// Copyright (c) HashiCorp, Inc. // SPDX-License-Identifier: MPL-2.0 package state import ( "errors" "fmt" "net" "reflect" "strings" "github.com/hashicorp/go-memdb" "github.com/hashicorp/consul/acl" "github.com/hashicorp/consul/agent/configentry" "github.com/hashicorp/consul/agent/structs" "github.com/hashicorp/consul/api" "github.com/hashicorp/consul/lib" "github.com/hashicorp/consul/lib/maps" "github.com/hashicorp/consul/types" ) const ( // indexServiceExtinction keeps track of the last raft index when the last instance // of any service was unregistered. This is used by blocking queries on missing services. indexServiceExtinction = "service_last_extinction" // indexNodeExtinction keeps track of the last raft index when the last instance // of any node was unregistered. This is used by blocking queries on missing nodes. indexNodeExtinction = "node_last_extinction" ) const ( // minUUIDLookupLen is used as a minimum length of a node name required before // we test to see if the name is actually a UUID and perform an ID-based node // lookup. minUUIDLookupLen = 2 ) var ( // startingVirtualIP is the start of the virtual IP range we assign to services. // The effective CIDR range is startingVirtualIP to (startingVirtualIP + virtualIPMaxOffset). startingVirtualIP = net.IP{240, 0, 0, 0} virtualIPMaxOffset = net.IP{15, 255, 255, 254} ErrNodeNotFound = errors.New("node not found") ) func resizeNodeLookupKey(s string) string { l := len(s) if l%2 != 0 { return s[0 : l-1] } return s } // Nodes is used to pull the full list of nodes for use during snapshots. func (s *Snapshot) Nodes() (memdb.ResultIterator, error) { iter, err := s.tx.Get(tableNodes, indexID) if err != nil { return nil, err } return iter, nil } // Services is used to pull the full list of services for a given node for use // during snapshots. func (s *Snapshot) Services(node string, entMeta *acl.EnterpriseMeta, peerName string) (memdb.ResultIterator, error) { // TODO: accept non-pointer value if entMeta == nil { entMeta = structs.NodeEnterpriseMetaInDefaultPartition() } return s.tx.Get(tableServices, indexNode, Query{ Value: node, EnterpriseMeta: *entMeta, PeerName: peerName, }) } // Checks is used to pull the full list of checks for a given node for use // during snapshots. func (s *Snapshot) Checks(node string, entMeta *acl.EnterpriseMeta, peerName string) (memdb.ResultIterator, error) { // TODO: accept non-pointer value if entMeta == nil { entMeta = structs.NodeEnterpriseMetaInDefaultPartition() } return s.tx.Get(tableChecks, indexNode, Query{ Value: node, EnterpriseMeta: *entMeta, PeerName: peerName, }) } // ServiceVirtualIPs is used to pull the service virtual IP mappings for use during snapshots. func (s *Snapshot) ServiceVirtualIPs() (memdb.ResultIterator, error) { iter, err := s.tx.Get(tableServiceVirtualIPs, indexID) if err != nil { return nil, err } return iter, nil } // FreeVirtualIPs is used to pull the freed virtual IPs for use during snapshots. func (s *Snapshot) FreeVirtualIPs() (memdb.ResultIterator, error) { iter, err := s.tx.Get(tableFreeVirtualIPs, indexID) if err != nil { return nil, err } return iter, nil } // Registration is used to make sure a node, service, and check registration is // performed within a single transaction to avoid race conditions on state // updates. func (s *Restore) Registration(idx uint64, req *structs.RegisterRequest) error { return s.store.ensureRegistrationTxn(s.tx, idx, true, req, true) } func (s *Restore) ServiceVirtualIP(req ServiceVirtualIP) error { if err := s.tx.Insert(tableServiceVirtualIPs, req); err != nil { return err } if err := updateVirtualIPMaxIndexes(s.tx, req.ModifyIndex, req.Service.ServiceName.PartitionOrDefault(), req.Service.Peer); err != nil { return err } return nil } func (s *Restore) FreeVirtualIP(req FreeVirtualIP) error { return s.tx.Insert(tableFreeVirtualIPs, req) } // EnsureRegistration is used to make sure a node, service, and check // registration is performed within a single transaction to avoid race // conditions on state updates. func (s *Store) EnsureRegistration(idx uint64, req *structs.RegisterRequest) error { tx := s.db.WriteTxn(idx) defer tx.Abort() if err := s.ensureRegistrationTxn(tx, idx, false, req, false); err != nil { return err } return tx.Commit() } func (s *Store) ensureCheckIfNodeMatches( tx WriteTxn, idx uint64, preserveIndexes bool, node string, nodePartition string, nodePeerName string, check *structs.HealthCheck, ) error { if !strings.EqualFold(check.PeerName, nodePeerName) { return fmt.Errorf("check peer name %q does not match node peer name %q", check.PeerName, nodePeerName) } if !strings.EqualFold(check.Node, node) || !acl.EqualPartitions(nodePartition, check.PartitionOrDefault()) { return fmt.Errorf("check node %q does not match node %q", printNodeName(check.Node, check.PartitionOrDefault()), printNodeName(node, nodePartition), ) } if err := s.ensureCheckTxn(tx, idx, preserveIndexes, check); err != nil { return fmt.Errorf("failed inserting check on node %q: %v", printNodeName(check.Node, check.PartitionOrDefault()), err) } return nil } func printNodeName(nodeName, partition string) string { if acl.IsDefaultPartition(partition) { return nodeName } return partition + "/" + nodeName } // ensureRegistrationTxn is used to make sure a node, service, and check // registration is performed within a single transaction to avoid race // conditions on state updates. func (s *Store) ensureRegistrationTxn(tx WriteTxn, idx uint64, preserveIndexes bool, req *structs.RegisterRequest, restore bool) error { if err := validateRegisterRequestPeerNamesTxn(tx, req, restore); err != nil { return err } if _, err := validateRegisterRequestTxn(tx, req, restore); err != nil { return err } // Create a node structure. node := &structs.Node{ ID: req.ID, Node: req.Node, Address: req.Address, Datacenter: req.Datacenter, Partition: req.PartitionOrDefault(), TaggedAddresses: req.TaggedAddresses, Meta: req.NodeMeta, PeerName: req.PeerName, Locality: req.Locality, } if preserveIndexes { node.CreateIndex = req.CreateIndex node.ModifyIndex = req.ModifyIndex } // Since this gets called for all node operations (service and check // updates) and churn on the node itself is basically none after the // node updates itself the first time, it's worth seeing if we need to // modify the node at all so we prevent watch churn and useless writes // and modify index bumps on the node. { existing, err := tx.First(tableNodes, indexID, Query{ Value: node.Node, EnterpriseMeta: *node.GetEnterpriseMeta(), PeerName: node.PeerName, }) if err != nil { return fmt.Errorf("node lookup failed: %s", err) } if existing == nil || req.ChangesNode(existing.(*structs.Node)) { if err := s.ensureNodeTxn(tx, idx, preserveIndexes, node); err != nil { return fmt.Errorf("failed inserting node: %s", err) } } } // Add the service, if any. We perform a similar check as we do for the // node info above to make sure we actually need to update the service // definition in order to prevent useless churn if nothing has changed. if req.Service != nil { existing, err := tx.First(tableServices, indexID, NodeServiceQuery{ EnterpriseMeta: req.Service.EnterpriseMeta, Node: req.Node, Service: req.Service.ID, PeerName: req.PeerName, }) if err != nil { return fmt.Errorf("failed service lookup: %s", err) } if existing == nil || !(existing.(*structs.ServiceNode).ToNodeService()).IsSame(req.Service) { if err := ensureServiceTxn(tx, idx, req.Node, preserveIndexes, req.Service); err != nil { return fmt.Errorf("failed inserting service: %s", err) } } } // Add the checks, if any. if req.Check != nil { err := s.ensureCheckIfNodeMatches(tx, idx, preserveIndexes, req.Node, req.PartitionOrDefault(), req.PeerName, req.Check) if err != nil { return err } } for _, check := range req.Checks { err := s.ensureCheckIfNodeMatches(tx, idx, preserveIndexes, req.Node, req.PartitionOrDefault(), req.PeerName, check) if err != nil { return err } } return nil } func validateRegisterRequestPeerNamesTxn(_ ReadTxn, args *structs.RegisterRequest, _ bool) error { var ( peerNames = make(map[string]struct{}) ) if args.Service != nil { if args.Service.PeerName == "" { args.Service.PeerName = args.PeerName } peerName := args.Service.PeerName // TODO(peering): validate the peering exists (skip check on restore) peerNames[peerName] = struct{}{} } validateCheck := func(chk *structs.HealthCheck) error { if chk.PeerName == "" { chk.PeerName = args.PeerName } peerName := chk.PeerName // TODO(peering): validate the peering exists (skip check on restore) peerNames[peerName] = struct{}{} return nil } if args.Check != nil { if err := validateCheck(args.Check); err != nil { return err } } for _, chk := range args.Checks { if err := validateCheck(chk); err != nil { return err } } { // TODO(peering): validate the node's peering exists (skip check on restore) peerName := args.PeerName peerNames[peerName] = struct{}{} } if len(peerNames) > 1 { return fmt.Errorf("Cannot register services and checks for multiple peer names in one registration request") } else if len(peerNames) == 0 { return fmt.Errorf("No peer names are present on the registration request; this makes no sense") } return nil } // EnsureNode is used to upsert node registration or modification. func (s *Store) EnsureNode(idx uint64, node *structs.Node) error { tx := s.db.WriteTxn(idx) defer tx.Abort() // Call the node upsert if err := s.ensureNodeTxn(tx, idx, false, node); err != nil { return err } return tx.Commit() } // ensureNoNodeWithSimilarNameTxn checks that no other node has conflict in its name // If allowClashWithoutID then, getting a conflict on another node without ID will be allowed func ensureNoNodeWithSimilarNameTxn(tx ReadTxn, node *structs.Node, allowClashWithoutID bool) error { // Retrieve all of the nodes q := Query{ PeerName: node.PeerName, EnterpriseMeta: *node.GetEnterpriseMeta(), } enodes, err := tx.Get(tableNodes, indexID+"_prefix", q) if err != nil { return fmt.Errorf("Cannot lookup all nodes: %s", err) } for nodeIt := enodes.Next(); nodeIt != nil; nodeIt = enodes.Next() { enode := nodeIt.(*structs.Node) if strings.EqualFold(node.Node, enode.Node) && node.ID != enode.ID { // Look up the existing node's Serf health check to see if it's failed. // If it is, the node can be renamed. enodeCheck, err := tx.First(tableChecks, indexID, NodeCheckQuery{ EnterpriseMeta: *node.GetEnterpriseMeta(), Node: enode.Node, CheckID: string(structs.SerfCheckID), PeerName: enode.PeerName, }) if err != nil { return fmt.Errorf("Cannot get status of node %s: %s", enode.Node, err) } // Get the node health. If there's no Serf health check, we consider it safe to rename // the node as it's likely an external node registration not managed by Consul. var nodeHealthy bool if enodeCheck != nil { enodeSerfCheck, ok := enodeCheck.(*structs.HealthCheck) if ok { nodeHealthy = enodeSerfCheck.Status != api.HealthCritical } } if !(enode.ID == "" && allowClashWithoutID) && nodeHealthy { return fmt.Errorf("Node name %s is reserved by node %s with name %s (%s)", node.Node, enode.ID, enode.Node, enode.Address) } } } return nil } // ensureNodeCASTxn updates a node only if the existing index matches the given index. // Returns a bool indicating if a write happened and any error. func (s *Store) ensureNodeCASTxn(tx WriteTxn, idx uint64, node *structs.Node) (bool, error) { // Retrieve the existing entry. existing, err := getNodeTxn(tx, node.Node, node.GetEnterpriseMeta(), node.PeerName) if err != nil { return false, err } // Check if the we should do the set. A ModifyIndex of 0 means that // we are doing a set-if-not-exists. if node.ModifyIndex == 0 && existing != nil { return false, nil } if node.ModifyIndex != 0 && existing == nil { return false, nil } if existing != nil && node.ModifyIndex != 0 && node.ModifyIndex != existing.ModifyIndex { return false, nil } // Perform the update. if err := s.ensureNodeTxn(tx, idx, false, node); err != nil { return false, err } return true, nil } // ensureNodeTxn is the inner function called to actually create a node // registration or modify an existing one in the state store. It allows // passing in a memdb transaction so it may be part of a larger txn. func (s *Store) ensureNodeTxn(tx WriteTxn, idx uint64, preserveIndexes bool, node *structs.Node) error { // See if there's an existing node with this UUID, and make sure the // name is the same. var n *structs.Node if node.ID != "" { existing, err := getNodeIDTxn(tx, node.ID, node.GetEnterpriseMeta(), node.PeerName) if err != nil { return fmt.Errorf("node lookup failed: %s", err) } if existing != nil { n = existing if !strings.EqualFold(n.Node, node.Node) { // Lets first get all nodes and check whether name do match, we do not allow clash on nodes without ID dupNameError := ensureNoNodeWithSimilarNameTxn(tx, node, false) if dupNameError != nil { return fmt.Errorf("Error while renaming Node ID: %q (%s): %s", node.ID, node.Address, dupNameError) } // We are actually renaming a node, remove its reference first err := s.deleteNodeTxn(tx, idx, n.Node, n.GetEnterpriseMeta(), n.PeerName) if err != nil { return fmt.Errorf("Error while renaming Node ID: %q (%s) from %s to %s: %w", node.ID, node.Address, n.Node, node.Node, err) } } } else { // We allow to "steal" another node name that would have no ID // It basically means that we allow upgrading a node without ID and add the ID dupNameError := ensureNoNodeWithSimilarNameTxn(tx, node, true) if dupNameError != nil { return fmt.Errorf("Error while renaming Node ID: %q: %s", node.ID, dupNameError) } } } // TODO: else Node.ID == "" should be forbidden in future Consul releases // See https://github.com/hashicorp/consul/pull/3983 for context // Check for an existing node by name to support nodes with no IDs. if n == nil { existing, err := tx.First(tableNodes, indexID, Query{ Value: node.Node, EnterpriseMeta: *node.GetEnterpriseMeta(), PeerName: node.PeerName, }) if err != nil { return fmt.Errorf("node name lookup failed: %s", err) } if existing != nil { n = existing.(*structs.Node) } // WARNING, for compatibility reasons with tests, we do not check // for case insensitive matches, which may lead to DB corruption // See https://github.com/hashicorp/consul/pull/3983 for context } // Get the indexes. if n != nil { node.CreateIndex = n.CreateIndex node.ModifyIndex = n.ModifyIndex // We do not need to update anything if node.IsSame(n) { return nil } node.ModifyIndex = idx } else if !preserveIndexes || node.CreateIndex == 0 { // If this isn't a snapshot or there were no saved indexes, set CreateIndex // and ModifyIndex from the given index. Prior to 1.9.0/1.8.3/1.7.7, nodes // were not saved with an index, so this is to avoid ending up with a 0 index // when loading a snapshot from an older version. node.CreateIndex = idx node.ModifyIndex = idx } // Insert the node and update the index. return catalogInsertNode(tx, node) } // GetNode is used to retrieve a node registration by node name ID. func (s *Store) GetNode(nodeNameOrID string, entMeta *acl.EnterpriseMeta, peerName string) (uint64, *structs.Node, error) { tx := s.db.Txn(false) defer tx.Abort() // TODO: accept non-pointer value if entMeta == nil { entMeta = structs.NodeEnterpriseMetaInDefaultPartition() } // Get the table index. idx := catalogNodesMaxIndex(tx, entMeta, peerName) // Retrieve the node from the state store node, err := getNodeTxn(tx, nodeNameOrID, entMeta, peerName) if err != nil { return 0, nil, fmt.Errorf("node lookup failed: %s", err) } return idx, node, nil } func getNodeTxn(tx ReadTxn, nodeNameOrID string, entMeta *acl.EnterpriseMeta, peerName string) (*structs.Node, error) { node, err := tx.First(tableNodes, indexID, Query{ Value: nodeNameOrID, EnterpriseMeta: *entMeta, PeerName: peerName, }) if err != nil { return nil, fmt.Errorf("node lookup failed: %s", err) } if node != nil { return node.(*structs.Node), nil } return nil, nil } func getNodeIDTxn(tx ReadTxn, id types.NodeID, entMeta *acl.EnterpriseMeta, peerName string) (*structs.Node, error) { node, err := tx.First(tableNodes, indexUUID+"_prefix", Query{ Value: string(id), EnterpriseMeta: *entMeta, PeerName: peerName, }) if err != nil { return nil, fmt.Errorf("node lookup by ID failed: %s", err) } if node != nil { return node.(*structs.Node), nil } return nil, nil } // GetNodeID is used to retrieve a node registration by node ID. func (s *Store) GetNodeID(id types.NodeID, entMeta *acl.EnterpriseMeta, peerName string) (uint64, *structs.Node, error) { tx := s.db.Txn(false) defer tx.Abort() // TODO: accept non-pointer value if entMeta == nil { entMeta = structs.NodeEnterpriseMetaInDefaultPartition() } idx := catalogNodesMaxIndex(tx, entMeta, peerName) // Retrieve the node from the state store node, err := getNodeIDTxn(tx, id, entMeta, peerName) return idx, node, err } // Nodes is used to return all of the known nodes. func (s *Store) Nodes(ws memdb.WatchSet, entMeta *acl.EnterpriseMeta, peerName string) (uint64, structs.Nodes, error) { tx := s.db.Txn(false) defer tx.Abort() // TODO: accept non-pointer value if entMeta == nil { entMeta = structs.NodeEnterpriseMetaInDefaultPartition() } idx := catalogNodesMaxIndex(tx, entMeta, peerName) // Retrieve all of the nodes q := Query{ PeerName: peerName, EnterpriseMeta: *entMeta, } nodes, err := tx.Get(tableNodes, indexID+"_prefix", q) if err != nil { return 0, nil, fmt.Errorf("failed nodes lookup: %s", err) } ws.Add(nodes.WatchCh()) // Create and return the nodes list. var results structs.Nodes for node := nodes.Next(); node != nil; node = nodes.Next() { results = append(results, node.(*structs.Node)) } return idx, results, nil } // NodesByMeta is used to return all nodes with the given metadata key/value pairs. func (s *Store) NodesByMeta(ws memdb.WatchSet, filters map[string]string, entMeta *acl.EnterpriseMeta, peerName string) (uint64, structs.Nodes, error) { tx := s.db.Txn(false) defer tx.Abort() // TODO: accept non-pointer value if entMeta == nil { entMeta = structs.NodeEnterpriseMetaInDefaultPartition() } idx := catalogNodesMaxIndex(tx, entMeta, peerName) if len(filters) == 0 { return idx, nil, nil // NodesByMeta is never called with an empty map, but just in case make it return no results. } // Retrieve all of the nodes. We'll do a lookup of just ONE KV pair, which // over-matches if multiple pairs are requested, but then in the loop below // we'll finish filtering. var firstKey, firstValue string for firstKey, firstValue = range filters { break } nodes, err := tx.Get(tableNodes, indexMeta, KeyValueQuery{ Key: firstKey, Value: firstValue, EnterpriseMeta: *entMeta, PeerName: peerName, }) if err != nil { return 0, nil, fmt.Errorf("failed nodes lookup: %s", err) } ws.Add(nodes.WatchCh()) // Create and return the nodes list. var results structs.Nodes for node := nodes.Next(); node != nil; node = nodes.Next() { n := node.