open-vault/builtin/credential/aws/backend.go

295 lines
10 KiB
Go
Raw Normal View History

Create unified aws auth backend (#2441) * Rename builtin/credential/aws-ec2 to aws The aws-ec2 authentication backend is being expanded and will become the generic aws backend. This is a small rename commit to keep the commit history clean. * Expand aws-ec2 backend to more generic aws This adds the ability to authenticate arbitrary AWS IAM principals using AWS's sts:GetCallerIdentity method. The AWS-EC2 auth backend is being to just AWS with the expansion. * Add missing aws auth handler to CLI This was omitted from the previous commit * aws auth backend general variable name cleanup Also fixed a bug where allowed auth types weren't being checked upon login, and added tests for it. * Update docs for the aws auth backend * Refactor aws bind validation * Fix env var override in aws backend test Intent is to override the AWS environment variables with the TEST_* versions if they are set, but the reverse was happening. * Update docs on use of IAM authentication profile AWS now allows you to change the instance profile of a running instance, so the use case of "a long-lived instance that's not in an instance profile" no longer means you have to use the the EC2 auth method. You can now just change the instance profile on the fly. * Fix typo in aws auth cli help * Respond to PR feedback * More PR feedback * Respond to additional PR feedback * Address more feedback on aws auth PR * Make aws auth_type immutable per role * Address more aws auth PR feedback * Address more iam auth PR feedback * Rename aws-ec2.html.md to aws.html.md Per PR feedback, to go along with new backend name. * Add MountType to logical.Request * Make default aws auth_type dependent upon MountType When MountType is aws-ec2, default to ec2 auth_type for backwards compatibility with legacy roles. Otherwise, default to iam. * Pass MountPoint and MountType back up to the core Previously the request router reset the MountPoint and MountType back to the empty string before returning to the core. This ensures they get set back to the correct values.
2017-04-24 19:15:50 +00:00
package awsauth
import (
2017-06-07 14:27:11 +00:00
"fmt"
"sync"
"time"
2017-06-07 14:27:11 +00:00
"github.com/aws/aws-sdk-go/aws/endpoints"
"github.com/aws/aws-sdk-go/service/ec2"
"github.com/aws/aws-sdk-go/service/iam"
"github.com/hashicorp/vault/logical"
"github.com/hashicorp/vault/logical/framework"
"github.com/patrickmn/go-cache"
)
func Factory(conf *logical.BackendConfig) (logical.Backend, error) {
b, err := Backend(conf)
if err != nil {
return nil, err
}
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
if err := b.Setup(conf); err != nil {
return nil, err
}
return b, nil
}
type backend struct {
*framework.Backend
// Lock to make changes to any of the backend's configuration endpoints.
configMutex sync.RWMutex
// Lock to make changes to role entries
roleMutex sync.RWMutex
// Lock to make changes to the blacklist entries
blacklistMutex sync.RWMutex
2016-05-05 18:28:46 +00:00
// Guards the blacklist/whitelist tidy functions
tidyBlacklistCASGuard uint32
tidyWhitelistCASGuard uint32
// Duration after which the periodic function of the backend needs to
// tidy the blacklist and whitelist entries.
tidyCooldownPeriod time.Duration
// nextTidyTime holds the time at which the periodic func should initiatite
// the tidy operations. This is set by the periodicFunc based on the value
// of tidyCooldownPeriod.
nextTidyTime time.Time
// Map to hold the EC2 client objects indexed by region and STS role.
// This avoids the overhead of creating a client object for every login request.
// When the credentials are modified or deleted, all the cached client objects
// will be flushed. The empty STS role signifies the master account
EC2ClientsMap map[string]map[string]*ec2.EC2
// Map to hold the IAM client objects indexed by region and STS role.
// This avoids the overhead of creating a client object for every login request.
// When the credentials are modified or deleted, all the cached client objects
// will be flushed. The empty STS role signifies the master account
IAMClientsMap map[string]map[string]*iam.IAM
2017-06-07 14:27:11 +00:00
// Map of AWS unique IDs to the full ARN corresponding to that unique ID
// This avoids the overhead of an AWS API hit for every login request
// using the IAM auth method when bound_iam_principal_arn contains a wildcard
iamUserIdToArnCache *cache.Cache
2017-06-07 14:27:11 +00:00
// AWS Account ID of the "default" AWS credentials
// This cache avoids the need to call GetCallerIdentity repeatedly to learn it
// We can't store this because, in certain pathological cases, it could change
// out from under us, such as a standby and active Vault server in different AWS
// accounts using their IAM instance profile to get their credentials.
defaultAWSAccountID string
resolveArnToUniqueIDFunc func(logical.Storage, string) (string, error)
}
func Backend(conf *logical.BackendConfig) (*backend, error) {
b := &backend{
// Setting the periodic func to be run once in an hour.
