open-consul/agent/consul/connect_ca_endpoint.go

249 lines
6.7 KiB
Go
Raw Normal View History

package consul
import (
"errors"
"fmt"
"time"
"github.com/hashicorp/go-hclog"
"github.com/hashicorp/go-memdb"
"github.com/hashicorp/consul/acl"
"github.com/hashicorp/consul/agent/connect"
"github.com/hashicorp/consul/agent/consul/state"
"github.com/hashicorp/consul/agent/structs"
)
var (
// Err strings. net/rpc doesn't have a way to transport typed/rich errors so
// we currently rely on sniffing the error string in a few cases where we need
// to change client behavior. These are the canonical error strings to use.
// Note though that client code can't use `err == consul.Err*` directly since
// the error returned by RPC will be a plain error.errorString created by
// net/rpc client so will not be the same _instance_ that this package
// variable points to. Clients need to compare using `err.Error() ==
// consul.ErrRateLimited.Error()` which is very sad. Short of replacing our
// RPC mechanism it's hard to know how to make that much better though.
ErrConnectNotEnabled = errors.New("Connect must be enabled in order to use this endpoint")
ErrRateLimited = errors.New("Rate limit reached, try again later")
ErrNotPrimaryDatacenter = errors.New("not the primary datacenter")
ErrStateReadOnly = errors.New("CA Provider State is read-only")
)
const (
// csrLimitWait is the maximum time we'll wait for a slot when CSR concurrency
// limiting or rate limiting is occurring. It's intentionally short so small
// batches of requests can be accommodated when server has capacity (assuming
// signing one cert takes much less than this) but failing requests fast when
// a thundering herd comes along.
csrLimitWait = 500 * time.Millisecond
)
// ConnectCA manages the Connect CA.
type ConnectCA struct {
// srv is a pointer back to the server.
srv *Server
logger hclog.Logger
}
// ConfigurationGet returns the configuration for the CA.
func (s *ConnectCA) ConfigurationGet(
args *structs.DCSpecificRequest,
reply *structs.CAConfiguration) error {
// Exit early if Connect hasn't been enabled.
if !s.srv.config.ConnectEnabled {
return ErrConnectNotEnabled
}
if done, err := s.srv.ForwardRPC("ConnectCA.ConfigurationGet", args, reply); done {
return err
}
// This action requires operator read access.
New ACLs (#4791) This PR is almost a complete rewrite of the ACL system within Consul. It brings the features more in line with other HashiCorp products. Obviously there is quite a bit left to do here but most of it is related docs, testing and finishing the last few commands in the CLI. I will update the PR description and check off the todos as I finish them over the next few days/week. Description At a high level this PR is mainly to split ACL tokens from Policies and to split the concepts of Authorization from Identities. A lot of this PR is mostly just to support CRUD operations on ACLTokens and ACLPolicies. These in and of themselves are not particularly interesting. The bigger conceptual changes are in how tokens get resolved, how backwards compatibility is handled and the separation of policy from identity which could lead the way to allowing for alternative identity providers. On the surface and with a new cluster the ACL system will look very similar to that of Nomads. Both have tokens and policies. Both have local tokens. The ACL management APIs for both are very similar. I even ripped off Nomad's ACL bootstrap resetting procedure. There are a few key differences though. Nomad requires token and policy replication where Consul only requires policy replication with token replication being opt-in. In Consul local tokens only work with token replication being enabled though. All policies in Nomad are globally applicable. In Consul all policies are stored and replicated globally but can be scoped to a subset of the datacenters. This allows for more granular access management. Unlike Nomad, Consul has legacy baggage in the form of the original ACL system. The ramifications of this are: A server running the new system must still support other clients using the legacy system. A client running the new system must be able to use the legacy RPCs when the servers in its datacenter are running the legacy system. The primary ACL DC's servers running in legacy mode needs to be a gate that keeps everything else in the entire multi-DC cluster running in legacy mode. So not only does this PR implement the new ACL system but has a legacy mode built in for when the cluster isn't ready for new ACLs. Also detecting that new ACLs can be used is automatic and requires no configuration on the part of administrators. This process is detailed more in the "Transitioning from Legacy to New ACL Mode" section below.
2018-10-19 16:04:07 +00:00
rule, err := s.srv.ResolveToken(args.Token)
if err != nil {
return err
}
if rule != nil && rule.OperatorWrite(nil) != acl.Allow {
return acl.ErrPermissionDenied
}
state := s.srv.fsm.State()
_, config, err := state.CAConfig(nil)
if err != nil {
return err
}
*reply = *config
return nil
}
// ConfigurationSet updates the configuration for the CA.
func (s *ConnectCA) ConfigurationSet(
args *structs.CARequest,
reply *interface{}) error {
// Exit early if Connect hasn't been enabled.
if !s.srv.config.ConnectEnabled {
return ErrConnectNotEnabled
}
if done, err := s.srv.ForwardRPC("ConnectCA.ConfigurationSet", args, reply); done {
return err
}
// This action requires operator write access.