(*structs.Node) if len(filters) <= 1 || structs.SatisfiesMetaFilters(n.Meta, filters) { results = append(results, n) } } return idx, results, nil } // DeleteNode is used to delete a given node by its ID. func (s *Store) DeleteNode(idx uint64, nodeName string, entMeta *acl.EnterpriseMeta, peerName string) error { tx := s.db.WriteTxn(idx) defer tx.Abort() // TODO: accept non-pointer value if entMeta == nil { entMeta = structs.NodeEnterpriseMetaInDefaultPartition() } // Call the node deletion. if err := s.deleteNodeTxn(tx, idx, nodeName, entMeta, peerName); err != nil { return err } return tx.Commit() } // deleteNodeCASTxn is used to try doing a node delete operation with a given // raft index. If the CAS index specified is not equal to the last observed index for // the given check, then the call is a noop, otherwise a normal check delete is invoked. func (s *Store) deleteNodeCASTxn(tx WriteTxn, idx, cidx uint64, nodeName string, entMeta *acl.EnterpriseMeta, peerName string) (bool, error) { // Look up the node. node, err := getNodeTxn(tx, nodeName, entMeta, peerName) if err != nil { return false, err } if node == nil { return false, nil } // If the existing index does not match the provided CAS // index arg, then we shouldn't update anything and can safely // return early here. if node.ModifyIndex != cidx { return false, nil } // Call the actual deletion if the above passed. if err := s.deleteNodeTxn(tx, idx, nodeName, entMeta, peerName); err != nil { return false, err } return true, nil } // deleteNodeTxn is the inner method used for removing a node from // the store within a given transaction. func (s *Store) deleteNodeTxn(tx WriteTxn, idx uint64, nodeName string, entMeta *acl.EnterpriseMeta, peerName string) error { // TODO: accept non-pointer value if entMeta == nil { entMeta = structs.DefaultEnterpriseMetaInDefaultPartition() } // Look up the node. nodeRaw, err := tx.First(tableNodes, indexID, Query{ Value: nodeName, EnterpriseMeta: *entMeta, PeerName: peerName, }) if err != nil { return fmt.Errorf("node lookup failed: %s", err) } if nodeRaw == nil { return nil } // Delete all services associated with the node and update the service index. services, err := tx.Get(tableServices, indexNode, Query{ Value: nodeName, EnterpriseMeta: *entMeta, PeerName: peerName, }) if err != nil { return fmt.Errorf("failed service lookup: %s", err) } var deleteServices []*structs.ServiceNode for service := services.Next(); service != nil; service = services.Next() { svc := service.(*structs.ServiceNode) deleteServices = append(deleteServices, svc) if err := catalogUpdateServiceIndexes(tx, idx, svc.ServiceName, &svc.EnterpriseMeta, svc.PeerName); err != nil { return err } if err := catalogUpdateServiceKindIndexes(tx, idx, svc.ServiceKind, &svc.EnterpriseMeta, svc.PeerName); err != nil { return err } } // Do the delete in a separate loop so we don't trash the iterator. for _, svc := range deleteServices { if err := s.deleteServiceTxn(tx, idx, nodeName, svc.ServiceID, &svc.EnterpriseMeta, svc.PeerName); err != nil { return err } } // Delete all checks associated with the node. This will invalidate // sessions as necessary. checks, err := tx.Get(tableChecks, indexNode, Query{ Value: nodeName, EnterpriseMeta: *entMeta, PeerName: peerName, }) if err != nil { return fmt.Errorf("failed check lookup: %s", err) } var deleteChecks []*structs.HealthCheck for check := checks.Next(); check != nil; check = checks.Next() { deleteChecks = append(deleteChecks, check.(*structs.HealthCheck)) } // Do the delete in a separate loop so we don't trash the iterator. for _, chk := range deleteChecks { if err := s.deleteCheckTxn(tx, idx, nodeName, chk.CheckID, &chk.EnterpriseMeta, chk.PeerName); err != nil { return err } } if peerName == "" { // Delete any coordinates associated with this node. coords, err := tx.Get(tableCoordinates, indexNode, Query{ Value: nodeName, EnterpriseMeta: *entMeta, PeerName: structs.DefaultPeerKeyword, }) if err != nil { return fmt.Errorf("failed coordinate lookup: %s", err) } var coordsToDelete []*structs.Coordinate for coord := coords.Next(); coord != nil; coord = coords.Next() { coordsToDelete = append(coordsToDelete, coord.(*structs.Coordinate)) } for _, coord := range coordsToDelete { if err := deleteCoordinateTxn(tx, idx, coord); err != nil { return fmt.Errorf("failed deleting coordinate: %s", err) } } } // Delete the node and update the index. if err := tx.Delete(tableNodes, nodeRaw); err != nil { return fmt.Errorf("failed deleting node: %s", err) } node := nodeRaw.(*structs.Node) if err := catalogUpdateNodesIndexes(tx, idx, entMeta, node.PeerName); err != nil { return fmt.Errorf("failed updating index: %s", err) } // Clean up node entry from index table if err := tx.Delete(tableIndex, &IndexEntry{Key: nodeIndexName(nodeName, entMeta, node.PeerName)}); err != nil { return fmt.Errorf("failed deleting nodeIndex %q: %w", nodeIndexName(nodeName, entMeta, node.PeerName), err) } if err := catalogUpdateNodeExtinctionIndex(tx, idx, entMeta, node.PeerName); err != nil { return err } if peerName == "" { // Invalidate any sessions for this node. toDelete, err := allNodeSessionsTxn(tx, nodeName, entMeta.PartitionOrDefault()) if err != nil { return err } for _, session := range toDelete { if err := s.deleteSessionTxn(tx, idx, session.ID, &session.EnterpriseMeta); err != nil { return fmt.Errorf("failed to delete session '%s': %v", session.ID, err) } } } return nil } // EnsureService is called to upsert creation of a given NodeService. func (s *Store) EnsureService(idx uint64, node string, svc *structs.NodeService) error { tx := s.db.WriteTxn(idx) defer tx.Abort() // Call the service registration upsert if err := ensureServiceTxn(tx, idx, node, false, svc); err != nil { return err } return tx.Commit() } var errCASCompareFailed = errors.New("compare-and-set: comparison failed") // ensureServiceCASTxn updates a service only if the existing index matches the given index. // Returns an error if the write didn't happen and nil if write was successful. func ensureServiceCASTxn(tx WriteTxn, idx uint64, node string, svc *structs.NodeService) error { // Retrieve the existing service. existing, err := tx.First(tableServices, indexID, NodeServiceQuery{ EnterpriseMeta: svc.EnterpriseMeta, Node: node, Service: svc.ID, PeerName: svc.PeerName, }) if err != nil { return fmt.Errorf("failed service lookup: %s", err) } // Check if the we should do the set. A ModifyIndex of 0 means that // we are doing a set-if-not-exists. if svc.ModifyIndex == 0 && existing != nil { return errCASCompareFailed } if svc.ModifyIndex != 0 && existing == nil { return errCASCompareFailed } e, ok := existing.(*structs.ServiceNode) if ok && svc.ModifyIndex != 0 && svc.ModifyIndex != e.ModifyIndex { return errCASCompareFailed } return ensureServiceTxn(tx, idx, node, false, svc) } // ensureServiceTxn is used to upsert a service registration within an // existing memdb transaction. func ensureServiceTxn(tx WriteTxn, idx uint64, node string, preserveIndexes bool, svc *structs.NodeService) error { // Check for existing service existing, err := tx.First(tableServices, indexID, NodeServiceQuery{ EnterpriseMeta: svc.EnterpriseMeta, Node: node, Service: svc.ID, PeerName: svc.PeerName, }) if err != nil { return fmt.Errorf("failed service lookup: %s", err) } if err = structs.ValidateServiceMetadata(svc.Kind, svc.Meta, false); err != nil { return fmt.Errorf("Invalid Service Meta for node %s and serviceID %s: %v", node, svc.ID, err) } if svc.PeerName == "" { // Do not associate non-typical services with gateways or consul services if svc.Kind == structs.ServiceKindTypical && svc.Service != "consul" { // Check if this service is covered by a gateway's wildcard specifier, we force the service kind to a gateway-service here as that take precedence sn := structs.NewServiceName(svc.Service, &svc.EnterpriseMeta) if err = checkGatewayWildcardsAndUpdate(tx, idx, &sn, svc, structs.GatewayServiceKindService); err != nil { return fmt.Errorf("failed updating gateway mapping: %s", err) } if err = checkGatewayAndUpdate(tx, idx, &sn, structs.GatewayServiceKindService); err != nil { return fmt.Errorf("failed updating gateway mapping: %s", err) } } // Only upsert KindServiceName if service is local if err := upsertKindServiceName(tx, idx, svc.Kind, svc.CompoundServiceName()); err != nil { return fmt.Errorf("failed to persist service name: %v", err) } } // Update upstream/downstream mappings if it's a connect service if svc.Kind == structs.ServiceKindConnectProxy || svc.Connect.Native { if err = updateMeshTopology(tx, idx, node, svc, existing); err != nil { return fmt.Errorf("failed updating upstream/downstream association") } service := svc.Service if svc.Kind == structs.ServiceKindConnectProxy { service = svc.Proxy.DestinationServiceName } sn := structs.ServiceName{Name: service, EnterpriseMeta: svc.EnterpriseMeta} if err = checkGatewayWildcardsAndUpdate(tx, idx, &sn, svc, structs.GatewayServiceKindService); err != nil { return fmt.Errorf("failed updating gateway mapping: %s", err) } if svc.PeerName == "" && sn.Name != "" { if err := upsertKindServiceName(tx, idx, structs.ServiceKindConnectEnabled, sn); err != nil { return fmt.Errorf("failed to persist service name as connect-enabled: %v", err) } } // Update the virtual IP for the service supported, err := virtualIPsSupported(tx, nil) if err != nil { return err } if supported && sn.Name != "" { psn := structs.PeeredServiceName{Peer: svc.PeerName, ServiceName: sn} vip, err := assignServiceVirtualIP(tx, idx, psn) if err != nil { return fmt.Errorf("failed updating virtual IP: %s", err) } if svc.TaggedAddresses == nil { svc.TaggedAddresses = make(map[string]structs.ServiceAddress) } svc.TaggedAddresses[structs.TaggedAddressVirtualIP] = structs.ServiceAddress{Address: vip, Port: svc.Port} } } if svc.PeerName == "" { // If there's a terminating gateway config entry for this service, populate the tagged addresses // with virtual IP mappings. termGatewayVIPsSupported, err := terminatingGatewayVirtualIPsSupported(tx, nil) if err != nil { return err } if termGatewayVIPsSupported && svc.Kind == structs.ServiceKindTerminatingGateway { _, conf, err := configEntryTxn(tx, nil, structs.TerminatingGateway, svc.Service, &svc.EnterpriseMeta) if err != nil { return fmt.Errorf("failed to retrieve terminating gateway config: %s", err) } if conf != nil { termGatewayConf := conf.(*structs.TerminatingGatewayConfigEntry) addrs, err := getTermGatewayVirtualIPs(tx, idx, termGatewayConf.Services, &svc.EnterpriseMeta) if err != nil { return err } if svc.TaggedAddresses == nil { svc.TaggedAddresses = make(map[string]structs.ServiceAddress) } for key, addr := range addrs { svc.TaggedAddresses[key] = addr } } } } // Create the service node entry and populate the indexes. Note that // conversion doesn't populate any of the node-specific information. // That's always populated when we read from the state store. entry := svc.ToServiceNode(node) // Get the node n, err := tx.First(tableNodes, indexID, Query{ Value: node, EnterpriseMeta: svc.EnterpriseMeta, PeerName: svc.PeerName, }) if err != nil { return fmt.Errorf("failed node lookup: %s", err) } if n == nil { return ErrMissingNode } if existing != nil { serviceNode := existing.(*structs.ServiceNode) entry.CreateIndex = serviceNode.CreateIndex entry.ModifyIndex = serviceNode.ModifyIndex // We cannot return here because: we want to keep existing behavior (ex: failed node lookup -> ErrMissingNode) // It might be modified in future, but it requires changing many unit tests // Enforcing saving the entry also ensures that if we add default values in .ToServiceNode() // those values will be saved even if node is not really modified for a while. if entry.IsSameService(serviceNode) { return nil } } if !preserveIndexes { entry.ModifyIndex = idx if existing == nil { entry.CreateIndex = idx } } // Insert the service and update the index return catalogInsertService(tx, entry) } // assignServiceVirtualIP assigns a virtual IP to the target service and updates // the global virtual IP counter if necessary. func assignServiceVirtualIP(tx WriteTxn, idx uint64, psn structs.PeeredServiceName) (string, error) { serviceVIP, err := tx.First(tableServiceVirtualIPs, indexID, psn) if err != nil { return "", fmt.Errorf("failed service virtual IP lookup: %s", err) } // Service already has a virtual IP assigned, nothing to do. if serviceVIP != nil { sVIP := serviceVIP.(ServiceVirtualIP).IP result, err := addIPOffset(startingVirtualIP, sVIP) if err != nil { return "", err } return result.String(), nil } // Get the next available virtual IP, drawing from any freed from deleted services // first and then falling back to the global counter if none are available. latestVIP, err := tx.First(tableFreeVirtualIPs, indexCounterOnly, false) if err != nil { return "", fmt.Errorf("failed virtual IP index lookup: %s", err) } if latestVIP == nil { latestVIP, err = tx.First(tableFreeVirtualIPs, indexCounterOnly, true) if err != nil { return "", fmt.Errorf("failed virtual IP index lookup: %s", err) } } if latestVIP != nil { if err := tx.Delete(tableFreeVirtualIPs, latestVIP); err != nil { return "", fmt.Errorf("failed updating freed virtual IP table: %v", err) } } var latest FreeVirtualIP if latestVIP == nil { latest = FreeVirtualIP{ IP: net.IPv4zero, IsCounter: true, } } else { latest = latestVIP.(FreeVirtualIP) } // Store the next virtual IP from the counter if there aren't any freed IPs to draw from. // Then increment to store the next free virtual IP. newEntry := FreeVirtualIP{ IP: latest.IP, IsCounter: latest.IsCounter, } if latest.IsCounter { newEntry.IP = make(net.IP, len(latest.IP)) copy(newEntry.IP, latest.IP) for i := len(newEntry.IP) - 1; i >= 0; i-- { newEntry.IP[i]++ if newEntry.IP[i] != 0 { break } } // Out of virtual IPs, fail registration. if newEntry.IP.Equal(virtualIPMaxOffset) { return "", fmt.Errorf("cannot allocate any more unique service virtual IPs") } if err := tx.Insert(tableFreeVirtualIPs, newEntry); err != nil { return "", fmt.Errorf("failed updating freed virtual IP table: %v", err) } } assignedVIP := ServiceVirtualIP{ Service: psn, IP: newEntry.IP, RaftIndex: structs.RaftIndex{ ModifyIndex: idx, CreateIndex: idx, }, } if err := tx.Insert(tableServiceVirtualIPs, assignedVIP); err != nil { return "", fmt.Errorf("failed inserting service virtual IP entry: %s", err) } if err := updateVirtualIPMaxIndexes(tx, idx, psn.ServiceName.PartitionOrDefault(), psn.Peer); err != nil { return "", err } result, err := addIPOffset(startingVirtualIP, assignedVIP.IP) if err != nil { return "", err } return result.String(), nil } // AssignManualServiceVIPs attempts to associate a list of manual virtual IP addresses with a given service name. // Any IP addresses given will be removed from other services in the same partition. This is done to ensure // that a manual VIP can only exist once for a given partition. // This function returns: // - a bool indicating whether the given service exists. // - a list of service names that had ip addresses removed from them. // - an error indicating success or failure of the call. func (s *Store) AssignManualServiceVIPs(idx uint64, psn structs.PeeredServiceName, ips []string) (bool, []structs.PeeredServiceName, error) { tx := s.db.WriteTxn(idx) defer tx.Abort() // First remove the given IPs from any existing services, to avoid duplicate assignments. assignedIPs := map[string]struct{}{} for _, ip := range ips { assignedIPs[ip] = struct{}{} } modifiedEntries := make(map[structs.PeeredServiceName]struct{}) for ip := range assignedIPs { entry, err := tx.First(tableServiceVirtualIPs, indexManualVIPs, psn.ServiceName.PartitionOrDefault(), ip) if err != nil { return false, nil, fmt.Errorf("failed service virtual IP lookup: %s", err) } if entry == nil { continue } newEntry := entry.(ServiceVirtualIP) if newEntry.Service.ServiceName.Matches(psn.ServiceName) { continue } // Rebuild this entry's list of manual IPs, removing any that are present // in new list we're assigning. var filteredIPs []string for _, existingIP := range newEntry.ManualIPs { if _, ok := assignedIPs[existingIP]; !ok { filteredIPs = append(filteredIPs, existingIP) } } newEntry.ManualIPs = filteredIPs newEntry.ModifyIndex = idx if err := tx.Insert(tableServiceVirtualIPs, newEntry); err != nil { return false, nil, fmt.Errorf("failed inserting service virtual IP entry: %s", err) } modifiedEntries[newEntry.Service] = struct{}{} } entry, err := tx.First(tableServiceVirtualIPs, indexID, psn) if err != nil { return false, nil, fmt.Errorf("failed service virtual IP lookup: %s", err) } if entry == nil { return false, nil, nil } newEntry := entry.(ServiceVirtualIP) newEntry.ManualIPs = ips newEntry.ModifyIndex = idx if err := tx.Insert(tableServiceVirtualIPs, newEntry); err != nil { return false, nil, fmt.Errorf("failed inserting service virtual IP entry: %s", err) } if err := updateVirtualIPMaxIndexes(tx, idx, psn.ServiceName.PartitionOrDefault(), psn.Peer); err != nil { return false, nil, err } if err = tx.Commit(); err != nil { return false, nil, err } return true, maps.SliceOfKeys(modifiedEntries), nil } func updateVirtualIPMaxIndexes(txn WriteTxn, idx uint64, partition, peerName string) error { // update per-partition max index if err := indexUpdateMaxTxn(txn, idx, partitionedIndexEntryName(tableServiceVirtualIPs, partition)); err != nil { return fmt.Errorf("failed while updating partitioned index: %w", err) } if peerName != "" { // track a separate max index for imported services if err := indexUpdateMaxTxn(txn, idx, partitionedIndexEntryName(tableServiceVirtualIPs+".imported", partition)); err != nil { return fmt.Errorf("failed while updating partitioned index for imported services: %w", err) } } return nil } func addIPOffset(a, b net.IP) (net.IP, error) { a4 := a.To4() b4 := b.To4() if a4 == nil || b4 == nil { return nil, errors.New("ip is not ipv4") } var raw uint64 for i := 0; i < 4; i++ { raw = raw<<8 + uint64(a4[i]) + uint64(b4[i]) } return net.IPv4(byte(raw>>24), byte(raw>>16), byte(raw>>8), byte(raw)), nil } func virtualIPsSupported(tx ReadTxn, ws memdb.WatchSet) (bool, error) { _, entry, err := systemMetadataGetTxn(tx, ws, structs.SystemMetadataVirtualIPsEnabled) if err != nil { return false, fmt.Errorf("failed system metadata lookup: %s", err) } if entry == nil { return false, nil } return entry.Value != "", nil } func terminatingGatewayVirtualIPsSupported(tx ReadTxn, ws memdb.WatchSet) (bool, error) { _, entry, err := systemMetadataGetTxn(tx, ws, structs.SystemMetadataTermGatewayVirtualIPsEnabled) if err != nil { return false, fmt.Errorf("failed system metadata lookup: %s", err) } if entry == nil { return false, nil } return entry.Value != "", nil } // Services returns all services along with a list of associated tags. func (s *Store) Services(ws memdb.WatchSet, entMeta *acl.EnterpriseMeta, peerName string) (uint64, []*structs.ServiceNode, error) { tx := s.db.Txn(false) defer tx.Abort() // Get the table index. idx := catalogServicesMaxIndex(tx, entMeta, peerName) // List all the services. services, err := catalogServiceListNoWildcard(tx, entMeta, peerName) if err != nil { return 0, nil, fmt.Errorf("failed querying services: %s", err) } ws.Add(services.WatchCh()) var result []*structs.ServiceNode for service := services.Next(); service != nil; service = services.Next() { result = append(result, service.(*structs.ServiceNode)) } return idx, result, nil } func (s *Store) ServiceList(ws memdb.WatchSet, entMeta *acl.EnterpriseMeta, peerName string) (uint64, structs.ServiceList, error) { tx := s.db.Txn(false) defer tx.Abort() return serviceListTxn(tx, ws, entMeta, peerName) } func serviceListTxn(tx ReadTxn, ws memdb.WatchSet, entMeta *acl.EnterpriseMeta, peerName string) (uint64, structs.ServiceList, error) { if entMeta == nil { entMeta = structs.NodeEnterpriseMetaInDefaultPartition() } idx := catalogServicesMaxIndex(tx, entMeta, peerName) services, err := tx.Get(tableServices, indexID+"_prefix", Query{ EnterpriseMeta: *entMeta, PeerName: peerName, }) if err != nil { return 0, nil, fmt.Errorf("failed querying services: %s", err) } ws.Add(services.WatchCh()) unique := make(map[structs.ServiceName]struct{}) for service := services.Next(); service != nil; service = services.Next() { svc := service.