99e0a124cb
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.
258 lines
7 KiB
Go
258 lines
7 KiB
Go
package consul
|
|
|
|
import (
|
|
"fmt"
|
|
"time"
|
|
|
|
"github.com/armon/go-metrics"
|
|
"github.com/hashicorp/consul/acl"
|
|
"github.com/hashicorp/consul/agent/consul/state"
|
|
"github.com/hashicorp/consul/agent/structs"
|
|
"github.com/hashicorp/go-memdb"
|
|
"github.com/hashicorp/go-uuid"
|
|
)
|
|
|
|
// Session endpoint is used to manipulate sessions for KV
|
|
type Session struct {
|
|
srv *Server
|
|
}
|
|
|
|
// Apply is used to apply a modifying request to the data store. This should
|
|
// only be used for operations that modify the data
|
|
func (s *Session) Apply(args *structs.SessionRequest, reply *string) error {
|
|
if done, err := s.srv.forward("Session.Apply", args, args, reply); done {
|
|
return err
|
|
}
|
|
defer metrics.MeasureSince([]string{"session", "apply"}, time.Now())
|
|
|
|
// Verify the args
|
|
if args.Session.ID == "" && args.Op == structs.SessionDestroy {
|
|
return fmt.Errorf("Must provide ID")
|
|
}
|
|
if args.Session.Node == "" && args.Op == structs.SessionCreate {
|
|
return fmt.Errorf("Must provide Node")
|
|
}
|
|
|
|
// Fetch the ACL token, if any, and apply the policy.
|
|
rule, err := s.srv.ResolveToken(args.Token)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
if rule != nil && s.srv.config.ACLEnforceVersion8 {
|
|
switch args.Op {
|
|
case structs.SessionDestroy:
|
|
state := s.srv.fsm.State()
|
|
_, existing, err := state.SessionGet(nil, args.Session.ID)
|
|
if err != nil {
|
|
return fmt.Errorf("Session lookup failed: %v", err)
|
|
}
|
|
if existing == nil {
|
|
return fmt.Errorf("Unknown session %q", args.Session.ID)
|
|
}
|
|
if !rule.SessionWrite(existing.Node) {
|
|
return acl.ErrPermissionDenied
|
|
}
|
|
|
|
case structs.SessionCreate:
|
|
if !rule.SessionWrite(args.Session.Node) {
|
|
return acl.ErrPermissionDenied
|
|
}
|
|
|
|
default:
|
|
return fmt.Errorf("Invalid session operation %q", args.Op)
|
|
}
|
|
}
|
|
|
|
// Ensure that the specified behavior is allowed
|
|
switch args.Session.Behavior {
|
|
case "":
|
|
// Default behavior to Release for backwards compatibility
|
|
args.Session.Behavior = structs.SessionKeysRelease
|
|
case structs.SessionKeysRelease:
|
|
case structs.SessionKeysDelete:
|
|
default:
|
|
return fmt.Errorf("Invalid Behavior setting '%s'", args.Session.Behavior)
|
|
}
|
|
|
|
// Ensure the Session TTL is valid if provided
|
|
if args.Session.TTL != "" {
|
|
ttl, err := time.ParseDuration(args.Session.TTL)
|
|
if err != nil {
|
|
return fmt.Errorf("Session TTL '%s' invalid: %v", args.Session.TTL, err)
|
|
}
|
|
|
|
if ttl != 0 && (ttl < s.srv.config.SessionTTLMin || ttl > structs.SessionTTLMax) {
|
|
return fmt.Errorf("Invalid Session TTL '%d', must be between [%v=%v]",
|
|
ttl, s.srv.config.SessionTTLMin, structs.SessionTTLMax)
|
|
}
|
|
}
|
|
|
|
// If this is a create, we must generate the Session ID. This must
|
|
// be done prior to appending to the raft log, because the ID is not
|
|
// deterministic. Once the entry is in the log, the state update MUST
|
|
// be deterministic or the followers will not converge.
