work on raft snapshot APIs * Always redirect snapshot install/download requests * More work on the snapshot APIs * Cleanup code a bit * On restore handle special cases * Use the seal to encrypt the sha sum file * Add sealer mechanism and fix some bugs * Call restore while state lock is held * Send restore cb trigger through raft log * Make error messages nicer * Add test helpers * Add snapshot test * Add shamir unseal test * Add more raft snapshot API tests * Fix locking * Change working to initalize * Add underlying raw object to test cluster core * Move leaderUUID to core * Add raft TLS rotation logic (#950) * Add TLS rotation logic * Cleanup logic a bit * Add/Remove from follower state on add/remove peer * add comments * Update more comments * Update request_forwarding_service.proto * Make sure we populate all nodes in the followerstate obj * Update times * Apply review feedback * Add more raft config setting (#947) * Add performance config setting * Add more config options and fix tests * Test Raft Recovery (#944) * Test raft recovery * Leave out a node during recovery * remove unused struct * Update physical/raft/snapshot_test.go * Update physical/raft/snapshot_test.go * fix vendoring * Switch to new raft interface * Remove unused files * Switch a gogo -> proto instance * Remove unneeded vault dep in go.sum * Update helper/testhelpers/testhelpers.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * Update vault/cluster/cluster.go * track active key within the keyring itself (#6915) * track active key within the keyring itself * lookup and store using the active key ID * update docstring * minor refactor * Small text fixes (#6912) * Update physical/raft/raft.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * review feedback * Move raft logical system into separate file * Update help text a bit * Enforce cluster addr is set and use it for raft bootstrapping * Fix tests * fix http test panic * Pull in latest raft-snapshot library * Add comment
2019-06-20 19:14:58 +00:00
default:
2018-10-23 06:34:02 +00:00
config, err = seal.BarrierConfig(ctx)
}
if err != nil {
return nil, err
}
// Check if we don't have enough keys to unlock, proceed through the rest of
// the call only if we have met the threshold
if len(c.unlockInfo.Parts) < config.SecretThreshold {
2016-08-19 20:45:17 +00:00
if c.logger.IsDebug() {
c.logger.Debug("cannot unseal, not enough keys", "keys", len(c.unlockInfo.Parts), "threshold", config.SecretThreshold, "nonce", c.unlockInfo.Nonce)
2016-08-19 20:45:17 +00:00
}
return nil, nil
2015-03-11 18:43:36 +00:00
}
// Best-effort memzero of unlock parts once we're done with them
defer func() {
for i := range c.unlockInfo.Parts {
memzero(c.unlockInfo.Parts[i])
}
c.unlockInfo = nil
}()
// Recover the split key. recoveredKey is the shamir combined
// key, or the single provided key if the threshold is 1.
var recoveredKey []byte
2018-10-23 06:34:02 +00:00
var masterKey []byte
var recoveryKey []byte
2015-03-11 18:43:36 +00:00
if config.SecretThreshold == 1 {
recoveredKey = make([]byte, len(c.unlockInfo.Parts[0]))
copy(recoveredKey, c.unlockInfo.Parts[0])
2015-03-11 18:43:36 +00:00
} else {
recoveredKey, err = shamir.Combine(c.unlockInfo.Parts)
2015-03-11 18:43:36 +00:00
if err != nil {
return nil, errwrap.Wrapf("failed to compute master key: {{err}}", err)
2015-03-11 18:43:36 +00:00
}
}
2018-10-23 06:34:02 +00:00
if seal.RecoveryKeySupported() && (useRecoveryKeys || c.migrationSeal != nil) {
// Verify recovery key
2018-10-23 06:34:02 +00:00
if err := seal.VerifyRecoveryKey(ctx, recoveredKey); err != nil {
return nil, err
}
2018-10-23 06:34:02 +00:00
recoveryKey = recoveredKey
// Get stored keys and shamir combine into single master key. Unsealing with
// recovery keys currently does not support: 1) mixed stored and non-stored
// keys setup, nor 2) seals that support recovery keys but not stored keys.
2018-03-20 18:54:10 +00:00
// If insufficient shares are provided, shamir.Combine will error, and if
// no stored keys are found it will return masterKey as nil.
2018-10-23 06:34:02 +00:00
if seal.StoredKeysSupported() {
masterKeyShares, err := seal.GetStoredKeys(ctx)
if err != nil {
return nil, errwrap.Wrapf("unable to retrieve stored keys: {{err}}", err)
}
switch len(masterKeyShares) {
case 0:
return nil, errors.New("seal returned no master key shares")
case 1:
2018-10-23 06:34:02 +00:00
masterKey = masterKeyShares[0]
default:
2018-10-23 06:34:02 +00:00
masterKey, err = shamir.Combine(masterKeyShares)
if err != nil {
return nil, errwrap.Wrapf("failed to compute master key: {{err}}", err)
}
}
2018-10-23 06:34:02 +00:00
}
} else {
masterKey = recoveredKey
}
// If we have a migration seal, now's the time!
if c.migrationSeal != nil {
// Unseal the barrier so we can rekey
if err := c.barrier.Unseal(ctx, masterKey); err != nil {
return nil, errwrap.Wrapf("error unsealing barrier with constructed master key: {{err}}", err)
}
defer c.barrier.Seal()
switch {
case c.migrationSeal.RecoveryKeySupported() && c.seal.RecoveryKeySupported():
// Set the recovery and barrier keys to be the same.
recoveryKey, err := c.migrationSeal.RecoveryKey(ctx)
if err != nil {
return nil, errwrap.Wrapf("error getting recovery key to set on new seal: {{err}}", err)
}
if err := c.seal.SetRecoveryKey(ctx, recoveryKey); err != nil {
return nil, errwrap.Wrapf("error setting new recovery key information during migrate: {{err}}", err)
}
barrierKeys, err := c.migrationSeal.GetStoredKeys(ctx)
if err != nil {
return nil, errwrap.Wrapf("error getting stored keys to set on new seal: {{err}}", err)
}
if err := c.seal.SetStoredKeys(ctx, barrierKeys); err != nil {
return nil, errwrap.Wrapf("error setting new barrier key information during migrate: {{err}}", err)
}
case c.migrationSeal.RecoveryKeySupported():
// Auto to Shamir, since recovery key isn't supported on new seal
// In this case we have to ensure that the recovery information was
// set properly.
if recoveryKey == nil {
return nil, errors.New("did not get expected recovery information to set new seal during migration")
}
// We have recovery keys; we're going to use them as the new
// barrier key.