(*structs.ServiceNode) unique[svc.CompoundServiceName().ServiceName] = struct{}{} } results := make(structs.ServiceList, 0, len(unique)) for sn := range unique { results = append(results, structs.ServiceName{Name: sn.Name, EnterpriseMeta: sn.EnterpriseMeta}) } return idx, results, nil } // ServicesByNodeMeta returns all services, filtered by the given node metadata. func (s *Store) ServicesByNodeMeta(ws memdb.WatchSet, filters map[string]string, entMeta *acl.EnterpriseMeta, peerName string) (uint64, []*structs.ServiceNode, error) { tx := s.db.Txn(false) defer tx.Abort() // TODO: accept non-pointer value if entMeta == nil { entMeta = structs.NodeEnterpriseMetaInDefaultPartition() } // Get the table index. idx := catalogServicesMaxIndex(tx, entMeta, peerName) if nodeIdx := catalogNodesMaxIndex(tx, entMeta, peerName); nodeIdx > idx { idx = nodeIdx } if len(filters) == 0 { return idx, nil, nil // ServicesByNodeMeta is never called with an empty map, but just in case make it return no results. } // Retrieve all of the nodes. We'll do a lookup of just ONE KV pair, which // over-matches if multiple pairs are requested, but then in the loop below // we'll finish filtering. var firstKey, firstValue string for firstKey, firstValue = range filters { break } nodes, err := tx.Get(tableNodes, indexMeta, KeyValueQuery{ Key: firstKey, Value: firstValue, EnterpriseMeta: *entMeta, PeerName: peerName, }) if err != nil { return 0, nil, fmt.Errorf("failed nodes lookup: %s", err) } ws.Add(nodes.WatchCh()) // We don't want to track an unlimited number of services, so we pull a // top-level watch to use as a fallback. allServices, err := catalogServiceListNoWildcard(tx, entMeta, peerName) if err != nil { return 0, nil, fmt.Errorf("failed services lookup: %s", err) } allServicesCh := allServices.WatchCh() var result structs.ServiceNodes for node := nodes.Next(); node != nil; node = nodes.Next() { n := node.(*structs.Node) if len(filters) > 1 && !structs.SatisfiesMetaFilters(n.Meta, filters) { continue } // List all the services on the node services, err := catalogServiceListByNode(tx, n.Node, entMeta, n.PeerName, false) if err != nil { return 0, nil, fmt.Errorf("failed querying services: %s", err) } ws.AddWithLimit(watchLimit, services.WatchCh(), allServicesCh) for service := services.Next(); service != nil; service = services.Next() { result = append(result, service.(*structs.ServiceNode)) } } return idx, result, nil } // maxIndexForService return the maximum Raft Index for a service // If the index is not set for the service, it will return the missing // service index. // The service_last_extinction is set to the last raft index when a service // was unregistered (or 0 if no services were ever unregistered). This // allows blocking queries to // - return when the last instance of a service is removed // - block until an instance for this service is available, or another // service is unregistered. func maxIndexForService(tx ReadTxn, serviceName string, serviceExists, checks bool, entMeta *acl.EnterpriseMeta, peerName string) uint64 { idx, _ := maxIndexAndWatchChForService(tx, serviceName, serviceExists, checks, entMeta, peerName) return idx } // maxIndexAndWatchChForService return the maximum Raft Index for a service. If // the index is not set for the service, it will return the missing service // index. The service_last_extinction is set to the last raft index when a // service was unregistered (or 0 if no services were ever unregistered). This // allows blocking queries to // - return when the last instance of a service is removed // - block until an instance for this service is available, or another // service is unregistered. // // It also _may_ return a watch chan to add to a WatchSet. It will only return // one if the service exists, and has a service index. If it doesn't then nil is // returned for the chan. This allows for blocking watchers to _only_ watch this // one chan in the common case, falling back to watching all touched MemDB // indexes in more complicated cases. func maxIndexAndWatchChForService(tx ReadTxn, serviceName string, serviceExists, checks bool, entMeta *acl.EnterpriseMeta, peerName string) (uint64, <-chan struct{}) { if !serviceExists { res, err := catalogServiceLastExtinctionIndex(tx, entMeta, peerName) if missingIdx, ok := res.(*IndexEntry); ok && err == nil { // Note safe to only watch the extinction index as it's not updated when new instances come along so return nil watchCh return missingIdx.Value, nil } } ch, res, err := catalogServiceMaxIndex(tx, serviceName, entMeta, peerName) if idx, ok := res.(*IndexEntry); ok && err == nil { return idx.Value, ch } return catalogMaxIndex(tx, entMeta, peerName, checks), nil } // Wrapper for maxIndexAndWatchChForService that operates on a list of ServiceNodes func maxIndexAndWatchChsForServiceNodes(tx ReadTxn, nodes structs.ServiceNodes, watchChecks bool) (uint64, []<-chan struct{}) { var watchChans []<-chan struct{} var maxIdx uint64 seen := make(map[structs.ServiceName]bool) for i := 0; i < len(nodes); i++ { sn := structs.NewServiceName(nodes[i].ServiceName, &nodes[i].EnterpriseMeta) if ok := seen[sn]; !ok { idx, svcCh := maxIndexAndWatchChForService(tx, sn.Name, true, watchChecks, &sn.EnterpriseMeta, nodes[i].PeerName) if idx > maxIdx { maxIdx = idx } if svcCh != nil { watchChans = append(watchChans, svcCh) } seen[sn] = true } } return maxIdx, watchChans } // ConnectServiceNodes returns the nodes associated with a Connect // compatible destination for the given service name. This will include // both proxies and native integrations. func (s *Store) ConnectServiceNodes(ws memdb.WatchSet, serviceName string, entMeta *acl.EnterpriseMeta, peerName string) (uint64, structs.ServiceNodes, error) { tx := s.db.ReadTxn() defer tx.Abort() // TODO: accept non-pointer value if entMeta == nil { entMeta = structs.DefaultEnterpriseMetaInDefaultPartition() } q := Query{ Value: serviceName, PeerName: peerName, EnterpriseMeta: *entMeta, } return serviceNodesTxn(tx, ws, indexConnect, q) } // ServiceNodes returns the nodes associated with a given service name. func (s *Store) ServiceNodes(ws memdb.WatchSet, serviceName string, entMeta *acl.EnterpriseMeta, peerName string) (uint64, structs.ServiceNodes, error) { tx := s.db.ReadTxn() defer tx.Abort() // TODO: accept non-pointer value if entMeta == nil { entMeta = structs.DefaultEnterpriseMetaInDefaultPartition() } q := Query{ Value: serviceName, PeerName: peerName, EnterpriseMeta: *entMeta, } return serviceNodesTxn(tx, ws, indexService, q) } func serviceNodesTxn(tx ReadTxn, ws memdb.WatchSet, index string, q Query) (uint64, structs.ServiceNodes, error) { connect := index == indexConnect serviceName := q.Value services, err := tx.Get(tableServices, index, q) if err != nil { return 0, nil, fmt.Errorf("failed service lookup: %s", err) } ws.Add(services.WatchCh()) var results structs.ServiceNodes for service := services.Next(); service != nil; service = services.Next() { results = append(results, service.(*structs.ServiceNode)) } // If we are querying for Connect nodes, the associated proxy might be a gateway. // Gateways are tracked in a separate table, and we append them to the result set. // We append rather than replace since it allows users to migrate a service // to the mesh with a mix of sidecars and gateways until all its instances have a sidecar. var idx uint64 if connect && q.PeerName == "" { // Look up gateway nodes associated with the service gwIdx, nodes, err := serviceGatewayNodes(tx, ws, serviceName, structs.ServiceKindTerminatingGateway, &q.EnterpriseMeta, structs.DefaultPeerKeyword) if err != nil { return 0, nil, fmt.Errorf("failed gateway nodes lookup: %v", err) } if idx < gwIdx { idx = gwIdx } // Watch for index changes to the gateway nodes svcIdx, chans := maxIndexAndWatchChsForServiceNodes(tx, nodes, false) if svcIdx > idx { idx = svcIdx } for _, ch := range chans { ws.Add(ch) } for i := 0; i < len(nodes); i++ { results = append(results, nodes[i]) } } // Fill in the node details. results, err = parseServiceNodes(tx, ws, results, &q.EnterpriseMeta, q.PeerName) if err != nil { return 0, nil, fmt.Errorf("failed parsing service nodes: %s", err) } // Get the table index. // TODO (gateways) (freddy) Why do we always consider the main service index here? // This doesn't seem to make sense for Connect when there's more than 1 result svcIdx := maxIndexForService(tx, serviceName, len(results) > 0, false, &q.EnterpriseMeta, q.PeerName) if idx < svcIdx { idx = svcIdx } return idx, results, nil } // ServiceTagNodes returns the nodes associated with a given service, filtering // out services that don't contain the given tags. func (s *Store) ServiceTagNodes(ws memdb.WatchSet, service string, tags []string, entMeta *acl.EnterpriseMeta, peerName string) (uint64, structs.ServiceNodes, error) { tx := s.db.Txn(false) defer tx.Abort() // TODO: accept non-pointer value if entMeta == nil { entMeta = structs.DefaultEnterpriseMetaInDefaultPartition() } services, err := tx.Get(tableServices, indexService, Query{ Value: service, EnterpriseMeta: *entMeta, PeerName: peerName, }) if err != nil { return 0, nil, fmt.Errorf("failed service lookup: %s", err) } ws.Add(services.WatchCh()) // Gather all the services and apply the tag filter. serviceExists := false var results structs.ServiceNodes for service := services.Next(); service != nil; service = services.Next() { svc := service.(*structs.ServiceNode) serviceExists = true if !serviceTagsFilter(svc, tags) { results = append(results, svc) } } // Fill in the node details. results, err = parseServiceNodes(tx, ws, results, entMeta, peerName) if err != nil { return 0, nil, fmt.Errorf("failed parsing service nodes: %s", err) } // Get the table index. idx := maxIndexForService(tx, service, serviceExists, false, entMeta, peerName) return idx, results, nil } // serviceTagFilter returns true (should filter) if the given service node // doesn't contain the given tag. func serviceTagFilter(sn *structs.ServiceNode, tag string) bool { tag = strings.ToLower(tag) // Look for the lower cased version of the tag. for _, t := range sn.ServiceTags { if strings.ToLower(t) == tag { return false } } // If we didn't hit the tag above then we should filter. return true } // serviceTagsFilter returns true (should filter) if the given service node // doesn't contain the given set of tags. func serviceTagsFilter(sn *structs.ServiceNode, tags []string) bool { for _, tag := range tags { if serviceTagFilter(sn, tag) { // If any one of the expected tags was not found, filter the service return true } } // If all tags were found, don't filter the service return false } // ServiceAddressNodes returns the nodes associated with a given service, filtering // out services that don't match the given serviceAddress func (s *Store) ServiceAddressNodes(ws memdb.WatchSet, address string, entMeta *acl.EnterpriseMeta, peerName string) (uint64, structs.ServiceNodes, error) { tx := s.db.Txn(false) defer tx.Abort() // List all the services. q := Query{ EnterpriseMeta: *entMeta, PeerName: peerName, } services, err := tx.Get(tableServices, indexID+"_prefix", q) if err != nil { return 0, nil, fmt.Errorf("failed service lookup: %s", err) } ws.Add(services.WatchCh()) // Gather all the services and apply the tag filter. var results structs.ServiceNodes for service := services.Next(); service != nil; service = services.Next() { svc := service.(*structs.ServiceNode) if svc.ServiceAddress == address { results = append(results, svc) } else { for _, addr := range svc.ServiceTaggedAddresses { if addr.Address == address { results = append(results, svc) break } } } } // Fill in the node details. results, err = parseServiceNodes(tx, ws, results, entMeta, peerName) if err != nil { return 0, nil, fmt.Errorf("failed parsing service nodes: %s", err) } return 0, results, nil } // parseServiceNodes iterates over a services query and fills in the node details, // returning a ServiceNodes slice. func parseServiceNodes(tx ReadTxn, ws memdb.WatchSet, services structs.ServiceNodes, entMeta *acl.EnterpriseMeta, peerName string) (structs.ServiceNodes, error) { // We don't want to track an unlimited number of nodes, so we pull a // top-level watch to use as a fallback. q := Query{ EnterpriseMeta: *entMeta, PeerName: peerName, } allNodes, err := tx.Get(tableNodes, indexID+"_prefix", q) if err != nil { return nil, fmt.Errorf("failed nodes lookup: %s", err) } allNodesCh := allNodes.WatchCh() // Fill in the node data for each service instance. var results structs.ServiceNodes for _, sn := range services { // Note that we have to clone here because we don't want to // modify the node-related fields on the object in the database, // which is what we are referencing. s := sn.PartialClone() // Grab the corresponding node record. watchCh, n, err := tx.FirstWatch(tableNodes, indexID, Query{ Value: sn.Node, EnterpriseMeta: sn.EnterpriseMeta, PeerName: sn.PeerName, }) if err != nil { return nil, fmt.Errorf("failed node lookup: %s", err) } ws.AddWithLimit(watchLimit, watchCh, allNodesCh) // Populate the node-related fields. The tagged addresses may be // used by agents to perform address translation if they are // configured to do that. node := n.(*structs.Node) s.ID = node.ID s.Address = node.Address s.Datacenter = node.Datacenter s.TaggedAddresses = node.TaggedAddresses s.EnterpriseMeta.Merge(node.GetEnterpriseMeta()) s.NodeMeta = node.Meta results = append(results, s) } return results, nil } // NodeService is used to retrieve a specific service associated with the given // node. func (s *Store) NodeService(ws memdb.WatchSet, nodeName string, serviceID string, entMeta *acl.EnterpriseMeta, peerName string) (uint64, *structs.NodeService, error) { tx := s.db.Txn(false) defer tx.Abort() // Get the table index. idx := catalogServicesMaxIndex(tx, entMeta, peerName) // Query the service service, err := getNodeServiceTxn(tx, ws, nodeName, serviceID, entMeta, peerName) if err != nil { return 0, nil, fmt.Errorf("failed querying service for node %q: %s", nodeName, err) } return idx, service, nil } func getNodeServiceTxn(tx ReadTxn, ws memdb.WatchSet, nodeName, serviceID string, entMeta *acl.EnterpriseMeta, peerName string) (*structs.NodeService, error) { sn, err := getServiceNodeTxn(tx, ws, nodeName, serviceID, entMeta, peerName) if err != nil { return nil, err } if sn != nil { return sn.ToNodeService(), nil } return nil, nil } func getServiceNodeTxn(tx ReadTxn, ws memdb.WatchSet, nodeName, serviceID string, entMeta *acl.EnterpriseMeta, peerName string) (*structs.ServiceNode, error) { // TODO: pass non-pointer type for ent meta if entMeta == nil { entMeta = structs.DefaultEnterpriseMetaInDefaultPartition() } // Query the service watch, service, err := tx.FirstWatch(tableServices, indexID, NodeServiceQuery{ EnterpriseMeta: *entMeta, Node: nodeName, Service: serviceID, PeerName: peerName, }) if err != nil { return nil, fmt.Errorf("failed querying service for node %q: %s", nodeName, err) } ws.Add(watch) if service != nil { return service.(*structs.ServiceNode), nil } return nil, nil } // ServiceNode is used to retrieve a specific service by service ID and node ID or name. func (s *Store) ServiceNode(nodeID, nodeName, serviceID string, entMeta *acl.EnterpriseMeta, peerName string) (uint64, *structs.ServiceNode, error) { var ( node *structs.Node err error ) if nodeID != "" { _, node, err = s.GetNodeID(types.NodeID(nodeID), entMeta, peerName) if err != nil { return 0, nil, fmt.Errorf("Failure looking up node by ID %s: %w", nodeID, err) } } else if nodeName != "" { _, node, err = s.GetNode(nodeName, entMeta, peerName) if err != nil { return 0, nil, fmt.Errorf("Failure looking up node by name %s: %w", nodeName, err) } } else { return 0, nil, fmt.Errorf("Node ID or name required to lookup the service") } if node == nil { return 0, nil, ErrNodeNotFound } tx := s.db.Txn(false) defer tx.Abort() // Get the table index. idx := catalogServicesMaxIndex(tx, entMeta, peerName) // Query the service service, err := getServiceNodeTxn(tx, nil, node.Node, serviceID, entMeta, peerName) if err != nil { return 0, nil, fmt.Errorf("failed querying service for node %q: %w", node.Node, err) } if service != nil { service.ID = node.ID } return idx, service, nil } func (s *Store) nodeServices(ws memdb.WatchSet, nodeNameOrID string, entMeta *acl.EnterpriseMeta, peerName string, allowWildcard bool) (bool, uint64, *structs.Node, memdb.ResultIterator, error) { tx := s.db.Txn(false) defer tx.Abort() // TODO: accept non-pointer value for entMeta if entMeta == nil { entMeta = structs.DefaultEnterpriseMetaInDefaultPartition() } // Query the node by node name watchCh, n, err := tx.FirstWatch(tableNodes, indexID, Query{ Value: nodeNameOrID, EnterpriseMeta: *entMeta, PeerName: peerName, }) if err != nil { return true, 0, nil, nil, fmt.Errorf("node lookup failed: %s", err) } if n != nil { ws.Add(watchCh) } else { if len(nodeNameOrID) < minUUIDLookupLen { ws.Add(watchCh) return true, 0, nil, nil, nil } // Attempt to lookup the node by its node ID iter, err := tx.Get(tableNodes, indexUUID+"_prefix", Query{ Value: resizeNodeLookupKey(nodeNameOrID), EnterpriseMeta: *entMeta, PeerName: peerName, }) if err != nil { ws.Add(watchCh) idx := catalogNodeLastExtinctionIndex(tx, entMeta, peerName) return true, idx, nil, nil, nil } n = iter.Next() if n == nil { // No nodes matched, even with the Node ID: add a watch on the node name. ws.Add(watchCh) idx := catalogNodeLastExtinctionIndex(tx, entMeta, peerName) return true, idx, nil, nil, nil } idWatchCh := iter.WatchCh() if iter.Next() != nil { // More than one match present: Watch on the node name channel and return // an empty result (node lookups can not be ambiguous). ws.Add(watchCh) return true, 0, nil, nil, nil } ws.Add(idWatchCh) } node := n.(*structs.Node) nodeName := node.Node // Read all of the services services, err := catalogServiceListByNode(tx, nodeName, entMeta, peerName, allowWildcard) if err != nil { return true, 0, nil, nil, fmt.Errorf("failed querying services for node %q: %s", nodeName, err) } ws.Add(services.WatchCh()) // Get the table index. idx := catalogNodeMaxIndex(tx, nodeName, entMeta, peerName) return false, idx, node, services, nil } // NodeServices is used to query service registrations by node name or UUID. func (s *Store) NodeServices(ws memdb.WatchSet, nodeNameOrID string, entMeta *acl.EnterpriseMeta, peerName string) (uint64, *structs.NodeServices, error) { done, idx, node, services, err := s.nodeServices(ws, nodeNameOrID, entMeta, peerName, false) if done || err != nil { return idx, nil, err } // Initialize the node services struct ns := &structs.NodeServices{ Node: node, Services: make(map[string]*structs.NodeService), } if services != nil { // Add all of the services to the map. for service := services.Next(); service != nil; service = services.Next() { svc := service.(*structs.ServiceNode).ToNodeService() ns.Services[svc.ID] = svc } } return idx, ns, nil } // NodeServices is used to query service registrations by node name or UUID. func (s *Store) NodeServiceList(ws memdb.WatchSet, nodeNameOrID string, entMeta *acl.EnterpriseMeta, peerName string) (uint64, *structs.NodeServiceList, error) { done, idx, node, services, err := s.nodeServices(ws, nodeNameOrID, entMeta, peerName, true) if done || err != nil { return idx, nil, err } if idx == 0 { return 0, nil, nil } // Initialize the node services struct ns := &structs.NodeServiceList{ Node: node, } if services != nil { // Add all of the services to the map. for service := services.Next(); service != nil; service = services.Next() { svc := service.