// If there is a real need, this can be made configurable.
tidyCooldownPeriod: time.Hour,
EC2ClientsMap: make(map[string]map[string]*ec2.EC2),
IAMClientsMap: make(map[string]map[string]*iam.IAM),
iamUserIdToArnCache: cache.New(7*24*time.Hour, 24*time.Hour),
}
2017-06-07 14:27:11 +00:00
b.resolveArnToUniqueIDFunc = b.resolveArnToRealUniqueId
b.Backend = &framework.Backend{
PeriodicFunc: b.periodicFunc,
AuthRenew: b.pathLoginRenew,
Help: backendHelp,
PathsSpecial: &logical.Paths{
Unauthenticated: []string{
"login",
},
LocalStorage: []string{
"whitelist/identity/",
},
SealWrapStorage: []string{
"config/client",
},
},
2016-04-30 07:20:21 +00:00
Paths: []*framework.Path{
pathLogin(b),
pathListRole(b),
pathListRoles(b),
pathRole(b),
pathRoleTag(b),
pathConfigClient(b),
pathConfigCertificate(b),
pathConfigSts(b),
pathListSts(b),
pathConfigTidyRoletagBlacklist(b),
pathConfigTidyIdentityWhitelist(b),
pathListCertificates(b),
pathListRoletagBlacklist(b),
pathRoletagBlacklist(b),
pathTidyRoletagBlacklist(b),
pathListIdentityWhitelist(b),
pathIdentityWhitelist(b),
pathTidyIdentityWhitelist(b),
2016-04-30 07:20:21 +00:00
},
Invalidate: b.invalidate,
BackendType: logical.TypeCredential,
}
return b, nil
}
// periodicFunc performs the tasks that the backend wishes to do periodically.
// Currently this will be triggered once in a minute by the RollbackManager.
//
// The tasks being done currently by this function are to cleanup the expired
// entries of both blacklist role tags and whitelist identities. Tidying is done
// not once in a minute, but once in an hour, controlled by 'tidyCooldownPeriod'.
// Tidying of blacklist and whitelist are by default enabled. This can be
// changed using `config/tidy/roletags` and `config/tidy/identities` endpoints.
func (b *backend) periodicFunc(req *logical.Request) error {
// Run the tidy operations for the first time. Then run it when current
// time matches the nextTidyTime.
if b.nextTidyTime.IsZero() || !time.Now().Before(b.nextTidyTime) {
// safety_buffer defaults to 180 days for roletag blacklist
safety_buffer := 15552000
tidyBlacklistConfigEntry, err := b.lockedConfigTidyRoleTags(req.Storage)
if err != nil {
return err
}
skipBlacklistTidy := false
// check if tidying of role tags was configured
if tidyBlacklistConfigEntry != nil {
// check if periodic tidying of role tags was disabled
if tidyBlacklistConfigEntry.DisablePeriodicTidy {
skipBlacklistTidy = true
}
// overwrite the default safety_buffer with the configured value
safety_buffer = tidyBlacklistConfigEntry.SafetyBuffer
}
// tidy role tags if explicitly not disabled
if !skipBlacklistTidy {
2016-05-05 18:28:46 +00:00
b.tidyBlacklistRoleTag(req.Storage, safety_buffer)
}
// reset the safety_buffer to 72h
safety_buffer = 259200
tidyWhitelistConfigEntry, err := b.lockedConfigTidyIdentities(req.Storage)
if err != nil {
return err
}
skipWhitelistTidy := false
// check if tidying of identities was configured
if tidyWhitelistConfigEntry != nil {
// check if periodic tidying of identities was disabled
if tidyWhitelistConfigEntry.DisablePeriodicTidy {
skipWhitelistTidy = true
}
// overwrite the default safety_buffer with the configured value
safety_buffer = tidyWhitelistConfigEntry.SafetyBuffer
}
// tidy identities if explicitly not disabled
if !skipWhitelistTidy {
2016-05-05 18:28:46 +00:00
b.tidyWhitelistIdentity(req.Storage, safety_buffer)
}
// Update the time at which to run the tidy functions again.