if err := c.barrier.Rekey(ctx, recoveryKey); err != nil {
return nil, errwrap.Wrapf("error rekeying barrier during migration: {{err}}", err)
}
if err := c.barrier.Delete(ctx, StoredBarrierKeysPath); err != nil {
// Don't actually exit here as successful deletion isn't critical
c.logger.Error("error deleting stored barrier keys after migration; continuing anyways", "error", err)
}
masterKey = recoveryKey
case c.seal.RecoveryKeySupported():
2018-10-23 06:34:02 +00:00
// The new seal will have recovery keys; we set it to the existing
// master key, so barrier key shares -> recovery key shares
if err := c.seal.SetRecoveryKey(ctx, masterKey); err != nil {
return nil, errwrap.Wrapf("error setting new recovery key information: {{err}}", err)
}
// Generate a new master key
newMasterKey, err := c.barrier.GenerateKey()
if err != nil {
return nil, errwrap.Wrapf("error generating new master key: {{err}}", err)
}
// Rekey the barrier
if err := c.barrier.Rekey(ctx, newMasterKey); err != nil {
return nil, errwrap.Wrapf("error rekeying barrier during migration: {{err}}", err)
}
// Store the new master key
if err := c.seal.SetStoredKeys(ctx, [][]byte{newMasterKey}); err != nil {
return nil, errwrap.Wrapf("error storing new master key: {[err}}", err)
}
// Return the new key so it can be used to unlock the barrier
masterKey = newMasterKey
default:
return nil, errors.New("unhandled migration case (shamir to shamir)")
2018-10-23 06:34:02 +00:00
}
// At this point we've swapped things around and need to ensure we
// don't migrate again
c.migrationSeal = nil
// Ensure we populate the new values
bc, err := c.seal.BarrierConfig(ctx)
if err != nil {
return nil, errwrap.Wrapf("error fetching barrier config after migration: {{err}}", err)
}
if err := c.seal.SetBarrierConfig(ctx, bc); err != nil {
return nil, errwrap.Wrapf("error storing barrier config after migration: {{err}}", err)
}
2018-10-23 06:34:02 +00:00
if c.seal.RecoveryKeySupported() {
rc, err := c.seal.RecoveryConfig(ctx)
if err != nil {
2018-10-23 06:34:02 +00:00
return nil, errwrap.Wrapf("error fetching recovery config after migration: {{err}}", err)
}
if err := c.seal.SetRecoveryConfig(ctx, rc); err != nil {
return nil, errwrap.Wrapf("error storing recovery config after migration: {{err}}", err)
}
}
}
2018-10-23 06:34:02 +00:00
return masterKey, nil
}
// unsealInternal takes in the master key and attempts to unseal the barrier.
// N.B.: This must be called with the state write lock held.
func (c *Core) unsealInternal(ctx context.Context, masterKey []byte) (bool, error) {
defer memzero(masterKey)
2015-03-11 18:43:36 +00:00
// Attempt to unlock
if err := c.barrier.Unseal(ctx, masterKey); err != nil {
2015-03-11 18:43:36 +00:00
return false, err
}
2018-09-18 03:03:00 +00:00
if err := preUnsealInternal(ctx, c); err != nil {
return false, err
}
if err := c.startClusterListener(ctx); err != nil {
return false, err
}
Raft Storage Backend (#6888) * Work on raft backend * Add logstore locally * Add encryptor and unsealable interfaces * Add clustering support to raft * Remove client and handler * Bootstrap raft on init * Cleanup raft logic a bit * More raft work * Work on TLS config * More work on bootstrapping * Fix build * More work on bootstrapping * More bootstrapping work * fix build * Remove consul dep * Fix build * merged oss/master into raft-storage * Work on bootstrapping * Get bootstrapping to work * Clean up FMS and node-id * Update local node ID logic * Cleanup node-id change * Work on snapshotting * Raft: Add remove peer API (#906) * Add remove peer API * Add some comments * Fix existing snapshotting (#909) * Raft get peers API (#912) * Read raft configuration * address review feedback * Use the Leadership Transfer API to step-down the active node (#918) * Raft join and unseal using Shamir keys (#917) * Raft join using shamir * Store AEAD instead of master key * Split the raft join process to answer the challenge after a successful unseal * get the follower to standby state * Make unseal work * minor changes * Some input checks * reuse the shamir seal access instead of new default seal access * refactor joinRaftSendAnswer function * Synchronously send answer in auto-unseal case * Address review feedback * Raft snapshots (#910) * Fix existing snapshotting * implement the noop snapshotting * Add comments and switch log libraries * add some snapshot tests * add snapshot test file * add TODO * More work on raft snapshotting * progress on the ConfigStore strategy * Don't use two buckets * Update the snapshot store logic to hide the file logic * Add more backend tests * Cleanup code a bit * [WIP] Raft recovery (#938) * Add recovery functionality * remove fmt.Printfs * Fix a few fsm bugs * Add max size value for raft backend (#942) * Add max size value for raft backend * Include physical.ErrValueTooLarge in the message * Raft snapshot Take/Restore API (#926) * Inital work on raft snapshot APIs * Always redirect snapshot install/download requests * More work on the snapshot APIs * Cleanup code a bit * On restore handle special cases * Use the seal to encrypt the sha sum file * Add sealer mechanism and fix some bugs * Call restore while state lock is held * Send restore cb trigger through raft log * Make error messages nicer * Add test helpers * Add snapshot test * Add shamir unseal test * Add more raft snapshot API tests * Fix locking * Change working to initalize * Add underlying raw object to test cluster core * Move leaderUUID to core * Add raft TLS rotation logic (#950) * Add TLS rotation logic * Cleanup logic a bit * Add/Remove from follower state on add/remove peer * add comments * Update more comments * Update request_forwarding_service.proto * Make sure we populate all nodes in the followerstate obj * Update times * Apply review feedback * Add more raft config setting (#947) * Add performance config setting * Add more config options and fix tests * Test Raft Recovery (#944) * Test raft recovery * Leave out a node during recovery * remove unused struct * Update physical/raft/snapshot_test.go * Update physical/raft/snapshot_test.go * fix vendoring * Switch to new raft interface * Remove unused files * Switch a gogo -> proto instance * Remove unneeded vault dep in go.sum * Update helper/testhelpers/testhelpers.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * Update vault/cluster/cluster.go * track active key within the keyring itself (#6915) * track active key within the keyring itself * lookup and store using the active key ID * update docstring * minor refactor * Small text fixes (#6912) * Update physical/raft/raft.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * review feedback * Move raft logical system into separate file * Update help text a bit * Enforce cluster addr is set and use it for raft bootstrapping * Fix tests * fix http test panic * Pull in latest raft-snapshot library * Add comment
2019-06-20 19:14:58 +00:00
if err := c.startRaftStorage(ctx); err != nil {
return false, err
}
2015-04-14 21:06:15 +00:00
// Do post-unseal setup if HA is not enabled
if c.ha == nil {
// We still need to set up cluster info even if it's not part of a
2017-01-04 21:44:03 +00:00
// cluster right now. This also populates the cached cluster object.
if err := c.setupCluster(ctx); err != nil {
c.logger.Error("cluster setup failed", "error", err)
c.barrier.Seal()
c.logger.Warn("vault is sealed")
return false, err
}
2018-09-18 03:03:00 +00:00
ctx, ctxCancel := context.WithCancel(namespace.RootContext(nil))
if err := c.postUnseal(ctx, ctxCancel, standardUnsealStrategy{}); err != nil {
c.logger.Error("post-unseal setup failed", "error", err)
2015-04-14 21:06:15 +00:00
c.barrier.Seal()
c.logger.Warn("vault is sealed")
2015-04-14 21:06:15 +00:00
return false, err
}
c.standby = false
2015-04-14 21:06:15 +00:00
} else {
// Go to standby mode, wait until we are active to unseal
c.standbyDoneCh = make(chan struct{})
2016-02-29 02:35:32 +00:00
c.manualStepDownCh = make(chan struct{})
c.standbyStopCh = make(chan struct{})
go c.runStandby(c.standbyDoneCh, c.manualStepDownCh, c.standbyStopCh)
2015-03-11 22:19:41 +00:00
}
// Force a cache bust here, which will also run migration code
if c.seal.RecoveryKeySupported() {
c.seal.SetRecoveryConfig(ctx, nil)
}
// Success!