New ACLs (#4791) This PR is almost a complete rewrite of the ACL system within Consul. It brings the features more in line with other HashiCorp products. Obviously there is quite a bit left to do here but most of it is related docs, testing and finishing the last few commands in the CLI. I will update the PR description and check off the todos as I finish them over the next few days/week. Description At a high level this PR is mainly to split ACL tokens from Policies and to split the concepts of Authorization from Identities. A lot of this PR is mostly just to support CRUD operations on ACLTokens and ACLPolicies. These in and of themselves are not particularly interesting. The bigger conceptual changes are in how tokens get resolved, how backwards compatibility is handled and the separation of policy from identity which could lead the way to allowing for alternative identity providers. On the surface and with a new cluster the ACL system will look very similar to that of Nomads. Both have tokens and policies. Both have local tokens. The ACL management APIs for both are very similar. I even ripped off Nomad's ACL bootstrap resetting procedure. There are a few key differences though. Nomad requires token and policy replication where Consul only requires policy replication with token replication being opt-in. In Consul local tokens only work with token replication being enabled though. All policies in Nomad are globally applicable. In Consul all policies are stored and replicated globally but can be scoped to a subset of the datacenters. This allows for more granular access management. Unlike Nomad, Consul has legacy baggage in the form of the original ACL system. The ramifications of this are: A server running the new system must still support other clients using the legacy system. A client running the new system must be able to use the legacy RPCs when the servers in its datacenter are running the legacy system. The primary ACL DC's servers running in legacy mode needs to be a gate that keeps everything else in the entire multi-DC cluster running in legacy mode. So not only does this PR implement the new ACL system but has a legacy mode built in for when the cluster isn't ready for new ACLs. Also detecting that new ACLs can be used is automatic and requires no configuration on the part of administrators. This process is detailed more in the "Transitioning from Legacy to New ACL Mode" section below.
2018-10-19 16:04:07 +00:00
rule, err := s.srv.ResolveToken(args.Token)
if err != nil {
return err
}
if rule != nil && rule.OperatorWrite(nil) != acl.Allow {
return acl.ErrPermissionDenied
}
return s.srv.caManager.UpdateConfiguration(args)
}
// Roots returns the currently trusted root certificates.
func (s *ConnectCA) Roots(
args *structs.DCSpecificRequest,
reply *structs.IndexedCARoots) error {
// Forward if necessary
if done, err := s.srv.ForwardRPC("ConnectCA.Roots", args, reply); done {
return err
}
// Exit early if Connect hasn't been enabled.
if !s.srv.config.ConnectEnabled {
return ErrConnectNotEnabled
}
return s.srv.blockingQuery(
&args.QueryOptions, &reply.QueryMeta,
func(ws memdb.WatchSet, state *state.Store) error {
roots, err := s.srv.getCARoots(ws, state)
if err != nil {
return err
}
*reply = *roots
return nil
},
)
}
// Sign signs a certificate for a service.
func (s *ConnectCA) Sign(
args *structs.CASignRequest,
reply *structs.IssuedCert) error {
// Exit early if Connect hasn't been enabled.
if !s.srv.config.ConnectEnabled {
return ErrConnectNotEnabled
}
if done, err := s.srv.ForwardRPC("ConnectCA.Sign", args, reply); done {
return err
}
// Parse the CSR
csr, err := connect.ParseCSR(args.CSR)
if err != nil {
return err
}
// Parse the SPIFFE ID
spiffeID, err := connect.ParseCertURI(csr.URIs[0])
if err != nil {
return err
}
// Verify that the ACL token provided has permission to act as this service
rule, err := s.srv.ResolveToken(args.Token)
if err != nil {
return err
}
var authzContext acl.AuthorizerContext
var entMeta structs.EnterpriseMeta
serviceID, isService := spiffeID.(*connect.SpiffeIDService)
agentID, isAgent := spiffeID.(*connect.SpiffeIDAgent)
if !isService && !isAgent {
return fmt.Errorf("SPIFFE ID in CSR must be a service or agent ID")
}
if isService {
entMeta.Merge(serviceID.GetEnterpriseMeta())
entMeta.FillAuthzContext(&authzContext)
if rule != nil && rule.ServiceWrite(serviceID.Service, &authzContext) != acl.Allow {
return acl.ErrPermissionDenied
}
// Verify that the DC in the service URI matches us. We might relax this
// requirement later but being restrictive for now is safer.
if serviceID.Datacenter != s.srv.config.Datacenter {
return fmt.Errorf("SPIFFE ID in CSR from a different datacenter: %s, "+
"we are %s", serviceID.Datacenter, s.srv.config.Datacenter)
}
} else if isAgent {
structs.DefaultEnterpriseMeta().FillAuthzContext(&authzContext)
if rule != nil && rule.NodeWrite(agentID.Agent, &authzContext) != acl.Allow {
return acl.ErrPermissionDenied
}
}
cert, err := s.srv.SignCertificate(csr, spiffeID)
if err != nil {
return err
}
*reply = *cert
return nil
}
// SignIntermediate signs an intermediate certificate for a remote datacenter.
func (s *ConnectCA) SignIntermediate(
args *structs.CASignRequest,
reply *string) error {
// Exit early if Connect hasn't been enabled.
if !s.srv.config.ConnectEnabled {
return ErrConnectNotEnabled
}
if done, err := s.srv.ForwardRPC("ConnectCA.SignIntermediate", args, reply); done {
return err
}
// Verify we are allowed to serve this request
if s.srv.config.PrimaryDatacenter != s.srv.config.Datacenter {
return ErrNotPrimaryDatacenter
}
// This action requires operator write access.
rule, err := s.srv.ResolveToken(args.Token)
if err != nil {
return err
}
if rule != nil && rule.OperatorWrite(nil) != acl.Allow {
return acl.ErrPermissionDenied
}
provider, _ := s.srv.caManager.getCAProvider()
if provider == nil {
return fmt.Errorf("internal error: CA provider is nil")
}
csr, err := connect.ParseCSR(args.CSR)
if err != nil {
return err
}
cert, err := provider.SignIntermediate(csr)
if err != nil {
return err
}
*reply = cert
return nil
}