(*structs.ServiceNode).ToNodeService() ns.Services = append(ns.Services, svc) } } return idx, ns, nil } // DeleteService is used to delete a given service associated with a node. func (s *Store) DeleteService(idx uint64, nodeName, serviceID string, entMeta *acl.EnterpriseMeta, peerName string) error { tx := s.db.WriteTxn(idx) defer tx.Abort() // Call the service deletion if err := s.deleteServiceTxn(tx, idx, nodeName, serviceID, entMeta, peerName); err != nil { return err } return tx.Commit() } // deleteServiceCASTxn is used to try doing a service delete operation with a given // raft index. If the CAS index specified is not equal to the last observed index for // the given service, then the call is a noop, otherwise a normal delete is invoked. func (s *Store) deleteServiceCASTxn(tx WriteTxn, idx, cidx uint64, nodeName, serviceID string, entMeta *acl.EnterpriseMeta, peerName string) (bool, error) { // Look up the service. service, err := getNodeServiceTxn(tx, nil, nodeName, serviceID, entMeta, peerName) if err != nil { return false, fmt.Errorf("service lookup failed: %s", err) } if service == nil { return false, nil } // If the existing index does not match the provided CAS // index arg, then we shouldn't update anything and can safely // return early here. if service.ModifyIndex != cidx { return false, nil } // Call the actual deletion if the above passed. if err := s.deleteServiceTxn(tx, idx, nodeName, serviceID, entMeta, peerName); err != nil { return false, err } return true, nil } // deleteServiceTxn is the inner method called to remove a service // registration within an existing transaction. func (s *Store) deleteServiceTxn(tx WriteTxn, idx uint64, nodeName, serviceID string, entMeta *acl.EnterpriseMeta, peerName string) error { // TODO: pass non-pointer type for ent meta if entMeta == nil { entMeta = structs.DefaultEnterpriseMetaInDefaultPartition() } service, err := tx.First(tableServices, indexID, NodeServiceQuery{ EnterpriseMeta: *entMeta, Node: nodeName, Service: serviceID, PeerName: peerName, }) if err != nil { return fmt.Errorf("failed service lookup: %s", err) } if service == nil { return nil } // Delete any checks associated with the service. This will invalidate // sessions as necessary. nsq := NodeServiceQuery{ Node: nodeName, Service: serviceID, EnterpriseMeta: *entMeta, PeerName: peerName, } checks, err := tx.Get(tableChecks, indexNodeService, nsq) if err != nil { return fmt.Errorf("failed service check lookup: %s", err) } var deleteChecks []*structs.HealthCheck for check := checks.Next(); check != nil; check = checks.Next() { deleteChecks = append(deleteChecks, check.(*structs.HealthCheck)) } // Do the delete in a separate loop so we don't trash the iterator. for _, check := range deleteChecks { if err := s.deleteCheckTxn(tx, idx, nodeName, check.CheckID, &check.EnterpriseMeta, check.PeerName); err != nil { return err } } // Update the index. if err := catalogUpdateCheckIndexes(tx, idx, entMeta, peerName); err != nil { return err } // Delete the service and update the index if err := tx.Delete(tableServices, service); err != nil { return fmt.Errorf("failed deleting service: %s", err) } svc := service.(*structs.ServiceNode) if err := catalogUpdateServicesIndexes(tx, idx, entMeta, svc.PeerName); err != nil { return fmt.Errorf("failed updating services indexes: %w", err) } if err := catalogUpdateServiceKindIndexes(tx, idx, svc.ServiceKind, &svc.EnterpriseMeta, svc.PeerName); err != nil { return fmt.Errorf("failed updating service-kind indexes: %w", err) } // Update the node indexes as the service information is included in node catalog queries. if err := catalogUpdateNodesIndexes(tx, idx, entMeta, svc.PeerName); err != nil { return fmt.Errorf("failed updating nodes indexes: %w", err) } if err := catalogUpdateNodeIndexes(tx, idx, nodeName, entMeta, svc.PeerName); err != nil { return fmt.Errorf("failed updating node indexes: %w", err) } psn := svc.CompoundServiceName() if err := cleanupMeshTopology(tx, idx, svc); err != nil { return fmt.Errorf("failed to clean up mesh-topology associations for %q: %v", psn.String(), err) } q := Query{ Value: svc.ServiceName, EnterpriseMeta: *entMeta, PeerName: svc.PeerName, } if remainingService, err := tx.First(tableServices, indexService, q); err == nil { if remainingService != nil { // We have at least one remaining service, update the index if err := catalogUpdateServiceIndexes(tx, idx, svc.ServiceName, entMeta, svc.PeerName); err != nil { return err } } else { // There are no more service instances, cleanup the service. index _, serviceIndex, err := catalogServiceMaxIndex(tx, svc.ServiceName, entMeta, svc.PeerName) if err == nil && serviceIndex != nil { // we found service. index, garbage collect it if err := tx.Delete(tableIndex, serviceIndex); err != nil { return fmt.Errorf("[FAILED] deleting serviceIndex %s: %s", svc.ServiceName, err) } } if err := catalogUpdateServiceExtinctionIndex(tx, idx, entMeta, svc.PeerName); err != nil { return err } if err := freeServiceVirtualIP(tx, idx, psn, nil); err != nil { return fmt.Errorf("failed to clean up virtual IP for %q: %v", psn.String(), err) } if svc.PeerName == "" { if err := cleanupKindServiceName(tx, idx, psn.ServiceName, svc.ServiceKind); err != nil { return fmt.Errorf("failed to persist service name: %v", err) } } } } else { return fmt.Errorf("Could not find any service %s: %s", svc.ServiceName, err) } // Cleanup ConnectEnabled for this service if none exist. if svc.PeerName == "" && (svc.ServiceKind == structs.ServiceKindConnectProxy || svc.ServiceConnect.Native) { service := svc.ServiceName if svc.ServiceKind == structs.ServiceKindConnectProxy { service = svc.ServiceProxy.DestinationServiceName } sn := structs.ServiceName{Name: service, EnterpriseMeta: svc.EnterpriseMeta} connectEnabled, err := serviceHasConnectEnabledInstances(tx, sn.Name, &sn.EnterpriseMeta) if err != nil { return fmt.Errorf("failed to search for connect instances for service %q: %w", sn.Name, err) } if !connectEnabled { if err := cleanupKindServiceName(tx, idx, sn, structs.ServiceKindConnectEnabled); err != nil { return fmt.Errorf("failed to cleanup connect-enabled service name: %v", err) } } } if svc.PeerName == "" { sn := structs.ServiceName{Name: svc.ServiceName, EnterpriseMeta: svc.EnterpriseMeta} if err := cleanupGatewayWildcards(tx, idx, sn, false); err != nil { return fmt.Errorf("failed to clean up gateway-service associations for %q: %v", psn.String(), err) } } return nil } // freeServiceVirtualIP is used to free a virtual IP for a service after the last instance // is removed. func freeServiceVirtualIP( tx WriteTxn, idx uint64, psn structs.PeeredServiceName, excludeGateway *structs.ServiceName, ) error { supported, err := virtualIPsSupported(tx, nil) if err != nil { return err } if !supported { return nil } // Don't deregister the virtual IP if at least one instance of this service still exists. q := Query{ Value: psn.ServiceName.Name, EnterpriseMeta: psn.ServiceName.EnterpriseMeta, PeerName: psn.Peer, } if remainingService, err := tx.First(tableServices, indexService, q); err == nil { if remainingService != nil { return nil } } else { return fmt.Errorf("failed service lookup for %q: %s", psn.ServiceName.Name, err) } // Don't deregister the virtual IP if at least one resolver/router/splitter config entry still // references this service. configEntryVIPKinds := []string{ structs.ServiceResolver, structs.ServiceRouter, structs.ServiceSplitter, structs.ServiceDefaults, structs.ServiceIntentions, } for _, kind := range configEntryVIPKinds { _, entry, err := configEntryTxn(tx, nil, kind, psn.ServiceName.Name, &psn.ServiceName.EnterpriseMeta) if err != nil { return fmt.Errorf("failed config entry lookup for %s/%s: %s", kind, psn.ServiceName.Name, err) } if entry != nil { return nil } } // Don't deregister the virtual IP if at least one terminating gateway still references this service. termGatewaySupported, err := terminatingGatewayVirtualIPsSupported(tx, nil) if err != nil { return err } if termGatewaySupported { svcGateways, err := tx.Get(tableGatewayServices, indexService, psn.ServiceName) if err != nil { return fmt.Errorf("failed gateway lookup for %q: %s", psn.ServiceName.Name, err) } for service := svcGateways.Next(); service != nil; service = svcGateways.Next() { if svc, ok := service.(*structs.GatewayService); ok && svc != nil { ignoreGateway := excludeGateway == nil || !svc.Gateway.Matches(*excludeGateway) if ignoreGateway && svc.GatewayKind == structs.ServiceKindTerminatingGateway { return nil } } } } serviceVIP, err := tx.First(tableServiceVirtualIPs, indexID, psn) if err != nil { return fmt.Errorf("failed service virtual IP lookup: %s", err) } // Service has no virtual IP assigned, nothing to do. if serviceVIP == nil { return nil } // Delete the service virtual IP and add it to the freed IPs list. if err := tx.Delete(tableServiceVirtualIPs, serviceVIP); err != nil { return fmt.Errorf("failed updating freed virtual IP table: %v", err) } newEntry := FreeVirtualIP{IP: serviceVIP.(ServiceVirtualIP).IP} if err := tx.Insert(tableFreeVirtualIPs, newEntry); err != nil { return fmt.Errorf("failed updating freed virtual IP table: %v", err) } if err := updateVirtualIPMaxIndexes(tx, idx, psn.ServiceName.PartitionOrDefault(), psn.Peer); err != nil { return err } return nil } // EnsureCheck is used to store a check registration in the db. func (s *Store) EnsureCheck(idx uint64, hc *structs.HealthCheck) error { tx := s.db.WriteTxn(idx) defer tx.Abort() // Call the check registration if err := s.ensureCheckTxn(tx, idx, false, hc); err != nil { return err } return tx.Commit() } // updateAllServiceIndexesOfNode updates the Raft index of all the services associated with this node func updateAllServiceIndexesOfNode(tx WriteTxn, idx uint64, nodeID string, entMeta *acl.EnterpriseMeta, peerName string) error { if peerName == "" { peerName = structs.LocalPeerKeyword } services, err := tx.Get(tableServices, indexNode, Query{ Value: nodeID, EnterpriseMeta: *entMeta.WithWildcardNamespace(), PeerName: peerName, }) if err != nil { return fmt.Errorf("failed updating services for node %s: %s", nodeID, err) } for service := services.Next(); service != nil; service = services.Next() { svc := service.(*structs.ServiceNode) if err := catalogUpdateServiceIndexes(tx, idx, svc.ServiceName, &svc.EnterpriseMeta, svc.PeerName); err != nil { return err } if err := catalogUpdateServiceKindIndexes(tx, idx, svc.ServiceKind, &svc.EnterpriseMeta, svc.PeerName); err != nil { return err } } return nil } // ensureCheckCASTxn updates a check only if the existing index matches the given index. // Returns a bool indicating if a write happened and any error. func (s *Store) ensureCheckCASTxn(tx WriteTxn, idx uint64, hc *structs.HealthCheck) (bool, error) { // Retrieve the existing entry. _, existing, err := getNodeCheckTxn(tx, hc.Node, hc.CheckID, &hc.EnterpriseMeta, hc.PeerName) if err != nil { return false, fmt.Errorf("failed health check lookup: %s", err) } // Check if the we should do the set. A ModifyIndex of 0 means that // we are doing a set-if-not-exists. if hc.ModifyIndex == 0 && existing != nil { return false, nil } if hc.ModifyIndex != 0 && existing == nil { return false, nil } if existing != nil && hc.ModifyIndex != 0 && hc.ModifyIndex != existing.ModifyIndex { return false, nil } // Perform the update. if err := s.ensureCheckTxn(tx, idx, false, hc); err != nil { return false, err } return true, nil } // ensureCheckTxn is used as the inner method to handle inserting // a health check into the state store. It ensures safety against inserting // checks with no matching node or service. func (s *Store) ensureCheckTxn(tx WriteTxn, idx uint64, preserveIndexes bool, hc *structs.HealthCheck) error { // Check if we have an existing health check existing, err := tx.First(tableChecks, indexID, NodeCheckQuery{ EnterpriseMeta: hc.EnterpriseMeta, Node: hc.Node, CheckID: string(hc.CheckID), PeerName: hc.PeerName, }) if err != nil { return fmt.Errorf("failed health check lookup: %s", err) } // Set the indexes if existing != nil { existingCheck := existing.(*structs.HealthCheck) hc.CreateIndex = existingCheck.CreateIndex hc.ModifyIndex = existingCheck.ModifyIndex } else if !preserveIndexes { hc.CreateIndex = idx } // Use the default check status if none was provided if hc.Status == "" { hc.Status = api.HealthCritical } // Get the node node, err := tx.First(tableNodes, indexID, Query{ Value: hc.Node, EnterpriseMeta: hc.EnterpriseMeta, PeerName: hc.PeerName, }) if err != nil { return fmt.Errorf("failed node lookup: %s", err) } if node == nil { return ErrMissingNode } modified := true // If the check is associated with a service, check that we have // a registration for the service. if hc.ServiceID != "" { service, err := tx.First(tableServices, indexID, NodeServiceQuery{ EnterpriseMeta: hc.EnterpriseMeta, Node: hc.Node, Service: hc.ServiceID, PeerName: hc.PeerName, }) if err != nil { return fmt.Errorf("failed service lookup: %s", err) } if service == nil { return ErrMissingService } // Copy in the service name and tags svc := service.(*structs.ServiceNode) hc.ServiceName = svc.ServiceName hc.ServiceTags = svc.ServiceTags if existing != nil && existing.(*structs.HealthCheck).IsSame(hc) { modified = false } else { if err = catalogUpdateServiceIndexes(tx, idx, svc.ServiceName, &svc.EnterpriseMeta, svc.PeerName); err != nil { return err } if err := catalogUpdateServiceKindIndexes(tx, idx, svc.ServiceKind, &svc.EnterpriseMeta, svc.PeerName); err != nil { return err } } } else { if existing != nil && existing.(*structs.HealthCheck).IsSame(hc) { modified = false } else { // Since the check has been modified, it impacts all services of node // Update the status for all the services associated with this node err = updateAllServiceIndexesOfNode(tx, idx, hc.Node, &hc.EnterpriseMeta, hc.PeerName) if err != nil { return err } } } // Delete any sessions for this check if the health is critical. if hc.Status == api.HealthCritical && hc.PeerName == "" { sessions, err := checkSessionsTxn(tx, hc) if err != nil { return err } // Delete the session in a separate loop so we don't trash the // iterator. for _, sess := range sessions { if err := s.deleteSessionTxn(tx, idx, sess.Session, &sess.EnterpriseMeta); err != nil { return fmt.Errorf("failed deleting session: %s", err) } } } if !modified { return nil } if !preserveIndexes { hc.ModifyIndex = idx } return catalogInsertCheck(tx, hc, idx) } // NodeCheck is used to retrieve a specific check associated with the given // node. func (s *Store) NodeCheck(nodeName string, checkID types.CheckID, entMeta *acl.EnterpriseMeta, peerName string) (uint64, *structs.HealthCheck, error) { tx := s.db.Txn(false) defer tx.Abort() return getNodeCheckTxn(tx, nodeName, checkID, entMeta, peerName) } // nodeCheckTxn is used as the inner method to handle reading a health check // from the state store. func getNodeCheckTxn(tx ReadTxn, nodeName string, checkID types.CheckID, entMeta *acl.EnterpriseMeta, peerName string) (uint64, *structs.HealthCheck, error) { // Get the table index. idx := catalogChecksMaxIndex(tx, entMeta, peerName) // TODO: accept non-pointer value if entMeta == nil { entMeta = structs.DefaultEnterpriseMetaInDefaultPartition() } // Return the check. check, err := tx.First(tableChecks, indexID, NodeCheckQuery{ EnterpriseMeta: *entMeta, Node: nodeName, CheckID: string(checkID), PeerName: peerName, }) if err != nil { return 0, nil, fmt.Errorf("failed check lookup: %s", err) } if check != nil { return idx, check.(*structs.HealthCheck), nil } return idx, nil, nil } // NodeChecks is used to retrieve checks associated with the // given node from the state store. func (s *Store) NodeChecks(ws memdb.WatchSet, nodeName string, entMeta *acl.EnterpriseMeta, peerName string) (uint64, structs.HealthChecks, error) { tx := s.db.Txn(false) defer tx.Abort() if entMeta == nil { entMeta = structs.DefaultEnterpriseMetaInDefaultPartition() } // Get the table index. idx := catalogChecksMaxIndex(tx, entMeta, peerName) // Return the checks. iter, err := catalogListChecksByNode(tx, Query{ Value: nodeName, EnterpriseMeta: *entMeta, PeerName: peerName, }) if err != nil { return 0, nil, fmt.Errorf("failed check lookup: %s", err) } ws.Add(iter.WatchCh()) var results structs.HealthChecks for check := iter.Next(); check != nil; check = iter.Next() { results = append(results, check.(*structs.HealthCheck)) } return idx, results, nil } // ServiceChecks is used to get all checks associated with a // given service ID. The query is performed against a service // _name_ instead of a service ID. func (s *Store) ServiceChecks(ws memdb.WatchSet, serviceName string, entMeta *acl.EnterpriseMeta, peerName string) (uint64, structs.HealthChecks, error) { tx := s.db.Txn(false) defer tx.Abort() // Get the table index. idx := catalogChecksMaxIndex(tx, entMeta, peerName) if entMeta == nil { entMeta = structs.DefaultEnterpriseMetaInDefaultPartition() } q := Query{ Value: serviceName, EnterpriseMeta: *entMeta, PeerName: peerName, } iter, err := tx.Get(tableChecks, indexService, q) if err != nil { return 0, nil, fmt.Errorf("failed check lookup: %s", err) } ws.Add(iter.WatchCh()) var results structs.HealthChecks for check := iter.Next(); check != nil; check = iter.Next() { results = append(results, check.(*structs.HealthCheck)) } return idx, results, nil } // ServiceChecksByNodeMeta is used to get all checks associated with a // given service ID, filtered by the given node metadata values. The query // is performed against a service _name_ instead of a service ID. func (s *Store) ServiceChecksByNodeMeta(ws memdb.WatchSet, serviceName string, filters map[string]string, entMeta *acl.EnterpriseMeta, peerName string) (uint64, structs.HealthChecks, error) { tx := s.db.Txn(false) defer tx.Abort() // Get the table index. idx := maxIndexForService(tx, serviceName, true, true, entMeta, peerName) if entMeta == nil { entMeta = structs.DefaultEnterpriseMetaInDefaultPartition() } q := Query{ Value: serviceName, EnterpriseMeta: *entMeta, PeerName: peerName, } iter, err := tx.Get(tableChecks, indexService, q) if err != nil { return 0, nil, fmt.Errorf("failed check lookup: %s", err) } ws.Add(iter.WatchCh()) return parseChecksByNodeMeta(tx, ws, idx, iter, filters, entMeta, peerName) } // ChecksInState is used to query the state store for all checks // which are in the provided state. func (s *Store) ChecksInState(ws memdb.WatchSet, state string, entMeta *acl.EnterpriseMeta, peerName string) (uint64, structs.HealthChecks, error) { tx := s.db.Txn(false) defer tx.Abort() idx, iter, err := checksInStateTxn(tx, ws, state, entMeta, peerName) if err != nil { return 0, nil, err } var results structs.HealthChecks for check := iter.Next(); check != nil; check = iter.Next() { results = append(results, check.(*structs.HealthCheck)) } return idx, results, nil } // ChecksInStateByNodeMeta is used to query the state store for all checks // which are in the provided state, filtered by the given node metadata values. func (s *Store) ChecksInStateByNodeMeta(ws memdb.WatchSet, state string, filters map[string]string, entMeta *acl.EnterpriseMeta, peerName string) (uint64, structs.HealthChecks, error) { tx := s.db.Txn(false) defer tx.Abort() idx, iter, err := checksInStateTxn(tx, ws, state, entMeta, peerName) if err != nil { return 0, nil, err } return parseChecksByNodeMeta(tx, ws, idx, iter, filters, entMeta, peerName) } func checksInStateTxn(tx ReadTxn, ws memdb.WatchSet, state string, entMeta *acl.EnterpriseMeta, peerName string) (uint64, memdb.ResultIterator, error) { // Get the table index. idx := catalogChecksMaxIndex(tx, entMeta, peerName) if entMeta == nil { entMeta = structs.DefaultEnterpriseMetaInDefaultPartition() } // Query all checks if HealthAny is passed, otherwise use the index. var iter memdb.ResultIterator var err error if state == api.HealthAny { q := Query{ EnterpriseMeta: *entMeta, PeerName: peerName, } iter, err = tx.Get(tableChecks, indexID+"_prefix", q) } else { q := Query{ Value: state, EnterpriseMeta: *entMeta, PeerName: peerName, } iter, err = tx.Get(tableChecks, indexStatus, q) } if err != nil { return 0, nil, fmt.Errorf("failed check lookup: %s", err) } ws.Add(iter.WatchCh()) return idx, iter, err } // parseChecksByNodeMeta is a helper function used to deduplicate some // repetitive code for returning health checks filtered by node metadata fields. func parseChecksByNodeMeta( tx ReadTxn, ws memdb.WatchSet, idx uint64, iter memdb.ResultIterator, filters map[string]string, entMeta *acl.EnterpriseMeta, peerName string, ) (uint64, structs.HealthChecks, error) { if entMeta == nil { entMeta = structs.DefaultEnterpriseMetaInDefaultPartition() } // We don't want to track an unlimited number of nodes, so we pull a // top-level watch to use as a fallback. q := Query{ EnterpriseMeta: *entMeta, PeerName: peerName, } allNodes, err := tx.Get(tableNodes, indexID+"_prefix", q) if err != nil { return 0, nil, fmt.Errorf("failed nodes lookup: %s", err) } allNodesCh := allNodes.WatchCh() // Only take results for nodes that satisfy the node metadata filters. var results structs.HealthChecks for check := iter.Next(); check != nil; check = iter.Next() { healthCheck := check.