atomic.StoreUint32(c.sealed, 0)
Raft Storage Backend (#6888) * Work on raft backend * Add logstore locally * Add encryptor and unsealable interfaces * Add clustering support to raft * Remove client and handler * Bootstrap raft on init * Cleanup raft logic a bit * More raft work * Work on TLS config * More work on bootstrapping * Fix build * More work on bootstrapping * More bootstrapping work * fix build * Remove consul dep * Fix build * merged oss/master into raft-storage * Work on bootstrapping * Get bootstrapping to work * Clean up FMS and node-id * Update local node ID logic * Cleanup node-id change * Work on snapshotting * Raft: Add remove peer API (#906) * Add remove peer API * Add some comments * Fix existing snapshotting (#909) * Raft get peers API (#912) * Read raft configuration * address review feedback * Use the Leadership Transfer API to step-down the active node (#918) * Raft join and unseal using Shamir keys (#917) * Raft join using shamir * Store AEAD instead of master key * Split the raft join process to answer the challenge after a successful unseal * get the follower to standby state * Make unseal work * minor changes * Some input checks * reuse the shamir seal access instead of new default seal access * refactor joinRaftSendAnswer function * Synchronously send answer in auto-unseal case * Address review feedback * Raft snapshots (#910) * Fix existing snapshotting * implement the noop snapshotting * Add comments and switch log libraries * add some snapshot tests * add snapshot test file * add TODO * More work on raft snapshotting * progress on the ConfigStore strategy * Don't use two buckets * Update the snapshot store logic to hide the file logic * Add more backend tests * Cleanup code a bit * [WIP] Raft recovery (#938) * Add recovery functionality * remove fmt.Printfs * Fix a few fsm bugs * Add max size value for raft backend (#942) * Add max size value for raft backend * Include physical.ErrValueTooLarge in the message * Raft snapshot Take/Restore API (#926) * Inital work on raft snapshot APIs * Always redirect snapshot install/download requests * More work on the snapshot APIs * Cleanup code a bit * On restore handle special cases * Use the seal to encrypt the sha sum file * Add sealer mechanism and fix some bugs * Call restore while state lock is held * Send restore cb trigger through raft log * Make error messages nicer * Add test helpers * Add snapshot test * Add shamir unseal test * Add more raft snapshot API tests * Fix locking * Change working to initalize * Add underlying raw object to test cluster core * Move leaderUUID to core * Add raft TLS rotation logic (#950) * Add TLS rotation logic * Cleanup logic a bit * Add/Remove from follower state on add/remove peer * add comments * Update more comments * Update request_forwarding_service.proto * Make sure we populate all nodes in the followerstate obj * Update times * Apply review feedback * Add more raft config setting (#947) * Add performance config setting * Add more config options and fix tests * Test Raft Recovery (#944) * Test raft recovery * Leave out a node during recovery * remove unused struct * Update physical/raft/snapshot_test.go * Update physical/raft/snapshot_test.go * fix vendoring * Switch to new raft interface * Remove unused files * Switch a gogo -> proto instance * Remove unneeded vault dep in go.sum * Update helper/testhelpers/testhelpers.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * Update vault/cluster/cluster.go * track active key within the keyring itself (#6915) * track active key within the keyring itself * lookup and store using the active key ID * update docstring * minor refactor * Small text fixes (#6912) * Update physical/raft/raft.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * review feedback * Move raft logical system into separate file * Update help text a bit * Enforce cluster addr is set and use it for raft bootstrapping * Fix tests * fix http test panic * Pull in latest raft-snapshot library * Add comment
2019-06-20 19:14:58 +00:00
if c.logger.IsInfo() {
c.logger.Info("vault is unsealed")
}
if c.ha != nil {
sd, ok := c.ha.(physical.ServiceDiscovery)
if ok {
if err := sd.NotifySealedStateChange(); err != nil {
2016-08-19 20:45:17 +00:00
if c.logger.IsWarn() {
c.logger.Warn("failed to notify unsealed status", "error", err)
2016-08-19 20:45:17 +00:00
}
}
}
}
2015-03-11 18:43:36 +00:00
return true, nil
2015-03-09 23:33:27 +00:00
}
2015-03-10 00:45:34 +00:00
// SealWithRequest takes in a logical.Request, acquires the lock, and passes
// through to sealInternal
func (c *Core) SealWithRequest(httpCtx context.Context, req *logical.Request) error {
defer metrics.MeasureSince([]string{"core", "seal-with-request"}, time.Now())
if c.Sealed() {
return nil
}
c.stateLock.RLock()
// This will unlock the read lock
// We use background context since we may not be active
2018-09-18 03:03:00 +00:00
ctx, cancel := context.WithCancel(namespace.RootContext(nil))
defer cancel()
go func() {
select {
case <-ctx.Done():
case <-httpCtx.Done():
cancel()
}
}()
// This will unlock the read lock
return c.sealInitCommon(ctx, req)
}
// Seal takes in a token and creates a logical.Request, acquires the lock, and
// passes through to sealInternal
func (c *Core) Seal(token string) error {
2015-04-08 23:43:17 +00:00
defer metrics.MeasureSince([]string{"core", "seal"}, time.Now())
if c.Sealed() {
return nil
2015-03-10 00:45:34 +00:00
}
2015-03-31 16:59:02 +00:00
c.stateLock.RLock()
req := &logical.Request{
Operation: logical.UpdateOperation,
Path: "sys/seal",
ClientToken: token,
}
// This will unlock the read lock
// We use background context since we may not be active
2018-09-18 03:03:00 +00:00
return c.sealInitCommon(namespace.RootContext(nil), req)
}
// sealInitCommon is common logic for Seal and SealWithRequest and is used to
// re-seal the Vault. This requires the Vault to be unsealed again to perform
// any further operations. Note: this function will read-unlock the state lock.
func (c *Core) sealInitCommon(ctx context.Context, req *logical.Request) (retErr error) {
defer metrics.MeasureSince([]string{"core", "seal-internal"}, time.Now())
var unlocked bool
defer func() {
if !unlocked {
c.stateLock.RUnlock()
}
}()
if req == nil {
retErr = multierror.Append(retErr, errors.New("nil request to seal"))
return retErr
}
// Since there is no token store in standby nodes, sealing cannot be done.