|
|
if args.Op == structs.SessionCreate {
|
|
// Generate a new session ID, verify uniqueness
|
|
state := s.srv.fsm.State()
|
|
for {
|
|
var err error
|
|
if args.Session.ID, err = uuid.GenerateUUID(); err != nil {
|
|
s.srv.logger.Printf("[ERR] consul.session: UUID generation failed: %v", err)
|
|
return err
|
|
}
|
|
_, sess, err := state.SessionGet(nil, args.Session.ID)
|
|
if err != nil {
|
|
s.srv.logger.Printf("[ERR] consul.session: Session lookup failed: %v", err)
|
|
return err
|
|
}
|
|
if sess == nil {
|
|
break
|
|
}
|
|
}
|
|
}
|
|
|
|
// Apply the update
|
|
resp, err := s.srv.raftApply(structs.SessionRequestType, args)
|
|
if err != nil {
|
|
s.srv.logger.Printf("[ERR] consul.session: Apply failed: %v", err)
|
|
return err
|
|
}
|
|
|
|
if args.Op == structs.SessionCreate && args.Session.TTL != "" {
|
|
// If we created a session with a TTL, reset the expiration timer
|
|
s.srv.resetSessionTimer(args.Session.ID, &args.Session)
|
|
} else if args.Op == structs.SessionDestroy {
|
|
// If we destroyed a session, it might potentially have a TTL,
|
|
// and we need to clear the timer
|
|
s.srv.clearSessionTimer(args.Session.ID)
|
|
}
|
|
|
|
if respErr, ok := resp.(error); ok {
|
|
return respErr
|
|
}
|
|
|
|
// Check if the return type is a string
|
|
if respString, ok := resp.(string); ok {
|
|
*reply = respString
|
|
}
|
|
return nil
|
|
}
|
|
|
|
// Get is used to retrieve a single session
|
|
func (s *Session) Get(args *structs.SessionSpecificRequest,
|
|
reply *structs.IndexedSessions) error {
|
|
if done, err := s.srv.forward("Session.Get", args, args, reply); done {
|
|
return err
|
|
}
|
|
|
|
return s.srv.blockingQuery(
|
|
&args.QueryOptions,
|
|
&reply.QueryMeta,
|
|
func(ws memdb.WatchSet, state *state.Store) error {
|
|
index, session, err := state.SessionGet(ws, args.Session)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
reply.Index = index
|
|
if session != nil {
|
|
reply.Sessions = structs.Sessions{session}
|
|
} else {
|
|
reply.Sessions = nil
|
|
}
|
|
if err := s.srv.filterACL(args.Token, reply); err != nil {
|
|
return err
|
|
}
|
|
return nil
|
|
})
|
|
}
|
|
|
|
// List is used to list all the active sessions
|
|
func (s *Session) List(args *structs.DCSpecificRequest,
|
|
reply *structs.IndexedSessions) error {
|
|
if done, err := s.srv.forward("Session.List", args, args, reply); done {
|
|
return err
|
|
}
|
|
|
|
return s.srv.blockingQuery(
|
|
&args.QueryOptions,
|
|
&reply.QueryMeta,
|
|
func(ws memdb.WatchSet, state *state.Store) error {
|
|
index, sessions, err := state.SessionList(ws)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
reply.Index, reply.Sessions = index, sessions
|
|
if err := s.srv.filterACL(args.Token, reply); err != nil {
|
|
return err
|
|
}
|
|
return nil
|
|
})
|
|
}
|
|
|
|
// NodeSessions is used to get all the sessions for a particular node
|
|
func (s *Session) NodeSessions(args *structs.NodeSpecificRequest,
|
|
reply *structs.IndexedSessions) error {
|
|
if done, err := s.srv.forward("Session.NodeSessions", args, args, reply); done {
|
|
return err
|
|
}
|
|
|
|
return s.srv.blockingQuery(
|
|
&args.QueryOptions,
|
|
&reply.QueryMeta,
|
|
func(ws memdb.WatchSet, state *state.Store) error {
|
|
index, sessions, err := state.NodeSessions(ws, args.Node)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
reply.Index, reply.Sessions = index, sessions
|
|
if err := s.srv.filterACL(args.Token, reply); err != nil {
|
|
return err
|
|
}
|
|
return nil
|
|
})
|
|
}
|
|
|
|
// Renew is used to renew the TTL on a single session
|
|
func (s *Session) Renew(args *structs.SessionSpecificRequest,
|
|
reply *structs.IndexedSessions) error {
|
|
if done, err := s.srv.forward("Session.Renew", args, args, reply); done {
|
|
return err
|
|
}
|
|
defer metrics.MeasureSince([]string{"session", "renew"}, time.Now())
|
|
|
|
// Get the session, from local state.
|
|
state := s.srv.fsm.State()
|
|
index, session, err := state.SessionGet(nil, args.Session)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
reply.Index = index
|
|
if session == nil {
|
|
return nil
|
|
}
|
|
|
|
// Fetch the ACL token, if any, and apply the policy.
|
|
rule, err := s.srv.ResolveToken(args.Token)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
if rule != nil && s.srv.config.ACLEnforceVersion8 {
|
|
if !rule.SessionWrite(session.Node) {
|
|
return acl.ErrPermissionDenied
|
|
}
|
|
}
|
|
|
|
// Reset the session TTL timer.
|
|
reply.Sessions = structs.Sessions{session}
|
|
if err := s.srv.resetSessionTimer(args.Session, session); err != nil {
|
|
s.srv.logger.Printf("[ERR] consul.session: Session renew failed: %v", err)
|
|
return err
|
|
}
|
|
|
|
return nil
|
|
}
|