(*structs.HealthCheck) watchCh, node, err := tx.FirstWatch(tableNodes, indexID, Query{ Value: healthCheck.Node, EnterpriseMeta: healthCheck.EnterpriseMeta, PeerName: healthCheck.PeerName, }) if err != nil { return 0, nil, fmt.Errorf("failed node lookup: %s", err) } if node == nil { return 0, nil, ErrMissingNode } // Add even the filtered nodes so we wake up if the node metadata // changes. ws.AddWithLimit(watchLimit, watchCh, allNodesCh) if structs.SatisfiesMetaFilters(node.(*structs.Node).Meta, filters) { results = append(results, healthCheck) } } return idx, results, nil } // DeleteCheck is used to delete a health check registration. func (s *Store) DeleteCheck(idx uint64, node string, checkID types.CheckID, entMeta *acl.EnterpriseMeta, peerName string) error { tx := s.db.WriteTxn(idx) defer tx.Abort() // Call the check deletion if err := s.deleteCheckTxn(tx, idx, node, checkID, entMeta, peerName); err != nil { return err } return tx.Commit() } // deleteCheckCASTxn is used to try doing a check delete operation with a given // raft index. If the CAS index specified is not equal to the last observed index for // the given check, then the call is a noop, otherwise a normal check delete is invoked. func (s *Store) deleteCheckCASTxn( tx WriteTxn, idx, cidx uint64, node string, checkID types.CheckID, entMeta *acl.EnterpriseMeta, peerName string, ) (bool, error) { // Try to retrieve the existing health check. _, hc, err := getNodeCheckTxn(tx, node, checkID, entMeta, peerName) if err != nil { return false, fmt.Errorf("check lookup failed: %s", err) } if hc == nil { return false, nil } // If the existing index does not match the provided CAS // index arg, then we shouldn't update anything and can safely // return early here. if hc.ModifyIndex != cidx { return false, nil } // Call the actual deletion if the above passed. if err := s.deleteCheckTxn(tx, idx, node, checkID, entMeta, peerName); err != nil { return false, err } return true, nil } // NodeServiceQuery is a type used to query the checks table. type NodeServiceQuery struct { Node string Service string PeerName string acl.EnterpriseMeta } func (q NodeServiceQuery) PeerOrEmpty() string { return q.PeerName } // NamespaceOrDefault exists because structs.EnterpriseMeta uses a pointer // receiver for this method. Remove once that is fixed. func (q NodeServiceQuery) NamespaceOrDefault() string { return q.EnterpriseMeta.NamespaceOrDefault() } // PartitionOrDefault exists because structs.EnterpriseMeta uses a pointer // receiver for this method. Remove once that is fixed. func (q NodeServiceQuery) PartitionOrDefault() string { return q.EnterpriseMeta.PartitionOrDefault() } // deleteCheckTxn is the inner method used to call a health // check deletion within an existing transaction. func (s *Store) deleteCheckTxn(tx WriteTxn, idx uint64, node string, checkID types.CheckID, entMeta *acl.EnterpriseMeta, peerName string) error { if entMeta == nil { entMeta = structs.DefaultEnterpriseMetaInDefaultPartition() } // Try to retrieve the existing health check. hc, err := tx.First(tableChecks, indexID, NodeCheckQuery{ EnterpriseMeta: *entMeta, Node: node, CheckID: string(checkID), PeerName: peerName, }) if err != nil { return fmt.Errorf("check lookup failed: %s", err) } if hc == nil { return nil } existing := hc.(*structs.HealthCheck) if existing != nil { // When no service is linked to this service, update all services of node if existing.ServiceID != "" { if err := catalogUpdateServiceIndexes(tx, idx, existing.ServiceName, &existing.EnterpriseMeta, existing.PeerName); err != nil { return err } svcRaw, err := tx.First(tableServices, indexID, NodeServiceQuery{ EnterpriseMeta: existing.EnterpriseMeta, Node: existing.Node, Service: existing.ServiceID, PeerName: existing.PeerName, }) if err != nil { return fmt.Errorf("failed retrieving service from state store: %v", err) } svc := svcRaw.(*structs.ServiceNode) if err := catalogUpdateServiceKindIndexes(tx, idx, svc.ServiceKind, &svc.EnterpriseMeta, svc.PeerName); err != nil { return err } } else { if err := updateAllServiceIndexesOfNode(tx, idx, existing.Node, &existing.EnterpriseMeta, existing.PeerName); err != nil { return fmt.Errorf("Failed to update services linked to deleted healthcheck: %s", err) } if err := catalogUpdateServicesIndexes(tx, idx, entMeta, existing.PeerName); err != nil { return err } } } // Delete the check from the DB and update the index. if err := tx.Delete(tableChecks, hc); err != nil { return fmt.Errorf("failed removing check: %s", err) } if err := catalogUpdateCheckIndexes(tx, idx, entMeta, peerName); err != nil { return err } if peerName == "" { // Delete any sessions for this check. sessions, err := checkSessionsTxn(tx, existing) if err != nil { return err } // Do the delete in a separate loop so we don't trash the iterator. for _, sess := range sessions { if err := s.deleteSessionTxn(tx, idx, sess.Session, &sess.EnterpriseMeta); err != nil { return fmt.Errorf("failed deleting session: %s", err) } } } return nil } // CombinedCheckServiceNodes is used to query all nodes and checks for both typical and Connect endpoints of a service func (s *Store) CombinedCheckServiceNodes(ws memdb.WatchSet, service structs.ServiceName, peerName string) (uint64, structs.CheckServiceNodes, error) { var ( resp structs.CheckServiceNodes maxIdx uint64 ) idx, csn, err := s.CheckServiceNodes(ws, service.Name, &service.EnterpriseMeta, peerName) if err != nil { return 0, nil, fmt.Errorf("failed to get downstream nodes for %q: %v", service, err) } if idx > maxIdx { maxIdx = idx } resp = append(resp, csn...) idx, csn, err = s.CheckConnectServiceNodes(ws, service.Name, &service.EnterpriseMeta, peerName) if err != nil { return 0, nil, fmt.Errorf("failed to get downstream connect nodes for %q: %v", service, err) } if idx > maxIdx { maxIdx = idx } resp = append(resp, csn...) return maxIdx, resp, nil } // CheckServiceNodes is used to query all nodes and checks for a given service. func (s *Store) CheckServiceNodes(ws memdb.WatchSet, serviceName string, entMeta *acl.EnterpriseMeta, peerName string) (uint64, structs.CheckServiceNodes, error) { return s.checkServiceNodes(ws, serviceName, false, entMeta, peerName) } // CheckConnectServiceNodes is used to query all nodes and checks for Connect // compatible endpoints for a given service. func (s *Store) CheckConnectServiceNodes(ws memdb.WatchSet, serviceName string, entMeta *acl.EnterpriseMeta, peerName string) (uint64, structs.CheckServiceNodes, error) { return s.checkServiceNodes(ws, serviceName, true, entMeta, peerName) } // CheckIngressServiceNodes is used to query all nodes and checks for ingress // endpoints for a given service. func (s *Store) CheckIngressServiceNodes(ws memdb.WatchSet, serviceName string, entMeta *acl.EnterpriseMeta) (uint64, structs.CheckServiceNodes, error) { tx := s.db.Txn(false) defer tx.Abort() maxIdx, nodes, err := serviceGatewayNodes(tx, ws, serviceName, structs.ServiceKindIngressGateway, entMeta, structs.DefaultPeerKeyword) if err != nil { return 0, nil, fmt.Errorf("failed gateway nodes lookup: %v", err) } // TODO(ingress) : Deal with incorporating index from mapping table // Watch for index changes to the gateway nodes idx, chans := maxIndexAndWatchChsForServiceNodes(tx, nodes, false) for _, ch := range chans { ws.Add(ch) } maxIdx = lib.MaxUint64(maxIdx, idx) // TODO(ingress): Test namespace functionality here // De-dup services to lookup names := make(map[structs.ServiceName]struct{}) for _, n := range nodes { names[n.CompoundServiceName().ServiceName] = struct{}{} } var results structs.CheckServiceNodes for sn := range names { idx, n, err := checkServiceNodesTxn(tx, ws, sn.Name, false, &sn.EnterpriseMeta, structs.DefaultPeerKeyword) if err != nil { return 0, nil, err } maxIdx = lib.MaxUint64(maxIdx, idx) results = append(results, n...) } return maxIdx, results, nil } func (s *Store) checkServiceNodes(ws memdb.WatchSet, serviceName string, connect bool, entMeta *acl.EnterpriseMeta, peerName string) (uint64, structs.CheckServiceNodes, error) { tx := s.db.Txn(false) defer tx.Abort() return checkServiceNodesTxn(tx, ws, serviceName, connect, entMeta, peerName) } func checkServiceNodesTxn(tx ReadTxn, ws memdb.WatchSet, serviceName string, connect bool, entMeta *acl.EnterpriseMeta, peerName string) (uint64, structs.CheckServiceNodes, error) { index := indexService if connect { index = indexConnect } // TODO: accept non-pointer if entMeta == nil { entMeta = structs.DefaultEnterpriseMetaInDefaultPartition() } q := Query{ Value: serviceName, EnterpriseMeta: *entMeta, PeerName: peerName, } iter, err := tx.Get(tableServices, index, q) if err != nil { return 0, nil, fmt.Errorf("failed service lookup: %s", err) } // Note we decide if we want to watch this iterator or not down below. We need // to see if it returned anything first. // Return the results. var results structs.ServiceNodes // For connect queries we need a list of any proxy service names in the result // set. Rather than have different code path for connect and non-connect, we // use the same one in both cases. For non-empty non-connect results, // serviceNames will always have exactly one element which is the same as // serviceName. For Connect there might be multiple different service names - // one for each service name a proxy is registered under, and the target // service name IFF there is at least one Connect-native instance of that // service. Either way there is usually only one distinct name if proxies are // named consistently but could be multiple. serviceNames := make(map[structs.ServiceName]struct{}, 2) for service := iter.Next(); service != nil; service = iter.Next() { sn := service.(*structs.ServiceNode) results = append(results, sn) name := structs.NewServiceName(sn.ServiceName, &sn.EnterpriseMeta) serviceNames[name] = struct{}{} } // If we are querying for Connect nodes, the associated proxy might be a terminating-gateway. // Gateways are tracked in a separate table, and we append them to the result set. // We append rather than replace since it allows users to migrate a service // to the mesh with a mix of sidecars and gateways until all its instances have a sidecar. var idx uint64 if connect && peerName == "" { // Look up gateway nodes associated with the service // TODO(peering): we'll have to do something here gwIdx, nodes, err := serviceGatewayNodes(tx, ws, serviceName, structs.ServiceKindTerminatingGateway, entMeta, structs.DefaultPeerKeyword) if err != nil { return 0, nil, fmt.Errorf("failed gateway nodes lookup: %v", err) } idx = lib.MaxUint64(idx, gwIdx) for i := 0; i < len(nodes); i++ { results = append(results, nodes[i]) name := structs.NewServiceName(nodes[i].ServiceName, &nodes[i].EnterpriseMeta) serviceNames[name] = struct{}{} } } // watchOptimized tracks if we meet the necessary condition to optimize // WatchSet size. That is that every service name represented in the result // set must have a service-specific index we can watch instead of many radix // nodes for all the actual nodes touched. This saves us watching potentially // thousands of watch chans for large services which may need many goroutines. // It also avoids the performance cliff that is hit when watchLimit is hit // (~682 service instances). See // https://github.com/hashicorp/consul/issues/4984 watchOptimized := false if len(serviceNames) > 0 { // Assume optimization will work since it really should at this point. For // safety we'll sanity check this below for each service name. watchOptimized = true // Fetch indexes for all names services in result set. for n := range serviceNames { // We know service values should exist since the serviceNames map is only // populated if there is at least one result above. so serviceExists arg // below is always true. svcIdx, svcCh := maxIndexAndWatchChForService(tx, n.Name, true, true, &n.EnterpriseMeta, peerName) // Take the max index represented idx = lib.MaxUint64(idx, svcIdx) if svcCh != nil { // Watch the service-specific index for changes in liu of all iradix nodes // for checks etc. ws.Add(svcCh) } else { // Nil svcCh shouldn't really happen since all existent services should // have a service-specific index but just in case it does due to a bug, // fall back to the more expensive old way of watching every radix node // we touch. watchOptimized = false } } } else { // If we have no results, we should use the index of the last service // extinction event so we don't go backwards when services deregister. We // use target serviceName here but it actually doesn't matter. No chan will // be returned as we can't use the optimization in this case (and don't need // to as there is only one chan to watch anyway). svcIdx, _ := maxIndexAndWatchChForService(tx, serviceName, false, true, entMeta, peerName) idx = lib.MaxUint64(idx, svcIdx) } // Create a nil watchset to pass below, we'll only pass the real one if we // need to. Nil watchers are safe/allowed and saves some allocation too. var fallbackWS memdb.WatchSet if !watchOptimized { // We weren't able to use the optimization of watching only service indexes // for some reason. That means we need to fallback to watching everything we // touch in the DB as normal. We plumb the caller's watchset through (note // it's a map so this is a by-reference assignment.) fallbackWS = ws // We also need to watch the iterator from earlier too. fallbackWS.Add(iter.WatchCh()) } else if connect { // If this is a connect query then there is a subtlety to watch out for. // In addition to watching the proxy service indexes for changes above, we // need to still keep an eye on the connect service index in case a new // proxy with a new name registers - we are only watching proxy service // names we know about above so we'd miss that otherwise. Thankfully this // is only ever one extra chan to watch and will catch any changes to // proxy registrations for this target service. ws.Add(iter.WatchCh()) } return parseCheckServiceNodes(tx, fallbackWS, idx, results, entMeta, peerName, err) } // CheckServiceTagNodes is used to query all nodes and checks for a given // service, filtering out services that don't contain the given tag. func (s *Store) CheckServiceTagNodes(ws memdb.WatchSet, serviceName string, tags []string, entMeta *acl.EnterpriseMeta, peerName string) (uint64, structs.CheckServiceNodes, error) { tx := s.db.Txn(false) defer tx.Abort() // TODO: accept non-pointer value if entMeta == nil { entMeta = structs.DefaultEnterpriseMetaInDefaultPartition() } q := Query{Value: serviceName, EnterpriseMeta: *entMeta, PeerName: peerName} iter, err := tx.Get(tableServices, indexService, q) if err != nil { return 0, nil, fmt.Errorf("failed service lookup: %s", err) } ws.Add(iter.WatchCh()) // Return the results, filtering by tag. serviceExists := false var results structs.ServiceNodes for service := iter.Next(); service != nil; service = iter.Next() { svc := service.(*structs.ServiceNode) serviceExists = true if !serviceTagsFilter(svc, tags) { results = append(results, svc) } } // Get the table index. idx := maxIndexForService(tx, serviceName, serviceExists, true, entMeta, peerName) return parseCheckServiceNodes(tx, ws, idx, results, entMeta, peerName, err) } // GatewayServices is used to query all services associated with a gateway func (s *Store) GatewayServices(ws memdb.WatchSet, gateway string, entMeta *acl.EnterpriseMeta) (uint64, structs.GatewayServices, error) { tx := s.db.Txn(false) defer tx.Abort() iter, err := tx.Get(tableGatewayServices, indexGateway, structs.NewServiceName(gateway, entMeta)) if err != nil { return 0, nil, fmt.Errorf("failed gateway services lookup: %s", err) } ws.Add(iter.WatchCh()) maxIdx, results, err := s.collectGatewayServices(tx, ws, iter) if err != nil { return 0, nil, err } idx := maxIndexTxn(tx, tableGatewayServices) return lib.MaxUint64(maxIdx, idx), results, nil } // TODO: Find a way to consolidate this with CheckIngressServiceNodes // ServiceGateways is used to query all gateways associated with a service func (s *Store) ServiceGateways(ws memdb.WatchSet, service string, kind structs.ServiceKind, entMeta acl.EnterpriseMeta) (uint64, structs.CheckServiceNodes, error) { tx := s.db.Txn(false) defer tx.Abort() // tableGatewayServices is not peer-aware, and the existence of TG/IG gateways is scrubbed during peer replication. maxIdx, nodes, err := serviceGatewayNodes(tx, ws, service, kind, &entMeta, structs.DefaultPeerKeyword) // Watch for index changes to the gateway nodes idx, chans := maxIndexAndWatchChsForServiceNodes(tx, nodes, false) for _, ch := range chans { ws.Add(ch) } maxIdx = lib.MaxUint64(maxIdx, idx) return parseCheckServiceNodes(tx, ws, maxIdx, nodes, &entMeta, structs.DefaultPeerKeyword, err) } func (s *Store) VirtualIPForService(psn structs.PeeredServiceName) (string, error) { tx := s.db.Txn(false) defer tx.Abort() vip, err := tx.First(tableServiceVirtualIPs, indexID, psn) if err != nil { return "", fmt.Errorf("failed service virtual IP lookup: %s", err) } if vip == nil { return "", nil } return vip.(ServiceVirtualIP).IPWithOffset() } func (s *Store) ServiceVirtualIPs() (uint64, []ServiceVirtualIP, error) { tx := s.db.Txn(false) defer tx.Abort() return servicesVirtualIPsTxn(tx, nil) } func servicesVirtualIPsTxn(tx ReadTxn, ws memdb.WatchSet) (uint64, []ServiceVirtualIP, error) { iter, err := tx.Get(tableServiceVirtualIPs, indexID) if err != nil { return 0, nil, err } ws.Add(iter.WatchCh()) var vips []ServiceVirtualIP for raw := iter.Next(); raw != nil; raw = iter.Next() { vip := raw.(ServiceVirtualIP) vips = append(vips, vip) } idx := maxIndexWatchTxn(tx, nil, tableServiceVirtualIPs) return idx, vips, nil } func (s *Store) ServiceManualVIPs(psn structs.PeeredServiceName) (*ServiceVirtualIP, error) { tx := s.db.Txn(false) defer tx.Abort() return serviceVIPsTxn(tx, psn) } func serviceVIPsTxn(tx ReadTxn, psn structs.PeeredServiceName) (*ServiceVirtualIP, error) { vip, err := tx.First(tableServiceVirtualIPs, indexID, psn) if err != nil { return nil, fmt.Errorf("failed service virtual IP lookup: %s", err) } if vip == nil { return nil, nil } entry := vip.