// Ideally, the request has to be forwarded to leader node for validation
// and the operation should be performed. But for now, just returning with
// an error and recommending a vault restart, which essentially does the
// same thing.
if c.standby {
c.logger.Error("vault cannot seal when in standby mode; please restart instead")
retErr = multierror.Append(retErr, errors.New("vault cannot seal when in standby mode; please restart instead"))
return retErr
}
2018-09-18 03:03:00 +00:00
acl, te, entity, identityPolicies, err := c.fetchACLTokenEntryAndEntity(ctx, req)
2015-03-31 16:59:02 +00:00
if err != nil {
retErr = multierror.Append(retErr, err)
return retErr
2015-03-31 16:59:02 +00:00
}
// Audit-log the request before going any further
auth := &logical.Auth{
2018-09-18 03:03:00 +00:00
ClientToken: req.ClientToken,
2018-10-15 16:56:24 +00:00
Accessor: req.ClientTokenAccessor,
}
if te != nil {
2018-09-18 03:03:00 +00:00
auth.IdentityPolicies = identityPolicies[te.NamespaceID]
delete(identityPolicies, te.NamespaceID)
auth.ExternalNamespacePolicies = identityPolicies
auth.TokenPolicies = te.Policies
2018-09-18 03:03:00 +00:00
auth.Policies = append(te.Policies, identityPolicies[te.NamespaceID]...)
auth.Metadata = te.Meta
auth.DisplayName = te.DisplayName
auth.EntityID = te.EntityID
2018-10-15 16:56:24 +00:00
auth.TokenType = te.Type
}
logInput := &logical.LogInput{
Auth: auth,
Request: req,
}
if err := c.auditBroker.LogRequest(ctx, logInput, c.auditedHeaders); err != nil {
c.logger.Error("failed to audit request", "request_path", req.Path, "error", err)
retErr = multierror.Append(retErr, errors.New("failed to audit request, cannot continue"))
return retErr
}
if entity != nil && entity.Disabled {
c.logger.Warn("permission denied as the entity on the token is disabled")
retErr = multierror.Append(retErr, logical.ErrPermissionDenied)
return retErr
}
if te != nil && te.EntityID != "" && entity == nil {
c.logger.Warn("permission denied as the entity on the token is invalid")
retErr = multierror.Append(retErr, logical.ErrPermissionDenied)
return retErr
}
2016-02-29 02:35:32 +00:00
// Attempt to use the token (decrement num_uses)
// On error bail out; if the token has been revoked, bail out too
2016-02-29 02:35:32 +00:00
if te != nil {
te, err = c.tokenStore.UseToken(ctx, te)
if err != nil {
c.logger.Error("failed to use token", "error", err)
retErr = multierror.Append(retErr, ErrInternalError)
return retErr
}
if te == nil {
// Token is no longer valid
retErr = multierror.Append(retErr, logical.ErrPermissionDenied)
return retErr
}
2016-02-29 02:35:32 +00:00
}
2015-03-31 16:59:02 +00:00
// Verify that this operation is allowed
2018-01-19 07:43:36 +00:00
authResults := c.performPolicyChecks(ctx, acl, te, req, entity, &PolicyCheckOpts{
2017-10-23 20:03:36 +00:00
RootPrivsRequired: true,
})
if !authResults.Allowed {
2018-08-11 02:32:10 +00:00
retErr = multierror.Append(retErr, authResults.Error)
if authResults.Error.ErrorOrNil() == nil || authResults.DeniedError {
2018-08-11 02:32:10 +00:00
retErr = multierror.Append(retErr, logical.ErrPermissionDenied)
}
2018-08-11 02:32:10 +00:00
return retErr
}
if te != nil && te.NumUses == tokenRevocationPending {
// Token needs to be revoked. We do this immediately here because
// we won't have a token store after sealing.
leaseID, err := c.expiration.CreateOrFetchRevocationLeaseByToken(c.activeContext, te)
if err == nil {
err = c.expiration.Revoke(c.activeContext, leaseID)
}
if err != nil {
c.logger.Error("token needed revocation before seal but failed to revoke", "error", err)
retErr = multierror.Append(retErr, ErrInternalError)
}
}
// Unlock; sealing will grab the lock when needed
unlocked = true
c.stateLock.RUnlock()
2018-07-25 03:26:28 +00:00
sealErr := c.sealInternal()
if sealErr != nil {
retErr = multierror.Append(retErr, sealErr)
}
return
2015-06-18 01:23:59 +00:00
}
// UIEnabled returns if the UI is enabled
func (c *Core) UIEnabled() bool {
return c.uiConfig.Enabled()
}
// UIHeaders returns configured UI headers
func (c *Core) UIHeaders() (http.Header, error) {
return c.uiConfig.Headers(context.Background())
}
// sealInternal is an internal method used to seal the vault. It does not do
// any authorization checking.
2018-07-25 03:26:28 +00:00
func (c *Core) sealInternal() error {
2019-06-21 00:55:10 +00:00
return c.sealInternalWithOptions(true, false, true)
2018-07-25 03:26:28 +00:00
}
2019-06-21 00:55:10 +00:00
func (c *Core) sealInternalWithOptions(grabStateLock, keepHALock, shutdownRaft bool) error {
// Mark sealed, and if already marked return
if swapped := atomic.CompareAndSwapUint32(c.sealed, 0, 1); !swapped {
return nil
}
c.logger.Info("marked as sealed")
2017-03-01 23:16:47 +00:00
// Clear forwarding clients
c.requestForwardingConnectionLock.Lock()
c.clearForwardingClients()
c.requestForwardingConnectionLock.Unlock()
activeCtxCancel := c.activeContextCancelFunc.Load().(context.CancelFunc)
cancelCtxAndLock := func() {
doneCh := make(chan struct{})
go func() {
select {
case <-doneCh:
// Attempt to drain any inflight requests
case <-time.After(DefaultMaxRequestDuration):
if activeCtxCancel != nil {
activeCtxCancel()
}
}
}()
c.stateLock.Lock()
close(doneCh)
// Stop requests from processing
if activeCtxCancel != nil {
activeCtxCancel()
}
}
2015-04-14 21:06:15 +00:00
// Do pre-seal teardown if HA is not enabled
if c.ha == nil {
if grabStateLock {
cancelCtxAndLock()
defer c.stateLock.Unlock()
}
// Even in a non-HA context we key off of this for some things
c.standby = true
// Stop requests from processing
if activeCtxCancel != nil {
activeCtxCancel()
}
2015-04-14 21:06:15 +00:00
if err := c.preSeal(); err != nil {
c.logger.Error("pre-seal teardown failed", "error", err)
2015-04-14 21:06:15 +00:00
return fmt.Errorf("internal error")
}
} else {
// If we are keeping the lock we already have the state write lock
// held. Otherwise grab it here so that when stopCh is triggered we are
// locked.
2018-07-25 03:26:28 +00:00
if keepHALock {
atomic.StoreUint32(c.keepHALockOnStepDown, 1)
2018-07-25 03:26:28 +00:00
}
if grabStateLock {
cancelCtxAndLock()
defer c.stateLock.Unlock()
}
// If we are trying to acquire the lock, force it to return with nil so
// runStandby will exit
// If we are active, signal the standby goroutine to shut down and wait
// for completion. We have the state lock here so nothing else should
// be toggling standby status.