b.nextTidyTime = time.Now().Add(b.tidyCooldownPeriod)
}
return nil
}
func (b *backend) invalidate(key string) {
switch key {
case "config/client":
b.configMutex.Lock()
defer b.configMutex.Unlock()
b.flushCachedEC2Clients()
b.flushCachedIAMClients()
2017-06-07 14:27:11 +00:00
b.defaultAWSAccountID = ""
}
}
// Putting this here so we can inject a fake resolver into the backend for unit testing
// purposes
func (b *backend) resolveArnToRealUniqueId(s logical.Storage, arn string) (string, error) {
entity, err := parseIamArn(arn)
if err != nil {
return "", err
}
// This odd-looking code is here because IAM is an inherently global service. IAM and STS ARNs
// don't have regions in them, and there is only a single global endpoint for IAM; see
// http://docs.aws.amazon.com/general/latest/gr/rande.html#iam_region
// However, the ARNs do have a partition in them, because the GovCloud and China partitions DO
// have their own separate endpoints, and the partition is encoded in the ARN. If Amazon's Go SDK
// would allow us to pass a partition back to the IAM client, it would be much simpler. But it
// doesn't appear that's possible, so in order to properly support GovCloud and China, we do a
// circular dance of extracting the partition from the ARN, finding any arbitrary region in the
// partition, and passing that region back back to the SDK, so that the SDK can figure out the
// proper partition from the arbitrary region we passed in to look up the endpoint.
// Sigh
region := getAnyRegionForAwsPartition(entity.Partition)
if region == nil {
return "", fmt.Errorf("Unable to resolve partition %q to a region", entity.Partition)
}
iamClient, err := b.clientIAM(s, region.ID(), entity.AccountNumber)
if err != nil {
return "", err
}
switch entity.Type {
case "user":
userInfo, err := iamClient.GetUser(&iam.GetUserInput{UserName: &entity.FriendlyName})
if err != nil {
return "", err
}
if userInfo == nil {
return "", fmt.Errorf("got nil result from GetUser")
}
return *userInfo.User.UserId, nil
case "role":
roleInfo, err := iamClient.GetRole(&iam.GetRoleInput{RoleName: &entity.FriendlyName})
if err != nil {
return "", err
}
if roleInfo == nil {
return "", fmt.Errorf("got nil result from GetRole")
}
return *roleInfo.Role.RoleId, nil
case "instance-profile":
profileInfo, err := iamClient.GetInstanceProfile(&iam.GetInstanceProfileInput{InstanceProfileName: &entity.FriendlyName})
if err != nil {
return "", err
}
if profileInfo == nil {
return "", fmt.Errorf("got nil result from GetInstanceProfile")
}
return *profileInfo.InstanceProfile.InstanceProfileId, nil
default:
return "", fmt.Errorf("unrecognized error type %#v", entity.Type)
}
}
2017-06-07 14:27:11 +00:00
// Adapted from https://docs.aws.amazon.com/sdk-for-go/api/aws/endpoints/
// the "Enumerating Regions and Endpoint Metadata" section
func getAnyRegionForAwsPartition(partitionId string) *endpoints.Region {
resolver := endpoints.DefaultResolver()
partitions := resolver.(endpoints.EnumPartitions).Partitions()
for _, p := range partitions {
if p.ID() == partitionId {
for _, r := range p.Regions() {
return &r
}
}
}
return nil
}
const backendHelp = `
2016-06-01 14:36:58 +00:00
aws-ec2 auth backend takes in PKCS#7 signature of an AWS EC2 instance and a client
created nonce to authenticates the EC2 instance with Vault.
Authentication is backed by a preconfigured role in the backend. The role
represents the authorization of resources by containing Vault's policies.
2016-05-13 18:31:13 +00:00
Role can be created using 'role/<role>' endpoint.
If there is need to further restrict the capabilities of the role on the instance
that is using the role, 'role_tag' option can be enabled on the role, and a tag
2016-05-13 18:31:13 +00:00
can be generated using 'role/<role>/tag' endpoint. This tag represents the
subset of capabilities set on the role. When the 'role_tag' option is enabled on
the role, the login operation requires that a respective role tag is attached to
the EC2 instance which performs the login.
`