(ServiceVirtualIP) return &entry, nil } // VirtualIPsForAllImportedServices returns a slice of ServiceVirtualIP for all // VirtualIP-assignable services that have been imported by the partition represented in entMeta. // Namespace is ignored. func (s *Store) VirtualIPsForAllImportedServices(ws memdb.WatchSet, entMeta acl.EnterpriseMeta) (uint64, []ServiceVirtualIP, error) { tx := s.db.ReadTxn() defer tx.Abort() q := Query{ EnterpriseMeta: entMeta, // Wildcard peername is used by prefix index to fetch all remote peers for a partition. PeerName: "*", } iter, err := tx.Get(tableServiceVirtualIPs, indexID+"_prefix", q) if err != nil { return 0, nil, fmt.Errorf("failed service virtual IP lookup: %s", err) } ws.Add(iter.WatchCh()) idx := maxIndexTxn(tx, partitionedIndexEntryName(tableServiceVirtualIPs+".imported", entMeta.PartitionOrDefault())) var vips []ServiceVirtualIP for raw := iter.Next(); raw != nil; raw = iter.Next() { vip := raw.(ServiceVirtualIP) vips = append(vips, vip) } return idx, vips, nil } func (s *Store) ServiceNamesOfKind(ws memdb.WatchSet, kind structs.ServiceKind) (uint64, []*KindServiceName, error) { tx := s.db.Txn(false) defer tx.Abort() wildcardMeta := structs.WildcardEnterpriseMetaInPartition(structs.WildcardSpecifier) return serviceNamesOfKindTxn(tx, ws, kind, *wildcardMeta) } func serviceNamesOfKindTxn(tx ReadTxn, ws memdb.WatchSet, kind structs.ServiceKind, entMeta acl.EnterpriseMeta) (uint64, []*KindServiceName, error) { iter, err := tx.Get(tableKindServiceNames, indexKind, Query{Value: string(kind), EnterpriseMeta: entMeta}) if err != nil { return 0, nil, err } // TODO(peering): Maybe delete this watch and rely on the max idx tables below, to avoid waking up on unrelated changes ws.Add(iter.WatchCh()) var names []*KindServiceName for name := iter.Next(); name != nil; name = iter.Next() { ksn := name.(*KindServiceName) names = append(names, ksn) } var idx uint64 switch { case entMeta.PartitionOrDefault() == structs.WildcardSpecifier: idx = kindServiceNamesMaxIndex(tx, ws, kind.Normalized()) case entMeta.NamespaceOrDefault() == structs.WildcardSpecifier: idx = kindServiceNamesMaxIndex(tx, ws, partitionedIndexEntryName(kind.Normalized(), entMeta.PartitionOrDefault())) default: idx = kindServiceNamesMaxIndex(tx, ws, partitionedAndNamespacedIndexEntryName(kind.Normalized(), &entMeta)) } return idx, names, nil } // parseCheckServiceNodes is used to parse through a given set of services, // and query for an associated node and a set of checks. This is the inner // method used to return a rich set of results from a more simple query. // // TODO: idx parameter is not used except as a return value. Remove it. // TODO: err parameter is only used for early return. Remove it and check from the // caller. func parseCheckServiceNodes( tx ReadTxn, ws memdb.WatchSet, idx uint64, services structs.ServiceNodes, entMeta *acl.EnterpriseMeta, peerName string, err error, ) (uint64, structs.CheckServiceNodes, error) { if err != nil { return 0, nil, err } // Special-case the zero return value to nil, since this ends up in // external APIs. if len(services) == 0 { return idx, nil, nil } if entMeta == nil { entMeta = structs.DefaultEnterpriseMetaInDefaultPartition() } // We don't want to track an unlimited number of nodes, so we pull a // top-level watch to use as a fallback. allNodes, err := tx.Get(tableNodes, indexID+"_prefix", Query{ EnterpriseMeta: *entMeta, PeerName: peerName, }) if err != nil { return 0, nil, fmt.Errorf("failed nodes lookup: %s", err) } allNodesCh := allNodes.WatchCh() // We need a similar fallback for checks. Since services need the // status of node + service-specific checks, we pull in a top-level // watch over all checks. allChecks, err := tx.Get(tableChecks, indexID+"_prefix", Query{ EnterpriseMeta: *entMeta, PeerName: peerName, }) if err != nil { return 0, nil, fmt.Errorf("failed checks lookup: %s", err) } allChecksCh := allChecks.WatchCh() results := make(structs.CheckServiceNodes, 0, len(services)) for _, sn := range services { // Retrieve the node. watchCh, n, err := tx.FirstWatch(tableNodes, indexID, Query{ Value: sn.Node, EnterpriseMeta: sn.EnterpriseMeta, PeerName: sn.PeerName, }) if err != nil { return 0, nil, fmt.Errorf("failed node lookup: %s", err) } ws.AddWithLimit(watchLimit, watchCh, allNodesCh) if n == nil { return 0, nil, ErrMissingNode } node := n.(*structs.Node) // First add the node-level checks. These always apply to any // service on the node. var checks structs.HealthChecks q := NodeServiceQuery{ Node: sn.Node, Service: "", // node checks have no service EnterpriseMeta: *sn.EnterpriseMeta.WithWildcardNamespace(), PeerName: sn.PeerName, } iter, err := tx.Get(tableChecks, indexNodeService, q) if err != nil { return 0, nil, err } ws.AddWithLimit(watchLimit, iter.WatchCh(), allChecksCh) for check := iter.Next(); check != nil; check = iter.Next() { checks = append(checks, check.(*structs.HealthCheck)) } // Now add the service-specific checks. q = NodeServiceQuery{ Node: sn.Node, Service: sn.ServiceID, EnterpriseMeta: sn.EnterpriseMeta, PeerName: sn.PeerName, } iter, err = tx.Get(tableChecks, indexNodeService, q) if err != nil { return 0, nil, err } ws.AddWithLimit(watchLimit, iter.WatchCh(), allChecksCh) for check := iter.Next(); check != nil; check = iter.Next() { checks = append(checks, check.(*structs.HealthCheck)) } // Append to the results. results = append(results, structs.CheckServiceNode{ Node: node, Service: sn.ToNodeService(), Checks: checks, }) } return idx, results, nil } // NodeInfo is used to generate a dump of a single node. The dump includes // all services and checks which are registered against the node. func (s *Store) NodeInfo(ws memdb.WatchSet, node string, entMeta *acl.EnterpriseMeta, peerName string) (uint64, structs.NodeDump, error) { tx := s.db.Txn(false) defer tx.Abort() if entMeta == nil { entMeta = structs.NodeEnterpriseMetaInDefaultPartition() } // Get the table index. idx := catalogMaxIndex(tx, entMeta, peerName, true) // Query the node by the passed node nodes, err := tx.Get(tableNodes, indexID, Query{ Value: node, EnterpriseMeta: *entMeta, PeerName: peerName, }) if err != nil { return 0, nil, fmt.Errorf("failed node lookup: %s", err) } ws.Add(nodes.WatchCh()) return parseNodes(tx, ws, idx, nodes, entMeta, peerName) } // NodeDump is used to generate a dump of all nodes. This call is expensive // as it has to query every node, service, and check. The response can also // be quite large since there is currently no filtering applied. func (s *Store) NodeDump(ws memdb.WatchSet, entMeta *acl.EnterpriseMeta, peerName string) (uint64, structs.NodeDump, error) { tx := s.db.Txn(false) defer tx.Abort() if entMeta == nil { entMeta = structs.NodeEnterpriseMetaInDefaultPartition() } // Get the table index. idx := catalogMaxIndex(tx, entMeta, peerName, true) // Fetch all of the registered nodes q := Query{ EnterpriseMeta: *entMeta, PeerName: peerName, } nodes, err := tx.Get(tableNodes, indexID+"_prefix", q) if err != nil { return 0, nil, fmt.Errorf("failed node lookup: %s", err) } ws.Add(nodes.WatchCh()) return parseNodes(tx, ws, idx, nodes, entMeta, peerName) } func (s *Store) ServiceDump(ws memdb.WatchSet, kind structs.ServiceKind, useKind bool, entMeta *acl.EnterpriseMeta, peerName string) (uint64, structs.CheckServiceNodes, error) { tx := s.db.Txn(false) defer tx.Abort() if useKind { return serviceDumpKindTxn(tx, ws, kind, entMeta, peerName) } else { return serviceDumpAllTxn(tx, ws, entMeta, peerName) } } func serviceDumpAllTxn(tx ReadTxn, ws memdb.WatchSet, entMeta *acl.EnterpriseMeta, peerName string) (uint64, structs.CheckServiceNodes, error) { // Get the table index idx := catalogMaxIndexWatch(tx, ws, entMeta, "", true) if entMeta == nil { entMeta = structs.DefaultEnterpriseMetaInDefaultPartition() } q := Query{ EnterpriseMeta: *entMeta, PeerName: peerName, } services, err := tx.Get(tableServices, indexID+"_prefix", q) if err != nil { return 0, nil, fmt.Errorf("failed service lookup: %s", err) } var results structs.ServiceNodes for service := services.Next(); service != nil; service = services.Next() { sn := service.(*structs.ServiceNode) results = append(results, sn) } return parseCheckServiceNodes(tx, nil, idx, results, entMeta, peerName, err) } func serviceDumpKindTxn(tx ReadTxn, ws memdb.WatchSet, kind structs.ServiceKind, entMeta *acl.EnterpriseMeta, peerName string) (uint64, structs.CheckServiceNodes, error) { // unlike when we are dumping all services here we only need to watch the kind specific index entry for changing (or nodes, checks) // updating any services, nodes or checks will bump the appropriate service kind index so there is no need to watch any of the individual // entries idx := catalogServiceKindMaxIndex(tx, ws, kind, entMeta, peerName) if entMeta == nil { entMeta = structs.DefaultEnterpriseMetaInDefaultPartition() } q := Query{ Value: string(kind), EnterpriseMeta: *entMeta, PeerName: peerName, } services, err := tx.Get(tableServices, indexKind, q) if err != nil { return 0, nil, fmt.Errorf("failed service lookup: %s", err) } var results structs.ServiceNodes for service := services.Next(); service != nil; service = services.Next() { sn := service.(*structs.ServiceNode) results = append(results, sn) } return parseCheckServiceNodes(tx, nil, idx, results, entMeta, peerName, err) } // parseNodes takes an iterator over a set of nodes and returns a struct // containing the nodes along with all of their associated services // and/or health checks. // TODO(peering): support parsing by peerName func parseNodes(tx ReadTxn, ws memdb.WatchSet, idx uint64, iter memdb.ResultIterator, entMeta *acl.EnterpriseMeta, peerName string) (uint64, structs.NodeDump, error) { if entMeta == nil { entMeta = structs.DefaultEnterpriseMetaInDefaultPartition() } // We don't want to track an unlimited number of services, so we pull a // top-level watch to use as a fallback. q := Query{ EnterpriseMeta: *entMeta, PeerName: peerName, } allServices, err := tx.Get(tableServices, indexID+"_prefix", q) if err != nil { return 0, nil, fmt.Errorf("failed services lookup: %s", err) } allServicesCh := allServices.WatchCh() // We need a similar fallback for checks. allChecks, err := tx.Get(tableChecks, indexID+"_prefix", entMeta) if err != nil { return 0, nil, fmt.Errorf("failed checks lookup: %s", err) } allChecksCh := allChecks.WatchCh() var results structs.NodeDump for n := iter.Next(); n != nil; n = iter.Next() { node := n.(*structs.Node) // Create the wrapped node dump := &structs.NodeInfo{ ID: node.ID, Node: node.Node, Partition: node.Partition, PeerName: node.PeerName, Address: node.Address, TaggedAddresses: node.TaggedAddresses, Meta: node.Meta, } // Query the node services services, err := catalogServiceListByNode(tx, node.Node, entMeta, node.PeerName, true) if err != nil { return 0, nil, fmt.Errorf("failed services lookup: %s", err) } ws.AddWithLimit(watchLimit, services.WatchCh(), allServicesCh) for service := services.Next(); service != nil; service = services.Next() { ns := service.(*structs.ServiceNode).ToNodeService() dump.Services = append(dump.Services, ns) } // Query the service level checks checks, err := catalogListChecksByNode(tx, Query{ Value: node.Node, EnterpriseMeta: *entMeta, PeerName: node.PeerName, }) if err != nil { return 0, nil, fmt.Errorf("failed node lookup: %s", err) } ws.AddWithLimit(watchLimit, checks.WatchCh(), allChecksCh) for check := checks.Next(); check != nil; check = checks.Next() { hc := check.(*structs.HealthCheck) dump.Checks = append(dump.Checks, hc) } // Add the result to the slice results = append(results, dump) } return idx, results, nil } // checkSessionsTxn returns the IDs of all sessions associated with a health check func checkSessionsTxn(tx ReadTxn, hc *structs.HealthCheck) ([]*sessionCheck, error) { // TODO(peering): what are implications for imported health checks? mappings, err := tx.Get(tableSessionChecks, indexNodeCheck, MultiQuery{Value: []string{hc.Node, string(hc.CheckID)}, EnterpriseMeta: *structs.DefaultEnterpriseMetaInPartition(hc.PartitionOrDefault())}) if err != nil { return nil, fmt.Errorf("failed session checks lookup: %s", err) } var sessions []*sessionCheck for mapping := mappings.Next(); mapping != nil; mapping = mappings.Next() { sessions = append(sessions, mapping.(*sessionCheck)) } return sessions, nil } // updateGatewayServices associates services with gateways as specified in a gateway config entry func updateGatewayServices(tx WriteTxn, idx uint64, conf structs.ConfigEntry, entMeta *acl.EnterpriseMeta) error { var ( noChange bool gatewayServices structs.GatewayServices err error ) gateway := structs.NewServiceName(conf.GetName(), entMeta) switch conf.GetKind() { case structs.IngressGateway: noChange, gatewayServices, err = ingressConfigGatewayServices(tx, gateway, conf, entMeta) case structs.TerminatingGateway: noChange, gatewayServices, err = terminatingConfigGatewayServices(tx, gateway, conf, entMeta) default: return fmt.Errorf("config entry kind %q does not need gateway-services", conf.GetKind()) } // Return early if there is an error OR we don't have any services to update if err != nil || noChange { return err } // Update terminating gateway service virtual IPs vipsSupported, err := terminatingGatewayVirtualIPsSupported(tx, nil) if err != nil { return err } if vipsSupported && conf.GetKind() == structs.TerminatingGateway { gatewayConf := conf.(*structs.TerminatingGatewayConfigEntry) if err := updateTerminatingGatewayVirtualIPs(tx, idx, gatewayConf, entMeta); err != nil { return err } } // Delete all associated with gateway first, to avoid keeping mappings that were removed sn := structs.NewServiceName(conf.GetName(), entMeta) if _, err := tx.DeleteAll(tableGatewayServices, indexGateway, sn); err != nil { return fmt.Errorf("failed to truncate gateway services table: %v", err) } if err := truncateGatewayServiceTopologyMappings(tx, idx, sn, conf.GetKind()); err != nil { return fmt.Errorf("failed to truncate mesh topology for gateway: %v", err) } for _, svc := range gatewayServices { // If the service is a wildcard we need to target all services within the namespace if svc.Service.Name == structs.WildcardSpecifier { if err := updateGatewayNamespace(tx, idx, svc, entMeta); err != nil { return fmt.Errorf("failed to associate gateway %q with wildcard: %v", gateway.String(), err) } // Skip service-specific update below if there was a wildcard update continue } // Since this service was specified on its own, and not with a wildcard, // if there is an existing entry, we overwrite it. The service entry is the source of truth. // // By extension, if TLS creds are provided with a wildcard but are not provided in // the service entry, the service does not inherit the creds from the wildcard. err = updateGatewayService(tx, idx, svc) if err != nil { return err } } if err := indexUpdateMaxTxn(tx, idx, tableGatewayServices); err != nil { return fmt.Errorf("failed updating gateway-services index: %v", err) } return nil } func getTermGatewayVirtualIPs( tx WriteTxn, idx uint64, services []structs.LinkedService, entMeta *acl.EnterpriseMeta, ) (map[string]structs.ServiceAddress, error) { addrs := make(map[string]structs.ServiceAddress, len(services)) for _, s := range services { sn := structs.ServiceName{Name: s.Name, EnterpriseMeta: *entMeta} // Terminating Gateways cannot route to services in peered clusters psn := structs.PeeredServiceName{ServiceName: sn, Peer: structs.DefaultPeerKeyword} vip, err := assignServiceVirtualIP(tx, idx, psn) if err != nil { return nil, err } key := structs.ServiceGatewayVirtualIPTag(sn) addrs[key] = structs.ServiceAddress{Address: vip} } return addrs, nil } func updateTerminatingGatewayVirtualIPs(tx WriteTxn, idx uint64, conf *structs.TerminatingGatewayConfigEntry, entMeta *acl.EnterpriseMeta) error { // Build the current map of services with virtual IPs for this gateway services := conf.Services addrs, err := getTermGatewayVirtualIPs(tx, idx, services, entMeta) if err != nil { return err } // Find any deleted service entries by comparing the new config entry to the existing one. _, existing, err := configEntryTxn(tx, nil, conf.GetKind(), conf.GetName(), entMeta) if err != nil { return fmt.Errorf("failed to get config entry: %v", err) } var deletes []structs.ServiceName cfg, ok := existing.(*structs.TerminatingGatewayConfigEntry) if ok { for _, s := range cfg.Services { sn := structs.ServiceName{Name: s.Name, EnterpriseMeta: *entMeta} key := structs.ServiceGatewayVirtualIPTag(sn) if _, ok := addrs[key]; !ok { deletes = append(deletes, sn) } } } q := Query{Value: conf.GetName(), EnterpriseMeta: *entMeta} _, svcNodes, err := serviceNodesTxn(tx, nil, indexService, q) if err != nil { return err } // Update the tagged addrs for any existing instances of this terminating gateway. for _, s := range svcNodes { newAddrs := make(map[string]structs.ServiceAddress) for key, addr := range s.ServiceTaggedAddresses { if !strings.HasPrefix(key, structs.TaggedAddressVirtualIP+":") { newAddrs[key] = addr } } for key, addr := range addrs { newAddrs[key] = addr } // Don't need to update the service record if it's a no-op. if reflect.DeepEqual(newAddrs, s.ServiceTaggedAddresses) { continue } newSN := s.PartialClone() newSN.ServiceTaggedAddresses = newAddrs newSN.ModifyIndex = idx if err := catalogInsertService(tx, newSN); err != nil { return err } } // Check if we can delete any virtual IPs for the removed services. gatewayName := structs.NewServiceName(conf.GetName(), entMeta) for _, sn := range deletes { // If there's no existing service nodes, attempt to free the virtual IP. q := Query{Value: sn.Name, EnterpriseMeta: sn.EnterpriseMeta} _, nodes, err := serviceNodesTxn(tx, nil, indexConnect, q) if err != nil { return err } if len(nodes) == 0 { psn := structs.PeeredServiceName{Peer: structs.DefaultPeerKeyword, ServiceName: sn} if err := freeServiceVirtualIP(tx, idx, psn, &gatewayName); err != nil { return err } } } return nil } // ingressConfigGatewayServices constructs a list of GatewayService structs for // insertion into the memdb table, specific to ingress gateways. The boolean // returned indicates that there are no changes necessary to the memdb table. func ingressConfigGatewayServices( tx ReadTxn, gateway structs.ServiceName, conf structs.ConfigEntry, entMeta *acl.EnterpriseMeta, ) (bool, structs.GatewayServices, error) { entry, ok := conf.(*structs.IngressGatewayConfigEntry) if !ok { return false, nil, fmt.Errorf("unexpected config entry type: %T", conf) } // Check if service list matches the last known list for the config entry, if it does, skip the update _, c, err := configEntryTxn(tx, nil, conf.GetKind(), conf.GetName(), entMeta) if err != nil { return false, nil, fmt.Errorf("failed to get config entry: %v", err) } if cfg, ok := c.(*structs.IngressGatewayConfigEntry); ok && cfg != nil { if reflect.DeepEqual(cfg.Listeners, entry.Listeners) { // Services are the same, nothing to update return true, nil, nil } } var gatewayServices structs.GatewayServices for _, listener := range entry.Listeners { for _, service := range listener.Services { mapping := &structs.GatewayService{ Gateway: gateway, Service: service.ToServiceName(), GatewayKind: structs.ServiceKindIngressGateway, Hosts: service.Hosts, Port: listener.Port, Protocol: listener.Protocol, } gatewayServices = append(gatewayServices, mapping) } } return false, gatewayServices, nil } // terminatingConfigGatewayServices constructs a list of GatewayService structs // for insertion into the memdb table, specific to terminating gateways. The // boolean returned indicates that there are no changes necessary to the memdb // table. func terminatingConfigGatewayServices( tx ReadTxn, gateway structs.ServiceName, conf structs.ConfigEntry, entMeta *acl.EnterpriseMeta, ) (bool, structs.GatewayServices, error) { entry, ok := conf.(*structs.TerminatingGatewayConfigEntry) if !ok { return false, nil, fmt.Errorf("unexpected config entry type: %T", conf) } // Check if service list matches the last known list for the config entry, if it does, skip the update _, c, err := configEntryTxn(tx, nil, conf.GetKind(), conf.GetName(), entMeta) if err != nil { return false, nil, fmt.Errorf("failed to get config entry: %v", err) } if cfg, ok := c.