close(c.standbyStopCh)
c.logger.Debug("finished triggering standbyStopCh for runStandby")
// Wait for runStandby to stop
<-c.standbyDoneCh
atomic.StoreUint32(c.keepHALockOnStepDown, 0)
c.logger.Debug("runStandby done")
2015-03-13 18:16:24 +00:00
}
Raft Storage Backend (#6888) * Work on raft backend * Add logstore locally * Add encryptor and unsealable interfaces * Add clustering support to raft * Remove client and handler * Bootstrap raft on init * Cleanup raft logic a bit * More raft work * Work on TLS config * More work on bootstrapping * Fix build * More work on bootstrapping * More bootstrapping work * fix build * Remove consul dep * Fix build * merged oss/master into raft-storage * Work on bootstrapping * Get bootstrapping to work * Clean up FMS and node-id * Update local node ID logic * Cleanup node-id change * Work on snapshotting * Raft: Add remove peer API (#906) * Add remove peer API * Add some comments * Fix existing snapshotting (#909) * Raft get peers API (#912) * Read raft configuration * address review feedback * Use the Leadership Transfer API to step-down the active node (#918) * Raft join and unseal using Shamir keys (#917) * Raft join using shamir * Store AEAD instead of master key * Split the raft join process to answer the challenge after a successful unseal * get the follower to standby state * Make unseal work * minor changes * Some input checks * reuse the shamir seal access instead of new default seal access * refactor joinRaftSendAnswer function * Synchronously send answer in auto-unseal case * Address review feedback * Raft snapshots (#910) * Fix existing snapshotting * implement the noop snapshotting * Add comments and switch log libraries * add some snapshot tests * add snapshot test file * add TODO * More work on raft snapshotting * progress on the ConfigStore strategy * Don't use two buckets * Update the snapshot store logic to hide the file logic * Add more backend tests * Cleanup code a bit * [WIP] Raft recovery (#938) * Add recovery functionality * remove fmt.Printfs * Fix a few fsm bugs * Add max size value for raft backend (#942) * Add max size value for raft backend * Include physical.ErrValueTooLarge in the message * Raft snapshot Take/Restore API (#926) * Inital work on raft snapshot APIs * Always redirect snapshot install/download requests * More work on the snapshot APIs * Cleanup code a bit * On restore handle special cases * Use the seal to encrypt the sha sum file * Add sealer mechanism and fix some bugs * Call restore while state lock is held * Send restore cb trigger through raft log * Make error messages nicer * Add test helpers * Add snapshot test * Add shamir unseal test * Add more raft snapshot API tests * Fix locking * Change working to initalize * Add underlying raw object to test cluster core * Move leaderUUID to core * Add raft TLS rotation logic (#950) * Add TLS rotation logic * Cleanup logic a bit * Add/Remove from follower state on add/remove peer * add comments * Update more comments * Update request_forwarding_service.proto * Make sure we populate all nodes in the followerstate obj * Update times * Apply review feedback * Add more raft config setting (#947) * Add performance config setting * Add more config options and fix tests * Test Raft Recovery (#944) * Test raft recovery * Leave out a node during recovery * remove unused struct * Update physical/raft/snapshot_test.go * Update physical/raft/snapshot_test.go * fix vendoring * Switch to new raft interface * Remove unused files * Switch a gogo -> proto instance * Remove unneeded vault dep in go.sum * Update helper/testhelpers/testhelpers.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * Update vault/cluster/cluster.go * track active key within the keyring itself (#6915) * track active key within the keyring itself * lookup and store using the active key ID * update docstring * minor refactor * Small text fixes (#6912) * Update physical/raft/raft.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * review feedback * Move raft logical system into separate file * Update help text a bit * Enforce cluster addr is set and use it for raft bootstrapping * Fix tests * fix http test panic * Pull in latest raft-snapshot library * Add comment
2019-06-20 19:14:58 +00:00
// If the storage backend needs to be sealed
2019-06-21 00:55:10 +00:00
if shutdownRaft {
if raftStorage, ok := c.underlyingPhysical.(*raft.RaftBackend); ok {
if err := raftStorage.TeardownCluster(c.getClusterListener()); err != nil {
2019-06-21 00:55:10 +00:00
c.logger.Error("error stopping storage cluster", "error", err)
return err
}
Raft Storage Backend (#6888) * Work on raft backend * Add logstore locally * Add encryptor and unsealable interfaces * Add clustering support to raft * Remove client and handler * Bootstrap raft on init * Cleanup raft logic a bit * More raft work * Work on TLS config * More work on bootstrapping * Fix build * More work on bootstrapping * More bootstrapping work * fix build * Remove consul dep * Fix build * merged oss/master into raft-storage * Work on bootstrapping * Get bootstrapping to work * Clean up FMS and node-id * Update local node ID logic * Cleanup node-id change * Work on snapshotting * Raft: Add remove peer API (#906) * Add remove peer API * Add some comments * Fix existing snapshotting (#909) * Raft get peers API (#912) * Read raft configuration * address review feedback * Use the Leadership Transfer API to step-down the active node (#918) * Raft join and unseal using Shamir keys (#917) * Raft join using shamir * Store AEAD instead of master key * Split the raft join process to answer the challenge after a successful unseal * get the follower to standby state * Make unseal work * minor changes * Some input checks * reuse the shamir seal access instead of new default seal access * refactor joinRaftSendAnswer function * Synchronously send answer in auto-unseal case * Address review feedback * Raft snapshots (#910) * Fix existing snapshotting * implement the noop snapshotting * Add comments and switch log libraries * add some snapshot tests * add snapshot test file * add TODO * More work on raft snapshotting * progress on the ConfigStore strategy * Don't use two buckets * Update the snapshot store logic to hide the file logic * Add more backend tests * Cleanup code a bit * [WIP] Raft recovery (#938) * Add recovery functionality * remove fmt.Printfs * Fix a few fsm bugs * Add max size value for raft backend (#942) * Add max size value for raft backend * Include physical.ErrValueTooLarge in the message * Raft snapshot Take/Restore API (#926) * Inital work on raft snapshot APIs * Always redirect snapshot install/download requests * More work on the snapshot APIs * Cleanup code a bit * On restore handle special cases * Use the seal to encrypt the sha sum file * Add sealer mechanism and fix some bugs * Call restore while state lock is held * Send restore cb trigger through raft log * Make error messages nicer * Add test helpers * Add snapshot test * Add shamir unseal test * Add more raft snapshot API tests * Fix locking * Change working to initalize * Add underlying raw object to test cluster core * Move leaderUUID to core * Add raft TLS rotation logic (#950) * Add TLS rotation logic * Cleanup logic a bit * Add/Remove from follower state on add/remove peer * add comments * Update more comments * Update request_forwarding_service.proto * Make sure we populate all nodes in the followerstate obj * Update times * Apply review feedback * Add more raft config setting (#947) * Add performance config setting * Add more config options and fix tests * Test Raft Recovery (#944) * Test raft recovery * Leave out a node during recovery * remove unused struct * Update physical/raft/snapshot_test.go * Update physical/raft/snapshot_test.go * fix vendoring * Switch to new raft interface * Remove unused files * Switch a gogo -> proto instance * Remove unneeded vault dep in go.sum * Update helper/testhelpers/testhelpers.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * Update vault/cluster/cluster.go * track active key within the keyring itself (#6915) * track active key within the keyring itself * lookup and store using the active key ID * update docstring * minor refactor * Small text fixes (#6912) * Update physical/raft/raft.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * review feedback * Move raft logical system into separate file * Update help text a bit * Enforce cluster addr is set and use it for raft bootstrapping * Fix tests * fix http test panic * Pull in latest raft-snapshot library * Add comment
2019-06-20 19:14:58 +00:00
}
2019-06-21 00:55:10 +00:00
// Stop the cluster listener
c.stopClusterListener()
}
c.logger.Debug("sealing barrier")
if err := c.barrier.Seal(); err != nil {
c.logger.Error("error sealing barrier", "error", err)
return err
}
if c.ha != nil {
sd, ok := c.ha.(physical.ServiceDiscovery)
if ok {
if err := sd.NotifySealedStateChange(); err != nil {
2016-08-19 20:45:17 +00:00
if c.logger.IsWarn() {
c.logger.Warn("failed to notify sealed status", "error", err)
2016-08-19 20:45:17 +00:00
}
}
}
}
2018-09-18 03:03:00 +00:00
postSealInternal(c)
c.logger.Info("vault is sealed")
return nil
2015-03-10 00:45:34 +00:00
}
2015-03-11 22:19:41 +00:00
2018-09-18 03:03:00 +00:00
type UnsealStrategy interface {
unseal(context.Context, log.Logger, *Core) error
}
2018-09-18 03:03:00 +00:00
type standardUnsealStrategy struct{}
2017-01-06 20:42:18 +00:00
2018-09-18 03:03:00 +00:00
func (s standardUnsealStrategy) unseal(ctx context.Context, logger log.Logger, c *Core) error {
2017-03-01 23:16:47 +00:00
// Clear forwarding clients; we're active
c.requestForwardingConnectionLock.Lock()
c.clearForwardingClients()
c.requestForwardingConnectionLock.Unlock()
2018-09-18 03:03:00 +00:00
if err := postUnsealPhysical(c); err != nil {
return err
}
if err := enterprisePostUnseal(c); err != nil {
return err
}
2018-09-18 03:03:00 +00:00
if !c.IsDRSecondary() {
if err := c.ensureWrappingKey(ctx); err != nil {
return err
}
2017-01-04 21:44:03 +00:00
}
2018-11-07 01:21:24 +00:00
if err := c.setupPluginCatalog(ctx); err != nil {
return err
}
if err := c.loadMounts(ctx); err != nil {
2015-03-11 22:19:41 +00:00
return err
}
if err := c.setupMounts(ctx); err != nil {
return err
}
if err := c.setupPolicyStore(ctx); err != nil {
return err
}
if err := c.loadCORSConfig(ctx); err != nil {
2017-06-17 04:04:55 +00:00
return err
}
if err := c.loadCurrentRequestCounters(ctx, time.Now()); err != nil {
return err
}
if err := c.loadCredentials(ctx); err != nil {
return err
}
if err := c.setupCredentials(ctx); err != nil {
return err
2015-03-18 22:30:31 +00:00
}
2018-09-18 03:03:00 +00:00
if !c.IsDRSecondary() {
if err := c.startRollback(); err != nil {
return err
}
if err := c.setupExpiration(expireLeaseStrategyRevoke); err != nil {
return err
}
if err := c.loadAudits(ctx); err != nil {
return err
}
if err := c.setupAudits(ctx); err != nil {
return err
}
if err := c.loadIdentityStoreArtifacts(ctx); err != nil {
return err
}
if err := loadMFAConfigs(ctx, c); err != nil {
return err
}
if err := c.setupAuditedHeadersConfig(ctx); err != nil {
return err
}
} else {
c.auditBroker = NewAuditBroker(c.logger)
Lazy-load plugin mounts (#3255) * Lazy load plugins to avoid setup-unwrap cycle * Remove commented blocks * Refactor NewTestCluster, use single core cluster on basic plugin tests * Set c.pluginDirectory in TestAddTestPlugin for setupPluginCatalog to work properly * Add special path to mock plugin * Move ensureCoresSealed to vault/testing.go * Use same method for EnsureCoresSealed and Cleanup * Bump ensureCoresSealed timeout to 60s * Correctly handle nil opts on NewTestCluster * Add metadata flag to APIClientMeta, use meta-enabled plugin when mounting to bootstrap * Check metadata flag directly on the plugin process * Plumb isMetadataMode down to PluginRunner * Add NOOP shims when running in metadata mode * Remove unused flag from the APIMetadata object * Remove setupSecretPlugins and setupCredentialPlugins functions * Move when we setup rollback manager to after the plugins are initialized * Fix tests * Fix merge issue * start rollback manager after the credential setup * Add guards against running certain client and server functions while in metadata mode * Call initialize once a plugin is loaded on the fly * Add more tests, update basic secret/auth plugin tests to trigger lazy loading * Skip mount if plugin removed from catalog * Fixup * Remove commented line on LookupPlugin * Fail on mount operation if plugin is re-added to catalog and mount is on existing path * Check type and special paths on startBackend * Fix merge conflicts * Refactor PluginRunner run methods to use runCommon, fix TestSystemBackend_Plugin_auth
2017-09-01 05:02:03 +00:00
}
2018-09-18 03:03:00 +00:00
if c.getClusterListener() != nil && (c.ha != nil || shouldStartClusterListener(c)) {
if err := c.setupRaftActiveNode(ctx); err != nil {
return err
}
if err := c.startForwarding(ctx); err != nil {
2018-09-18 03:03:00 +00:00
return err
}
Raft Storage Backend (#6888) * Work on raft backend * Add logstore locally * Add encryptor and unsealable interfaces * Add clustering support to raft * Remove client and handler * Bootstrap raft on init * Cleanup raft logic a bit * More raft work * Work on TLS config * More work on bootstrapping * Fix build * More work on bootstrapping * More bootstrapping work * fix build * Remove consul dep * Fix build * merged oss/master into raft-storage * Work on bootstrapping * Get bootstrapping to work * Clean up FMS and node-id * Update local node ID logic * Cleanup node-id change * Work on snapshotting * Raft: Add remove peer API (#906) * Add remove peer API * Add some comments * Fix existing snapshotting (#909) * Raft get peers API (#912) * Read raft configuration * address review feedback * Use the Leadership Transfer API to step-down the active node (#918) * Raft join and unseal using Shamir keys (#917) * Raft join using shamir * Store AEAD instead of master key * Split the raft join process to answer the challenge after a successful unseal * get the follower to standby state * Make unseal work * minor changes * Some input checks * reuse the shamir seal access instead of new default seal access * refactor joinRaftSendAnswer function * Synchronously send answer in auto-unseal case * Address review feedback * Raft snapshots (#910) * Fix existing snapshotting * implement the noop snapshotting * Add comments and switch log libraries * add some snapshot tests * add snapshot test file * add TODO * More work on raft snapshotting * progress on the ConfigStore strategy * Don't use two buckets * Update the snapshot store logic to hide the file logic * Add more backend tests * Cleanup code a bit * [WIP] Raft recovery (#938) * Add recovery functionality * remove fmt.Printfs * Fix a few fsm bugs * Add max size value for raft backend (#942) * Add max size value for raft backend * Include physical.ErrValueTooLarge in the message * Raft snapshot Take/Restore API (#926) * Inital work on raft snapshot APIs * Always redirect snapshot install/download requests * More work on the snapshot APIs * Cleanup code a bit * On restore handle special cases * Use the seal to encrypt the sha sum file * Add sealer mechanism and fix some bugs * Call restore while state lock is held * Send restore cb trigger through raft log * Make error messages nicer * Add test helpers * Add snapshot test * Add shamir unseal test * Add more raft snapshot API tests * Fix locking * Change working to initalize * Add underlying raw object to test cluster core * Move leaderUUID to core * Add raft TLS rotation logic (#950) * Add TLS rotation logic * Cleanup logic a bit * Add/Remove from follower state on add/remove peer * add comments * Update more comments * Update request_forwarding_service.proto * Make sure we populate all nodes in the followerstate obj * Update times * Apply review feedback * Add more raft config setting (#947) * Add performance config setting * Add more config options and fix tests * Test Raft Recovery (#944) * Test raft recovery * Leave out a node during recovery * remove unused struct * Update physical/raft/snapshot_test.go * Update physical/raft/snapshot_test.go * fix vendoring * Switch to new raft interface * Remove unused files * Switch a gogo -> proto instance * Remove unneeded vault dep in go.sum * Update helper/testhelpers/testhelpers.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * Update vault/cluster/cluster.go * track active key within the keyring itself (#6915) * track active key within the keyring itself * lookup and store using the active key ID * update docstring * minor refactor * Small text fixes (#6912) * Update physical/raft/raft.go Co-Authored-By: Calvin Leung Huang <cleung2010@gmail.com> * review feedback * Move raft logical system into separate file * Update help text a bit * Enforce cluster addr is set and use it for raft bootstrapping * Fix tests * fix http test panic * Pull in latest raft-snapshot library * Add comment
2019-06-20 19:14:58 +00:00
}
2018-09-18 03:03:00 +00:00
c.clusterParamsLock.Lock()
defer c.clusterParamsLock.Unlock()
if err := startReplication(c); err != nil {
return err
}
2018-09-18 03:03:00 +00:00
return nil
}
// postUnseal is invoked on the active node after the barrier is unsealed, but before
2018-09-18 03:03:00 +00:00
// allowing any user operations. This allows us to setup any state that
// requires the Vault to be unsealed such as mount tables, logical backends,
// credential stores, etc.