(*structs.TerminatingGatewayConfigEntry); ok && cfg != nil { if reflect.DeepEqual(cfg.Services, entry.Services) { // Services are the same, nothing to update return true, nil, nil } } var gatewayServices structs.GatewayServices for _, svc := range entry.Services { kind, err := GatewayServiceKind(tx, svc.Name, &svc.EnterpriseMeta) if err != nil { return false, nil, fmt.Errorf("failed to get gateway service kind for service %s: %v", svc.Name, err) } mapping := &structs.GatewayService{ Gateway: gateway, Service: structs.NewServiceName(svc.Name, &svc.EnterpriseMeta), GatewayKind: structs.ServiceKindTerminatingGateway, KeyFile: svc.KeyFile, CertFile: svc.CertFile, CAFile: svc.CAFile, SNI: svc.SNI, ServiceKind: kind, } gatewayServices = append(gatewayServices, mapping) } return false, gatewayServices, nil } func GatewayServiceKind(tx ReadTxn, name string, entMeta *acl.EnterpriseMeta) (structs.GatewayServiceKind, error) { serviceIter, err := tx.First(tableServices, indexService, Query{ Value: name, EnterpriseMeta: *entMeta, }) if err != nil { return structs.GatewayServiceKindUnknown, err } if serviceIter != nil { return structs.GatewayServiceKindService, err } _, entry, err := configEntryTxn(tx, nil, structs.ServiceDefaults, name, entMeta) if err != nil { return structs.GatewayServiceKindUnknown, err } if entry != nil { sd, ok := entry.(*structs.ServiceConfigEntry) if !ok { return structs.GatewayServiceKindUnknown, fmt.Errorf("invalid config entry type %T", entry) } if sd.Destination != nil { return structs.GatewayServiceKindDestination, nil } } return structs.GatewayServiceKindUnknown, nil } // updateGatewayNamespace is used to target all services within a namespace func updateGatewayNamespace(tx WriteTxn, idx uint64, service *structs.GatewayService, entMeta *acl.EnterpriseMeta) error { if entMeta == nil { entMeta = structs.DefaultEnterpriseMetaInDefaultPartition() } q := Query{Value: string(structs.ServiceKindTypical), EnterpriseMeta: *entMeta} services, err := tx.Get(tableServices, indexKind, q) if err != nil { return fmt.Errorf("failed querying services: %s", err) } // Iterate over services in namespace and insert mapping for each for svc := services.Next(); svc != nil; svc = services.Next() { sn := svc.(*structs.ServiceNode) // Only associate non-consul services with gateways if sn.ServiceName == "consul" { continue } hasConnectInstance, hasNonConnectInstance, err := serviceHasConnectInstances(tx, sn.ServiceName, entMeta) if err != nil { return err } if service.GatewayKind == structs.ServiceKindIngressGateway && !hasConnectInstance { continue } if service.GatewayKind == structs.ServiceKindTerminatingGateway && !hasNonConnectInstance { continue } existing, err := tx.First(tableGatewayServices, indexID, service.Gateway, sn.CompoundServiceName().ServiceName, service.Port) if err != nil { return fmt.Errorf("gateway service lookup failed: %s", err) } if existing != nil { // If there's an existing service associated with this gateway then we skip it. // This means the service was specified on its own, and the service entry overrides the wildcard entry. continue } mapping := service.Clone() mapping.Service = structs.NewServiceName(sn.ServiceName, &service.Service.EnterpriseMeta) mapping.FromWildcard = true err = updateGatewayService(tx, idx, mapping) if err != nil { return err } } entries, err := tx.Get(tableConfigEntries, indexID+"_prefix", ConfigEntryKindQuery{Kind: structs.ServiceDefaults, EnterpriseMeta: *entMeta}) if err != nil { return fmt.Errorf("failed querying entries: %s", err) } for entry := entries.Next(); entry != nil; entry = entries.Next() { e := entry.(*structs.ServiceConfigEntry) if e.Destination == nil { continue } sn := structs.ServiceName{ Name: e.Name, EnterpriseMeta: e.EnterpriseMeta, } existing, err := tx.First(tableGatewayServices, indexID, service.Gateway, sn, service.Port) if err != nil { return fmt.Errorf("gateway service lookup failed: %s", err) } if existing != nil { // If there's an existing service associated with this gateway then we skip it. // This means the service was specified on its own, and the service entry overrides the wildcard entry. continue } mapping := service.Clone() mapping.Service = structs.NewServiceName(e.Name, &service.Service.EnterpriseMeta) mapping.ServiceKind = structs.GatewayServiceKindDestination mapping.FromWildcard = true err = updateGatewayService(tx, idx, mapping) if err != nil { return err } } // Also store a mapping for the wildcard so that the TLS creds can be pulled // for new services registered in its namespace err = updateGatewayService(tx, idx, service) if err != nil { return err } return nil } // serviceHasConnectInstances returns whether the service has at least one connect instance, // and at least one non-connect instance. func serviceHasConnectInstances(tx WriteTxn, serviceName string, entMeta *acl.EnterpriseMeta) (bool, bool, error) { hasConnectInstance := false query := Query{ Value: serviceName, EnterpriseMeta: *entMeta, } svc, err := tx.First(tableServices, indexConnect, query) if err != nil { return false, false, fmt.Errorf("failed service lookup: %s", err) } if svc != nil { hasConnectInstance = true } hasNonConnectInstance := false iter, err := tx.Get(tableServices, indexService, query) if err != nil { return false, false, fmt.Errorf("failed service lookup: %s", err) } for service := iter.Next(); service != nil; service = iter.Next() { sn := service.(*structs.ServiceNode) if !sn.ServiceConnect.Native { hasNonConnectInstance = true break } } return hasConnectInstance, hasNonConnectInstance, nil } // serviceHasConnectEnabledInstances returns whether the given service name // has a corresponding connect-proxy or connect-native instance. // This function is mostly a clone of `serviceHasConnectInstances`, but it has // an early return to improve performance and returns true if at least one // connect-native instance exists. func serviceHasConnectEnabledInstances(tx WriteTxn, serviceName string, entMeta *acl.EnterpriseMeta) (bool, error) { query := Query{ Value: serviceName, EnterpriseMeta: *entMeta, } svc, err := tx.First(tableServices, indexConnect, query) if err != nil { return false, fmt.Errorf("failed service lookup: %w", err) } if svc != nil { return true, nil } return false, nil } // updateGatewayService associates services with gateways after an eligible event // ie. Registering a service in a namespace targeted by a gateway func updateGatewayService(tx WriteTxn, idx uint64, mapping *structs.GatewayService) error { // Check if mapping already exists in table if it's already in the table // Avoid insert if nothing changed existing, err := tx.First(tableGatewayServices, indexID, mapping.Gateway, mapping.Service, mapping.Port) if err != nil { return fmt.Errorf("gateway service lookup failed: %s", err) } if gs, ok := existing.(*structs.GatewayService); ok && gs != nil { mapping.CreateIndex = gs.CreateIndex if gs.IsSame(mapping) { return nil } } else { // We have a new mapping mapping.CreateIndex = idx } mapping.ModifyIndex = idx if err := tx.Insert(tableGatewayServices, mapping); err != nil { return fmt.Errorf("failed inserting gateway service mapping: %s", err) } if err := indexUpdateMaxTxn(tx, idx, tableGatewayServices); err != nil { return fmt.Errorf("failed updating gateway-services index: %v", err) } if err := insertGatewayServiceTopologyMapping(tx, idx, mapping); err != nil { return fmt.Errorf("failed to reconcile mesh topology for gateway: %v", err) } return nil } // checkWildcardForGatewaysAndUpdate checks whether a service matches a // wildcard definition in gateway config entries and if so adds it the the // gateway-services table. func checkGatewayWildcardsAndUpdate(tx WriteTxn, idx uint64, svc *structs.ServiceName, ns *structs.NodeService, kind structs.GatewayServiceKind) error { sn := structs.ServiceName{Name: structs.WildcardSpecifier, EnterpriseMeta: svc.EnterpriseMeta} svcGateways, err := tx.Get(tableGatewayServices, indexService, sn) if err != nil { return fmt.Errorf("failed gateway lookup for %q: %s", svc.Name, err) } hasConnectInstance, hasNonConnectInstance, err := serviceHasConnectInstances(tx, svc.Name, &svc.EnterpriseMeta) if err != nil { return err } // If we were passed a NodeService, this might be the first registered instance of the service // so we need to count it as either a connect or non-connect instance. if ns != nil { if ns.Connect.Native || ns.Kind == structs.ServiceKindConnectProxy { hasConnectInstance = true } else { hasNonConnectInstance = true } } for service := svcGateways.Next(); service != nil; service = svcGateways.Next() { if wildcardSvc, ok := service.(*structs.GatewayService); ok && wildcardSvc != nil { if wildcardSvc.GatewayKind == structs.ServiceKindIngressGateway && !hasConnectInstance { continue } if wildcardSvc.GatewayKind == structs.ServiceKindTerminatingGateway && !hasNonConnectInstance && kind != structs.GatewayServiceKindDestination { continue } // Copy the wildcard mapping and modify it gatewaySvc := wildcardSvc.Clone() gatewaySvc.Service = structs.NewServiceName(svc.Name, &svc.EnterpriseMeta) gatewaySvc.FromWildcard = true gatewaySvc.ServiceKind = kind if err = updateGatewayService(tx, idx, gatewaySvc); err != nil { return fmt.Errorf("Failed to associate service %q with gateway %q", gatewaySvc.Service.String(), gatewaySvc.Gateway.String()) } } } return nil } // checkGatewayAndUpdate checks whether a service matches a // wildcard definition in gateway config entries and if so adds it the the // gateway-services table. func checkGatewayAndUpdate(tx WriteTxn, idx uint64, svc *structs.ServiceName, kind structs.GatewayServiceKind) error { sn := structs.ServiceName{Name: svc.Name, EnterpriseMeta: svc.EnterpriseMeta} svcGateways, err := tx.First(tableGatewayServices, indexService, sn) if err != nil { return fmt.Errorf("failed gateway lookup for %q: %s", svc.Name, err) } if service, ok := svcGateways.(*structs.GatewayService); ok && service != nil { // Copy the wildcard mapping and modify it gatewaySvc := service.Clone() gatewaySvc.Service = structs.NewServiceName(svc.Name, &svc.EnterpriseMeta) gatewaySvc.ServiceKind = kind if err = updateGatewayService(tx, idx, gatewaySvc); err != nil { return fmt.Errorf("Failed to associate service %q with gateway %q", gatewaySvc.Service.String(), gatewaySvc.Gateway.String()) } } return nil } func cleanupGatewayWildcards(tx WriteTxn, idx uint64, sn structs.ServiceName, cleaningUpDestination bool) error { // Clean up association between service name and gateways if needed gateways, err := tx.Get(tableGatewayServices, indexService, sn) if err != nil { return fmt.Errorf("failed gateway lookup for %q: %s", sn.Name, err) } mappings := make([]*structs.GatewayService, 0) for mapping := gateways.Next(); mapping != nil; mapping = gateways.Next() { if gs, ok := mapping.(*structs.GatewayService); ok && gs != nil { mappings = append(mappings, gs) } } // Check whether there are any connect or non-connect instances remaining for this service. // If there are no connect instances left, ingress gateways with a wildcard entry can remove // their association with it (same with terminating gateways if there are no non-connect // instances left). hasConnectInstance, hasNonConnectInstance, err := serviceHasConnectInstances(tx, sn.Name, &sn.EnterpriseMeta) if err != nil { return err } // If we're deleting a service instance but this service is defined as a destination via config entry, // keep the mapping around. hasDestination := false if !cleaningUpDestination { q := configentry.NewKindName(structs.ServiceDefaults, sn.Name, &sn.EnterpriseMeta) existing, err := tx.First(tableConfigEntries, indexID, q) if err != nil { return fmt.Errorf("failed config entry lookup: %s", err) } if existing != nil { if entry, ok := existing.(*structs.ServiceConfigEntry); ok && entry.Destination != nil { hasDestination = true } } } // Do the updates in a separate loop so we don't trash the iterator. for _, m := range mappings { // Only delete if association was created by a wildcard specifier. // Otherwise the service was specified in the config entry, and the association should be maintained // for when the service is re-registered if m.FromWildcard { if m.GatewayKind == structs.ServiceKindIngressGateway && hasConnectInstance { continue } if m.GatewayKind == structs.ServiceKindTerminatingGateway && (hasNonConnectInstance || hasDestination) { continue } if err := tx.Delete(tableGatewayServices, m); err != nil { return fmt.Errorf("failed to truncate gateway services table: %v", err) } if err := indexUpdateMaxTxn(tx, idx, tableGatewayServices); err != nil { return fmt.Errorf("failed updating gateway-services index: %v", err) } if err := deleteGatewayServiceTopologyMapping(tx, idx, m); err != nil { return fmt.Errorf("failed to reconcile mesh topology for gateway: %v", err) } } else { kind, err := GatewayServiceKind(tx, m.Service.Name, &m.Service.EnterpriseMeta) if err != nil { return fmt.Errorf("failed to get gateway service kind for service %s: %v", sn.Name, err) } checkGatewayAndUpdate(tx, idx, &structs.ServiceName{Name: m.Service.Name, EnterpriseMeta: m.Service.EnterpriseMeta}, kind) } } return nil } func (s *Store) DumpGatewayServices(ws memdb.WatchSet) (uint64, structs.GatewayServices, error) { tx := s.db.ReadTxn() defer tx.Abort() iter, err := tx.Get(tableGatewayServices, indexID) if err != nil { return 0, nil, fmt.Errorf("failed to dump gateway-services: %s", err) } ws.Add(iter.WatchCh()) maxIdx, results, err := s.collectGatewayServices(tx, ws, iter) if err != nil { return 0, nil, err } idx := maxIndexTxn(tx, tableGatewayServices) return lib.MaxUint64(maxIdx, idx), results, nil } func (s *Store) collectGatewayServices(tx ReadTxn, ws memdb.WatchSet, iter memdb.ResultIterator) (uint64, structs.GatewayServices, error) { var maxIdx uint64 var results structs.GatewayServices for obj := iter.Next(); obj != nil; obj = iter.Next() { gs := obj.(*structs.GatewayService) maxIdx = lib.MaxUint64(maxIdx, gs.ModifyIndex) if gs.Service.Name != structs.WildcardSpecifier { idx, matches, err := checkProtocolMatch(tx, ws, gs) if err != nil { return 0, nil, fmt.Errorf("failed checking protocol: %s", err) } maxIdx = lib.MaxUint64(maxIdx, idx) if matches { results = append(results, gs) } } } return maxIdx, results, nil } // TODO: How to handle index rolling back when a config entry is // deleted that references a service? // We might need something like the service_last_extinction index? func serviceGatewayNodes(tx ReadTxn, ws memdb.WatchSet, service string, kind structs.ServiceKind, entMeta *acl.EnterpriseMeta, peerName string) (uint64, structs.ServiceNodes, error) { if peerName != "" { return 0, nil, nil } // Look up gateway name associated with the service gws, err := tx.Get(tableGatewayServices, indexService, structs.NewServiceName(service, entMeta)) if err != nil { return 0, nil, fmt.Errorf("failed gateway lookup: %s", err) } // Adding this channel to the WatchSet means that the watch will fire if a config entry targeting the service is added. // Otherwise, if there's no associated gateway, then no watch channel would be returned ws.Add(gws.WatchCh()) var ret structs.ServiceNodes var maxIdx uint64 for gateway := gws.Next(); gateway != nil; gateway = gws.Next() { mapping := gateway.(*structs.GatewayService) // TODO(ingress): Test this conditional if mapping.GatewayKind != kind { continue } maxIdx = lib.MaxUint64(maxIdx, mapping.ModifyIndex) // Look up nodes for gateway q := Query{ Value: mapping.Gateway.Name, EnterpriseMeta: mapping.Gateway.EnterpriseMeta, } gwServices, err := tx.Get(tableServices, indexService, q) if err != nil { return 0, nil, fmt.Errorf("failed service lookup: %s", err) } var exists bool for svc := gwServices.Next(); svc != nil; svc = gwServices.Next() { sn := svc.(*structs.ServiceNode) ret = append(ret, sn) // Tracking existence to know whether we should check extinction index for service exists = true } // This prevents the index from sliding back if case all instances of the gateway service are deregistered svcIdx := maxIndexForService(tx, mapping.Gateway.Name, exists, false, &mapping.Gateway.EnterpriseMeta, structs.DefaultPeerKeyword) maxIdx = lib.MaxUint64(maxIdx, svcIdx) // Ensure that blocking queries wake up if the gateway-service mapping exists, but the gateway does not exist yet if !exists { ws.Add(gwServices.WatchCh()) } } return maxIdx, ret, nil } // metricsProtocolForIngressGateway determines the protocol that should be used when fetching metrics for an ingress gateway // Since ingress gateways may have listeners with different protocols, favor capturing all traffic by only returning HTTP // when all listeners are HTTP-like. func metricsProtocolForIngressGateway(tx ReadTxn, ws memdb.WatchSet, sn structs.ServiceName) (uint64, string, error) { idx, conf, err := configEntryTxn(tx, ws, structs.IngressGateway, sn.Name, &sn.EnterpriseMeta) if err != nil { return 0, "", fmt.Errorf("failed to get ingress-gateway config entry for %q: %v", sn.String(), err) } if conf == nil { return 0, "", nil } entry, ok := conf.