func (c *Core) postUnseal(ctx context.Context, ctxCancelFunc context.CancelFunc, unsealer UnsealStrategy) (retErr error) {
defer metrics.MeasureSince([]string{"core", "post_unseal"}, time.Now())
// Clear any out
c.postUnsealFuncs = nil
// Create a new request context
c.activeContext = ctx
c.activeContextCancelFunc.Store(ctxCancelFunc)
defer func() {
if retErr != nil {
ctxCancelFunc()
c.preSeal()
}
}()
c.logger.Info("post-unseal setup starting")
// Enable the cache
c.physicalCache.Purge(ctx)
if !c.cachingDisabled {
c.physicalCache.SetEnabled(true)
}
2018-09-18 03:03:00 +00:00
// Purge these for safety in case of a rekey
c.seal.SetBarrierConfig(ctx, nil)
if c.seal.RecoveryKeySupported() {
c.seal.SetRecoveryConfig(ctx, nil)
2017-10-23 20:03:36 +00:00
}
2018-09-18 03:03:00 +00:00
if err := unsealer.unseal(ctx, c.logger, c); err != nil {
return err
}
2017-04-04 00:52:29 +00:00
// Automatically re-encrypt the keys used for auto unsealing when the
// seal's encryption key changes. The regular rotation of cryptographic
// keys is a NIST recommendation. Access to prior keys for decryption
// is normally supported for a configurable time period. Re-encrypting
// the keys used for auto unsealing ensures Vault and its data will
// continue to be accessible even after prior seal keys are destroyed.
if seal, ok := c.seal.(*autoSeal); ok {
if err := seal.UpgradeKeys(c.activeContext); err != nil {
c.logger.Warn("post-unseal upgrade seal keys failed", "error", err)
}
}
2015-04-08 23:43:17 +00:00
c.metricsCh = make(chan struct{})
go c.emitMetrics(c.metricsCh)
// This is intentionally the last block in this function. We want to allow
// writes just before allowing client requests, to ensure everything has
// been set up properly before any writes can have happened.
for _, v := range c.postUnsealFuncs {
v()
}
c.logger.Info("post-unseal setup complete")
2015-03-11 22:19:41 +00:00
return nil
}
2015-03-13 18:16:24 +00:00
// preSeal is invoked before the barrier is sealed, allowing
// for any state teardown required.
2015-11-02 18:29:18 +00:00
func (c *Core) preSeal() error {
2015-04-08 23:43:17 +00:00
defer metrics.MeasureSince([]string{"core", "pre_seal"}, time.Now())
c.logger.Info("pre-seal teardown starting")
2015-05-28 19:07:52 +00:00
// Clear any pending funcs
c.postUnsealFuncs = nil
2015-05-28 19:07:52 +00:00
// Clear any rekey progress
2016-04-04 14:44:22 +00:00
c.barrierRekeyConfig = nil
c.recoveryRekeyConfig = nil
2015-05-28 19:07:52 +00:00
2015-04-08 23:43:17 +00:00
if c.metricsCh != nil {
close(c.metricsCh)
c.metricsCh = nil
}
2015-11-02 18:29:18 +00:00
var result error
2019-06-21 00:55:10 +00:00
c.stopForwarding()
c.stopRaftActiveNode()
2018-09-18 03:03:00 +00:00
c.clusterParamsLock.Lock()
if err := stopReplication(c); err != nil {
result = multierror.Append(result, errwrap.Wrapf("error stopping replication: {{err}}", err))
}
c.clusterParamsLock.Unlock()
if err := c.teardownAudits(); err != nil {
2016-08-19 20:45:17 +00:00
result = multierror.Append(result, errwrap.Wrapf("error tearing down audits: {{err}}", err))
}
if err := c.stopExpiration(); err != nil {
2016-08-19 20:45:17 +00:00
result = multierror.Append(result, errwrap.Wrapf("error stopping expiration: {{err}}", err))
}
if err := c.teardownCredentials(context.Background()); err != nil {
2016-08-19 20:45:17 +00:00
result = multierror.Append(result, errwrap.Wrapf("error tearing down credentials: {{err}}", err))
2015-03-18 22:30:31 +00:00
}
if err := c.teardownPolicyStore(); err != nil {
2016-08-19 20:45:17 +00:00
result = multierror.Append(result, errwrap.Wrapf("error tearing down policy store: {{err}}", err))
}
if err := c.stopRollback(); err != nil {
2016-08-19 20:45:17 +00:00
result = multierror.Append(result, errwrap.Wrapf("error stopping rollback: {{err}}", err))
}
if err := c.unloadMounts(context.Background()); err != nil {
2016-08-19 20:45:17 +00:00
result = multierror.Append(result, errwrap.Wrapf("error unloading mounts: {{err}}", err))
2015-03-13 18:16:24 +00:00
}
if err := enterprisePreSeal(c); err != nil {
result = multierror.Append(result, err)
}
2018-09-18 03:03:00 +00:00
preSealPhysical(c)
c.logger.Info("pre-seal teardown complete")
2015-11-02 18:29:18 +00:00
return result
2015-03-13 18:16:24 +00:00
}
2015-04-08 23:43:17 +00:00
func enterprisePostUnsealImpl(c *Core) error {
return nil
}
func enterprisePreSealImpl(c *Core) error {
return nil
}
func startReplicationImpl(c *Core) error {
return nil
}
func stopReplicationImpl(c *Core) error {
return nil
}
2018-03-20 18:54:10 +00:00
// emitMetrics is used to periodically expose metrics while running
2015-04-08 23:43:17 +00:00
func (c *Core) emitMetrics(stopCh chan struct{}) {
emitTimer := time.Tick(time.Second)
writeTimer := time.Tick(c.counters.syncInterval)
2015-04-08 23:43:17 +00:00
for {
select {
case <-emitTimer:
c.metricsMutex.Lock()
if c.expiration != nil {
c.expiration.emitMetrics()
}
c.metricsMutex.Unlock()
case <-writeTimer:
if stopped := grabLockOrStop(c.stateLock.RLock, c.stateLock.RUnlock, stopCh); stopped {
// Go through the loop again, this time the stop channel case
// should trigger
continue
}
if c.perfStandby {
syncCounter(c)
} else {
err := c.saveCurrentRequestCounters(context.Background(), time.Now())
if err != nil {
c.logger.Error("writing request counters to barrier", "err", err)
}
}
c.stateLock.RUnlock()
2015-04-08 23:43:17 +00:00
case <-stopCh:
return
}
}
}
2016-04-04 14:44:22 +00:00
func (c *Core) ReplicationState() consts.ReplicationState {
return consts.ReplicationState(atomic.LoadUint32(c.replicationState))
}
func (c *Core) ActiveNodeReplicationState() consts.ReplicationState {
return consts.ReplicationState(atomic.LoadUint32(c.activeNodeReplicationState))
}
2016-04-04 14:44:22 +00:00
func (c *Core) SealAccess() *SealAccess {
2017-10-23 20:03:36 +00:00
return NewSealAccess(c.seal)
2016-04-04 14:44:22 +00:00
}
// StorageType returns a string equal to the storage configuration's type.
func (c *Core) StorageType() string {
return c.storageType
}
2016-08-19 20:45:17 +00:00
func (c *Core) Logger() log.Logger {
return c.logger
}
func (c *Core) BarrierKeyLength() (min, max int) {
min, max = c.barrier.KeyLength()
max += shamir.ShareOverhead
return
}
func (c *Core) AuditedHeadersConfig() *AuditedHeadersConfig {
return c.auditedHeaders
}
2017-03-01 17:39:42 +00:00
2018-09-18 03:03:00 +00:00
func waitUntilWALShippedImpl(ctx context.Context, c *Core, index uint64) bool {
return true
}
func merkleRootImpl(c *Core) string {
return ""
}
func lastWALImpl(c *Core) uint64 {
return 0
}
func lastPerformanceWALImpl(c *Core) uint64 {
return 0
}
2017-03-01 17:42:10 +00:00
func lastRemoteWALImpl(c *Core) uint64 {
2017-03-01 17:39:42 +00:00
return 0
}
2017-10-23 20:03:36 +00:00
2018-10-23 06:34:02 +00:00
func (c *Core) PhysicalSealConfigs(ctx context.Context) (*SealConfig, *SealConfig, error) {
pe, err := c.physical.Get(ctx, barrierSealConfigPath)
if err != nil {
return nil, nil, errwrap.Wrapf("failed to fetch barrier seal configuration at migration check time: {{err}}", err)
}
if pe == nil {
return nil, nil, nil
}
barrierConf := new(SealConfig)
if err := jsonutil.DecodeJSON(pe.Value, barrierConf); err != nil {
return nil, nil, errwrap.Wrapf("failed to decode barrier seal configuration at migration check time: {{err}}", err)
}
err = barrierConf.Validate()
if err != nil {
return nil, nil, errwrap.Wrapf("failed to validate barrier seal configuration at migration check time: {{err}}", err)
}
// In older versions of vault the default seal would not store a type. This
// is here to offer backwards compatibility for older seal configs.
if barrierConf.Type == "" {
barrierConf.Type = seal.Shamir
}
2018-10-23 06:34:02 +00:00
var recoveryConf *SealConfig
pe, err = c.physical.Get(ctx, recoverySealConfigPlaintextPath)
if err != nil {
return nil, nil, errwrap.Wrapf("failed to fetch seal configuration at migration check time: {{err}}", err)
}
if pe != nil {
recoveryConf = &SealConfig{}
if err := jsonutil.DecodeJSON(pe.Value, recoveryConf); err != nil {
return nil, nil, errwrap.Wrapf("failed to decode seal configuration at migration check time: {{err}}", err)
}
err = recoveryConf.Validate()
if err != nil {
return nil, nil, errwrap.Wrapf("failed to validate seal configuration at migration check time: {{err}}", err)
}
// In older versions of vault the default seal would not store a type. This
// is here to offer backwards compatibility for older seal configs.
if recoveryConf.Type == "" {
recoveryConf.Type = seal.Shamir
}
2018-10-23 06:34:02 +00:00
}
return barrierConf, recoveryConf, nil
}
func (c *Core) SetSealsForMigration(migrationSeal, newSeal, unwrapSeal Seal) {
2018-10-23 06:34:02 +00:00
c.stateLock.Lock()
defer c.stateLock.Unlock()
c.unwrapSeal = unwrapSeal
if c.unwrapSeal != nil {
c.unwrapSeal.SetCore(c)
}
if newSeal != nil && migrationSeal != nil {
c.migrationSeal = migrationSeal
c.migrationSeal.SetCore(c)
c.seal = newSeal
c.seal.SetCore(c)
c.logger.Warn("entering seal migration mode; Vault will not automatically unseal even if using an autoseal", "from_barrier_type", c.migrationSeal.BarrierType(), "to_barrier_type", c.seal.BarrierType())
c.initSealsForMigration()
}
2018-10-23 06:34:02 +00:00
}
func (c *Core) IsInSealMigration() bool {
c.stateLock.RLock()
defer c.stateLock.RUnlock()
return c.migrationSeal != nil
}
2017-10-23 20:03:36 +00:00
func (c *Core) BarrierEncryptorAccess() *BarrierEncryptorAccess {
return NewBarrierEncryptorAccess(c.barrier)
}
func (c *Core) PhysicalAccess() *physical.PhysicalAccess {
return physical.NewPhysicalAccess(c.physical)
}
func (c *Core) RouterAccess() *RouterAccess {
return NewRouterAccess(c)
}
// IsDRSecondary returns if the current cluster state is a DR secondary.
func (c *Core) IsDRSecondary() bool {
return c.ReplicationState().HasState(consts.ReplicationDRSecondary)
}
func (c *Core) AddLogger(logger log.Logger) {
c.allLoggersLock.Lock()
defer c.allLoggersLock.Unlock()
c.allLoggers = append(c.allLoggers, logger)
}
func (c *Core) SetLogLevel(level log.Level) {
c.allLoggersLock.RLock()
defer c.allLoggersLock.RUnlock()
for _, logger := range c.allLoggers {
logger.SetLevel(level)
}
}
2018-11-07 01:21:24 +00:00
// SetConfig sets core's config object to the newly provided config.
func (c *Core) SetConfig(conf *server.Config) {
c.stateLock.Lock()
c.rawConfig = conf
c.stateLock.Unlock()
}
// SanitizedConfig returns a sanitized version of the current config.
// See server.Config.Sanitized for specific values omitted.
func (c *Core) SanitizedConfig() map[string]interface{} {
c.stateLock.RLock()
defer c.stateLock.RUnlock()
return c.rawConfig.Sanitized()
}
// MetricsHelper returns the global metrics helper which allows external
// packages to access Vault's internal metrics.
func (c *Core) MetricsHelper() *metricsutil.MetricsHelper {
return c.metricsHelper
}
2018-11-07 01:21:24 +00:00
// BuiltinRegistry is an interface that allows the "vault" package to use
// the registry of builtin plugins without getting an import cycle. It
// also allows for mocking the registry easily.
type BuiltinRegistry interface {
Contains(name string, pluginType consts.PluginType) bool
Get(name string, pluginType consts.PluginType) (func() (interface{}, error), bool)
Keys(pluginType consts.PluginType) []string
}