(*structs.IngressGatewayConfigEntry) if !ok { return 0, "", fmt.Errorf("unexpected config entry type: %T", conf) } counts := make(map[string]int) for _, l := range entry.Listeners { if structs.IsProtocolHTTPLike(l.Protocol) { counts["http"] += 1 } else { counts["tcp"] += 1 } } protocol := "tcp" if counts["tcp"] == 0 && counts["http"] > 0 { protocol = "http" } return idx, protocol, nil } // checkProtocolMatch filters out any GatewayService entries added from a wildcard with a protocol // that doesn't match the one configured in their discovery chain. func checkProtocolMatch(tx ReadTxn, ws memdb.WatchSet, svc *structs.GatewayService) (uint64, bool, error) { if svc.GatewayKind != structs.ServiceKindIngressGateway || !svc.FromWildcard { return 0, true, nil } idx, protocol, err := protocolForService(tx, ws, svc.Service) if err != nil { return 0, false, err } return idx, svc.Protocol == protocol, nil } // TODO(freddy) Split this up. The upstream/downstream logic is very similar. // TODO(freddy) Add comprehensive state store test func (s *Store) ServiceTopology( ws memdb.WatchSet, dc, service string, kind structs.ServiceKind, defaultAllow acl.EnforcementDecision, entMeta *acl.EnterpriseMeta, ) (uint64, *structs.ServiceTopology, error) { tx := s.db.ReadTxn() defer tx.Abort() sn := structs.NewServiceName(service, entMeta) var ( maxIdx uint64 protocol string err error fullyTransparent bool hasTransparent bool connectNative bool ) switch kind { case structs.ServiceKindIngressGateway: maxIdx, protocol, err = metricsProtocolForIngressGateway(tx, ws, sn) if err != nil { return 0, nil, fmt.Errorf("failed to fetch protocol for service %s: %v", sn.String(), err) } case structs.ServiceKindTypical: maxIdx, protocol, err = protocolForService(tx, ws, sn) if err != nil { return 0, nil, fmt.Errorf("failed to fetch protocol for service %s: %v", sn.String(), err) } // Fetch connect endpoints for the target service in order to learn if its proxies are configured as // transparent proxies. if entMeta == nil { entMeta = structs.DefaultEnterpriseMetaInDefaultPartition() } q := Query{ Value: service, EnterpriseMeta: *entMeta, PeerName: structs.TODOPeerKeyword, } idx, proxies, err := serviceNodesTxn(tx, ws, indexConnect, q) if err != nil { return 0, nil, fmt.Errorf("failed to fetch connect endpoints for service %s: %v", sn.String(), err) } if idx > maxIdx { maxIdx = idx } if len(proxies) == 0 { break } fullyTransparent = true for _, proxy := range proxies { switch proxy.ServiceProxy.Mode { case structs.ProxyModeTransparent: hasTransparent = true default: // Only consider the target proxy to be transparent when all instances are in that mode. // This is done because the flag is used to display warnings about proxies needing to enable // transparent proxy mode. If ANY instance isn't in the right mode then the warming applies. fullyTransparent = false } if proxy.ServiceConnect.Native { connectNative = true } } default: return 0, nil, fmt.Errorf("unsupported kind %q", kind) } idx, upstreamNames, err := upstreamsFromRegistrationTxn(tx, ws, sn) if err != nil { return 0, nil, err } if idx > maxIdx { maxIdx = idx } var upstreamSources = make(map[string]string) for _, un := range upstreamNames { upstreamSources[un.String()] = structs.TopologySourceRegistration } upstreamDecisions := make(map[string]structs.IntentionDecisionSummary) // Only transparent proxies / connect native services have upstreams from intentions if hasTransparent || connectNative { idx, intentionUpstreams, err := s.intentionTopologyTxn(tx, ws, sn, false, defaultAllow, structs.IntentionTargetService) if err != nil { return 0, nil, err } if idx > maxIdx { maxIdx = idx } for _, svc := range intentionUpstreams { if _, ok := upstreamSources[svc.Name.String()]; ok { // Avoid duplicating entry continue } upstreamDecisions[svc.Name.String()] = svc.Decision upstreamNames = append(upstreamNames, svc.Name) var source string switch { case svc.Decision.HasExact: source = structs.TopologySourceSpecificIntention case svc.Decision.DefaultAllow: source = structs.TopologySourceDefaultAllow default: source = structs.TopologySourceWildcardIntention } upstreamSources[svc.Name.String()] = source } } matchEntry := structs.IntentionMatchEntry{ Namespace: entMeta.NamespaceOrDefault(), Partition: entMeta.PartitionOrDefault(), Name: service, } _, srcIntentions, err := compatIntentionMatchOneTxn(tx, ws, matchEntry, structs.IntentionMatchSource, structs.IntentionTargetService) if err != nil { return 0, nil, fmt.Errorf("failed to query intentions for %s", sn.String()) } for _, un := range upstreamNames { opts := IntentionDecisionOpts{ Target: un.Name, Namespace: un.NamespaceOrDefault(), Partition: un.PartitionOrDefault(), Intentions: srcIntentions, MatchType: structs.IntentionMatchDestination, DefaultDecision: defaultAllow, AllowPermissions: false, } decision, err := s.IntentionDecision(opts) if err != nil { return 0, nil, fmt.Errorf("failed to get intention decision from (%s) to (%s): %v", sn.String(), un.String(), err) } upstreamDecisions[un.String()] = decision } idx, unfilteredUpstreams, err := s.combinedServiceNodesTxn(tx, ws, upstreamNames, structs.DefaultPeerKeyword) if err != nil { return 0, nil, fmt.Errorf("failed to get upstreams for %q: %v", sn.String(), err) } if idx > maxIdx { maxIdx = idx } var upstreams structs.CheckServiceNodes for _, upstream := range unfilteredUpstreams { sn := upstream.Service.CompoundServiceName() if upstream.Service.Kind == structs.ServiceKindConnectProxy { sn = structs.NewServiceName(upstream.Service.Proxy.DestinationServiceName, &upstream.Service.EnterpriseMeta) } // Avoid returning upstreams from intentions when none of the proxy instances of the target are in transparent mode or connect native. if !hasTransparent && !connectNative && upstreamSources[sn.String()] != structs.TopologySourceRegistration { continue } upstreams = append(upstreams, upstream) } var foundUpstreams = make(map[structs.ServiceName]struct{}) for _, csn := range upstreams { foundUpstreams[csn.Service.CompoundServiceName()] = struct{}{} } // Check upstream names that had no service instances to see if they are routing config. for _, un := range upstreamNames { if _, ok := foundUpstreams[un]; ok { continue } for _, kind := range serviceGraphKinds { idx, entry, err := configEntryTxn(tx, ws, kind, un.Name, &un.EnterpriseMeta) if err != nil { return 0, nil, err } if entry != nil { upstreamSources[un.String()] = structs.TopologySourceRoutingConfig } if idx > maxIdx { maxIdx = idx } } } idx, downstreamNames, err := s.downstreamsForServiceTxn(tx, ws, dc, sn) if err != nil { return 0, nil, err } if idx > maxIdx { maxIdx = idx } var downstreamSources = make(map[string]string) for _, dn := range downstreamNames { downstreamSources[dn.String()] = structs.TopologySourceRegistration } idx, intentionDownstreams, err := s.intentionTopologyTxn(tx, ws, sn, true, defaultAllow, structs.IntentionTargetService) if err != nil { return 0, nil, err } if idx > maxIdx { maxIdx = idx } downstreamDecisions := make(map[string]structs.IntentionDecisionSummary) for _, svc := range intentionDownstreams { if _, ok := downstreamSources[svc.Name.String()]; ok { // Avoid duplicating entry continue } downstreamNames = append(downstreamNames, svc.Name) downstreamDecisions[svc.Name.String()] = svc.Decision var source string switch { case svc.Decision.HasExact: source = structs.TopologySourceSpecificIntention case svc.Decision.DefaultAllow: source = structs.TopologySourceDefaultAllow default: source = structs.TopologySourceWildcardIntention } downstreamSources[svc.Name.String()] = source } _, dstIntentions, err := compatIntentionMatchOneTxn(tx, ws, matchEntry, structs.IntentionMatchDestination, structs.IntentionTargetService) if err != nil { return 0, nil, fmt.Errorf("failed to query intentions for %s", sn.String()) } for _, dn := range downstreamNames { opts := IntentionDecisionOpts{ Target: dn.Name, Namespace: dn.NamespaceOrDefault(), Partition: dn.PartitionOrDefault(), Intentions: dstIntentions, MatchType: structs.IntentionMatchSource, DefaultDecision: defaultAllow, AllowPermissions: false, } decision, err := s.IntentionDecision(opts) if err != nil { return 0, nil, fmt.Errorf("failed to get intention decision from (%s) to (%s): %v", dn.String(), sn.String(), err) } downstreamDecisions[dn.String()] = decision } idx, unfilteredDownstreams, err := s.combinedServiceNodesTxn(tx, ws, downstreamNames, structs.DefaultPeerKeyword) if err != nil { return 0, nil, fmt.Errorf("failed to get downstreams for %q: %v", sn.String(), err) } if idx > maxIdx { maxIdx = idx } // Store downstreams with at least one instance in transparent proxy mode. // This is to avoid returning downstreams from intentions when none of the downstreams are transparent proxies. tproxyMap := make(map[structs.ServiceName]struct{}) for _, downstream := range unfilteredDownstreams { if downstream.Service.Proxy.Mode == structs.ProxyModeTransparent { sn := structs.NewServiceName(downstream.Service.Proxy.DestinationServiceName, &downstream.Service.EnterpriseMeta) tproxyMap[sn] = struct{}{} } } var downstreams structs.CheckServiceNodes for _, downstream := range unfilteredDownstreams { sn := downstream.Service.CompoundServiceName() if downstream.Service.Kind == structs.ServiceKindConnectProxy { sn = structs.NewServiceName(downstream.Service.Proxy.DestinationServiceName, &downstream.Service.EnterpriseMeta) } if _, ok := tproxyMap[sn]; !ok && !downstream.Service.Connect.Native && downstreamSources[sn.String()] != structs.TopologySourceRegistration { // If downstream is not a transparent proxy or connect native, remove references delete(downstreamSources, sn.String()) delete(downstreamDecisions, sn.String()) continue } downstreams = append(downstreams, downstream) } resp := &structs.ServiceTopology{ TransparentProxy: fullyTransparent, MetricsProtocol: protocol, Upstreams: upstreams, Downstreams: downstreams, UpstreamDecisions: upstreamDecisions, DownstreamDecisions: downstreamDecisions, UpstreamSources: upstreamSources, DownstreamSources: downstreamSources, } return maxIdx, resp, nil } // combinedServiceNodesTxn returns typical and connect endpoints for a list of services. // This enabled aggregating checks statuses across both. func (s *Store) combinedServiceNodesTxn(tx ReadTxn, ws memdb.WatchSet, names []structs.ServiceName, peerName string) (uint64, structs.CheckServiceNodes, error) { var ( maxIdx uint64 resp structs.CheckServiceNodes ) for _, u := range names { // Collect typical then connect instances idx, csn, err := checkServiceNodesTxn(tx, ws, u.Name, false, &u.EnterpriseMeta, peerName) if err != nil { return 0, nil, err } if idx > maxIdx { maxIdx = idx } resp = append(resp, csn...) idx, csn, err = checkServiceNodesTxn(tx, ws, u.Name, true, &u.EnterpriseMeta, peerName) if err != nil { return 0, nil, err } if idx > maxIdx { maxIdx = idx } resp = append(resp, csn...) } return maxIdx, resp, nil } // downstreamsForServiceTxn will find all downstream services that could route traffic to the input service. // There are two factors at play. Upstreams defined in a proxy registration, and the discovery chain for those upstreams. func (s *Store) downstreamsForServiceTxn(tx ReadTxn, ws memdb.WatchSet, dc string, service structs.ServiceName) (uint64, []structs.ServiceName, error) { // First fetch services that have discovery chains that eventually route to the target service idx, sources, err := s.discoveryChainSourcesTxn(tx, ws, dc, service) if err != nil { return 0, nil, fmt.Errorf("failed to get sources for discovery chain target %q: %v", service.String(), err) } var maxIdx uint64 if idx > maxIdx { maxIdx = idx } var ( resp []structs.ServiceName seen = make(map[structs.ServiceName]bool) ) for _, s := range sources { // We then follow these sources one level down to the services defining them as an upstream. idx, downstreams, err := downstreamsFromRegistrationTxn(tx, ws, s) if err != nil { return 0, nil, fmt.Errorf("failed to get registration downstreams for %q: %v", s.String(), err) } if idx > maxIdx { maxIdx = idx } for _, d := range downstreams { if !seen[d] { resp = append(resp, d) seen[d] = true } } } return maxIdx, resp, nil } // upstreamsFromRegistrationTxn returns the ServiceNames of the upstreams defined across instances of the input func upstreamsFromRegistrationTxn(tx ReadTxn, ws memdb.WatchSet, sn structs.ServiceName) (uint64, []structs.ServiceName, error) { return linkedFromRegistrationTxn(tx, ws, sn, false) } // downstreamsFromRegistrationTxn returns the ServiceNames of downstream services based on registrations across instances of the input func downstreamsFromRegistrationTxn(tx ReadTxn, ws memdb.WatchSet, sn structs.ServiceName) (uint64, []structs.ServiceName, error) { return linkedFromRegistrationTxn(tx, ws, sn, true) } func linkedFromRegistrationTxn(tx ReadTxn, ws memdb.WatchSet, service structs.ServiceName, downstreams bool) (uint64, []structs.ServiceName, error) { // To fetch upstreams we query services that have the input listed as a downstream // To fetch downstreams we query services that have the input listed as an upstream index := indexDownstream if downstreams { index = indexUpstream } iter, err := tx.Get(tableMeshTopology, index, service) if err != nil { return 0, nil, fmt.Errorf("%q lookup failed: %v", tableMeshTopology, err) } ws.Add(iter.WatchCh()) var ( idx uint64 resp []structs.ServiceName ) for raw := iter.Next(); raw != nil; raw = iter.Next() { entry := raw.(*upstreamDownstream) if entry.ModifyIndex > idx { idx = entry.ModifyIndex } linked := entry.Upstream if downstreams { linked = entry.Downstream } resp = append(resp, linked) } // TODO (freddy) This needs a tombstone to avoid the index sliding back on mapping deletion // Using the table index here means that blocking queries will wake up more often than they should tableIdx := maxIndexTxn(tx, tableMeshTopology) if tableIdx > idx { idx = tableIdx } return idx, resp, nil } // updateMeshTopology creates associations between the input service and its upstreams in the topology table func updateMeshTopology(tx WriteTxn, idx uint64, node string, svc *structs.NodeService, existing interface{}) error { // TODO(peering): make this peering aware oldUpstreams := make(map[structs.ServiceName]bool) if e, ok := existing.(*structs.ServiceNode); ok { for _, u := range e.ServiceProxy.Upstreams { upstreamMeta := acl.NewEnterpriseMetaWithPartition(e.PartitionOrDefault(), u.DestinationNamespace) sn := structs.NewServiceName(u.DestinationName, &upstreamMeta) oldUpstreams[sn] = true } } // Despite the name "destination", this service name is downstream of the proxy downstream := structs.NewServiceName(svc.Proxy.DestinationServiceName, &svc.EnterpriseMeta) inserted := make(map[structs.ServiceName]bool) for _, u := range svc.Proxy.Upstreams { if u.DestinationType == structs.UpstreamDestTypePreparedQuery { continue } // TODO (freddy): Account for upstream datacenter upstreamMeta := acl.NewEnterpriseMetaWithPartition(svc.PartitionOrDefault(), u.DestinationNamespace) upstream := structs.NewServiceName(u.DestinationName, &upstreamMeta) obj, err := tx.First(tableMeshTopology, indexID, upstream, downstream) if err != nil { return fmt.Errorf("%q lookup failed: %v", tableMeshTopology, err) } sid := svc.CompoundServiceID() uid := structs.UniqueID(node, sid.String()) var mapping *upstreamDownstream if existing, ok := obj.(*upstreamDownstream); ok { mapping := existing.DeepCopy() mapping.Refs[uid] = struct{}{} mapping.ModifyIndex = idx inserted[upstream] = true } if mapping == nil { mapping = &upstreamDownstream{ Upstream: upstream, Downstream: downstream, Refs: map[string]struct{}{uid: {}}, RaftIndex: structs.RaftIndex{ CreateIndex: idx, ModifyIndex: idx, }, } } if err := tx.Insert(tableMeshTopology, mapping); err != nil { return fmt.Errorf("failed inserting %s mapping: %s", tableMeshTopology, err) } if err := indexUpdateMaxTxn(tx, idx, tableMeshTopology); err != nil { return fmt.Errorf("failed updating %s index: %v", tableMeshTopology, err) } inserted[upstream] = true } for u := range oldUpstreams { if !inserted[u] { if _, err := tx.DeleteAll(tableMeshTopology, indexID, u, downstream); err != nil { return fmt.Errorf("failed to truncate %s table: %v", tableMeshTopology, err) } if err := indexUpdateMaxTxn(tx, idx, tableMeshTopology); err != nil { return fmt.Errorf("failed updating %s index: %v", tableMeshTopology, err) } } } return nil } // cleanupMeshTopology removes a service from the mesh topology table // This is only safe to call when there are no more known instances of this proxy func cleanupMeshTopology(tx WriteTxn, idx uint64, service *structs.ServiceNode) error { if service.PeerName != "" { return nil } if service.ServiceKind != structs.ServiceKindConnectProxy { return nil } sn := structs.NewServiceName(service.ServiceProxy.DestinationServiceName, &service.EnterpriseMeta) sid := service.CompoundServiceID() uid := structs.UniqueID(service.Node, sid.String()) iter, err := tx.Get(tableMeshTopology, indexDownstream, sn) if err != nil { return fmt.Errorf("%q lookup failed: %v", tableMeshTopology, err) } mappings := make([]*upstreamDownstream, 0) for raw := iter.Next(); raw != nil; raw = iter.Next() { mappings = append(mappings, raw.(*upstreamDownstream)) } // Do the updates in a separate loop so we don't trash the iterator. for _, m := range mappings { copy := m.DeepCopy() // Bail early if there's no reference to the proxy ID we're deleting if _, ok := copy.Refs[uid]; !ok { continue } delete(copy.Refs, uid) if len(copy.Refs) == 0 { if err := tx.Delete(tableMeshTopology, m); err != nil { return fmt.Errorf("failed to truncate %s table: %v", tableMeshTopology, err) } if err := indexUpdateMaxTxn(tx, idx, tableMeshTopology); err != nil { return fmt.Errorf("failed updating %s index: %v", tableMeshTopology, err) } continue } if err := tx.Insert(tableMeshTopology, copy); err != nil { return fmt.Errorf("failed inserting %s mapping: %s", tableMeshTopology, err) } } return nil } func insertGatewayServiceTopologyMapping(tx WriteTxn, idx uint64, gs *structs.GatewayService) error { // Only ingress gateways are standalone items in the mesh topology viz if gs.GatewayKind != structs.ServiceKindIngressGateway || gs.Service.Name == structs.WildcardSpecifier { return nil } mapping := upstreamDownstream{ Upstream: gs.Service, Downstream: gs.Gateway, Refs: make(map[string]struct{}), RaftIndex: gs.RaftIndex, } if err := tx.Insert(tableMeshTopology, &mapping); err != nil { return fmt.Errorf("failed inserting %s mapping: %s", tableMeshTopology, err) } if err := indexUpdateMaxTxn(tx, idx, tableMeshTopology); err != nil { return fmt.Errorf("failed updating %s index: %v", tableMeshTopology, err) } return nil } func deleteGatewayServiceTopologyMapping(tx WriteTxn, idx uint64, gs *structs.GatewayService) error { // Only ingress gateways are standalone items in the mesh topology viz if gs.GatewayKind != structs.ServiceKindIngressGateway { return nil } if _, err := tx.DeleteAll(tableMeshTopology, indexID, gs.Service, gs.Gateway); err != nil { return fmt.Errorf("failed to truncate %s table: %v", tableMeshTopology, err) } if err := indexUpdateMaxTxn(tx, idx, tableMeshTopology); err != nil { return fmt.Errorf("failed updating %s index: %v", tableMeshTopology, err) } return nil } func truncateGatewayServiceTopologyMappings(tx WriteTxn, idx uint64, gateway structs.ServiceName, kind string) error { // Only ingress gateways are standalone items in the mesh topology viz if kind != string(structs.ServiceKindIngressGateway) { return nil } if _, err := tx.DeleteAll(tableMeshTopology, indexDownstream, gateway); err != nil { return fmt.Errorf("failed to truncate %s table: %v", tableMeshTopology, err) } if err := indexUpdateMaxTxn(tx, idx, tableMeshTopology); err != nil { return fmt.Errorf("failed updating %s index: %v", tableMeshTopology, err) } return nil } func upsertKindServiceName(tx WriteTxn, idx uint64, kind structs.ServiceKind, name structs.ServiceName) error { // TODO(peering): make this peering aware q := KindServiceNameQuery{Name: name.Name, Kind: kind, EnterpriseMeta: name.EnterpriseMeta} existing, err := tx.First(tableKindServiceNames, indexID, q) if err != nil { return err } // Service name is already known. Nothing to do. if existing != nil { return nil } ksn := KindServiceName{ Kind: kind, Service: name, RaftIndex: structs.RaftIndex{ CreateIndex: idx, ModifyIndex: idx, }, } if err := tx.Insert(tableKindServiceNames, &ksn); err != nil { return fmt.Errorf("failed inserting %s/%s into %s: %s", kind, name.String(), tableKindServiceNames, err) } return updateKindServiceNamesIndex(tx, idx, kind, name.EnterpriseMeta) } func cleanupKindServiceName(tx WriteTxn, idx uint64, name structs.ServiceName, kind structs.ServiceKind) error { q := KindServiceNameQuery{Name: name.Name, Kind: kind, EnterpriseMeta: name.EnterpriseMeta} if _, err := tx.DeleteAll(tableKindServiceNames, indexID, q); err != nil { return fmt.Errorf("failed to delete %s from %s: %s", name, tableKindServiceNames, err) } return updateKindServiceNamesIndex(tx, idx, kind, name.EnterpriseMeta) } // CatalogDump returns all the contents of the node, service and check tables. // In Enterprise, this will return entries across all partitions and namespaces. // TODO(peering) make this peering aware? func (s *Store) CatalogDump() (*structs.CatalogContents, error) { tx := s.db.Txn(false) contents := &structs.CatalogContents{} nodes, err := tx.Get(tableNodes, indexID) if err != nil { return nil, fmt.Errorf("failed nodes lookup: %s", err) } for node := nodes.Next(); node != nil; node = nodes.Next() { n := node.(*structs.Node) contents.Nodes = append(contents.Nodes, n) } services, err := tx.Get(tableServices, indexID) if err != nil { return nil, fmt.Errorf("failed services lookup: %s", err) } for service := services.Next(); service != nil; service = services.Next() { svc := service.(*structs.ServiceNode) contents.Services = append(contents.Services, svc) } checks, err := tx.Get(tableChecks, indexID) if err != nil { return nil, fmt.Errorf("failed checks lookup: %s", err) } for check := checks.Next(); check != nil; check = checks.Next() { c := check.(*structs.HealthCheck) contents.Checks = append(contents.Checks, c) } return contents, nil }