removed the request from the pool if the the proposal is bad

integration/smartbft/smartbft_test.go

@@ -19,6 +19,7 @@ import (
 	"path/filepath"
 	"strconv"
 	"strings"
+	"sync"
 	"syscall"
 	"time"
 
@@ -1809,6 +1810,115 @@ var _ = Describe("EndToEnd Smart BFT configuration test", func() {
 			By("invoking the chaincode")
 			invokeQuery(network, peer, network.Orderers[1], channel, 90)
 		})
+
+		It("send an update transaction to each orderer and deleting failed requests", func() {
+			channel := "testchannel1"
+			By("Create network")
+			networkConfig := nwo.MultiNodeSmartBFT()
+			networkConfig.Channels = nil
+			network = nwo.New(networkConfig, testDir, client, StartPort(), components)
+			network.GenerateConfigTree()
+			network.Bootstrap()
+			network.EventuallyTimeout *= 2
+
+			By("Start orderers")
+			var ordererRunners []*ginkgomon.Runner
+			for _, orderer := range network.Orderers {
+				runner := network.OrdererRunner(orderer,
+					"FABRIC_LOGGING_SPEC=orderer.common.cluster=debug:orderer.consensus.smartbft=debug:policies.ImplicitOrderer=debug",
+					"ORDERER_GENERAL_BACKOFF_MAXDELAY=20s")
+				ordererRunners = append(ordererRunners, runner)
+				proc := ifrit.Invoke(runner)
+				ordererProcesses = append(ordererProcesses, proc)
+				Eventually(proc.Ready(), network.EventuallyTimeout).Should(BeClosed())
+			}
+
+			By("Start peer")
+			peerRunner := network.PeerGroupRunner()
+			peerProcesses = ifrit.Invoke(peerRunner)
+			Eventually(peerProcesses.Ready(), network.EventuallyTimeout).Should(BeClosed())
+			peer := network.Peer("Org1", "peer0")
+
+			By("Join network to channel")
+			joinChannel(network, channel)
+
+			By("Waiting for followers to see the leader")
+			Eventually(ordererRunners[1].Err(), network.EventuallyTimeout, time.Second).Should(gbytes.Say("Message from 1"))
+			Eventually(ordererRunners[2].Err(), network.EventuallyTimeout, time.Second).Should(gbytes.Say("Message from 1"))
+			Eventually(ordererRunners[3].Err(), network.EventuallyTimeout, time.Second).Should(gbytes.Say("Message from 1"))
+
+			By("Joining peers to channel")
+			orderer1 := network.Orderers[0]
+			network.JoinChannel(channel, orderer1, network.PeersWithChannel(channel)...)
+
+			By("Deploying chaincode")
+			deployChaincode(network, channel, testDir)
+
+			By("querying the chaincode")
+			sess, err := network.PeerUserSession(peer, "User1", commands.ChaincodeQuery{
+				ChannelID: channel,
+				Name:      "mycc",
+				Ctor:      `{"Args":["query","a"]}`,
+			})
+			Expect(err).NotTo(HaveOccurred())
+			Eventually(sess, network.EventuallyTimeout).Should(gexec.Exit(0))
+			Expect(sess).To(gbytes.Say("100"))
+
+			By("invoking the chaincode")
+			invokeQuery(network, peer, network.Orderers[1], channel, 90)
+
+			By("Changing the channel config, e.g. maximum number of bytes in a 1 MB batch")
+			newAbsoluteMaxBytes := 1_000_000
+			config := nwo.GetConfig(network, peer, orderer1, channel)
+			updatedConfig := proto.Clone(config).(*common.Config)
+			batchSizeConfigValue := updatedConfig.ChannelGroup.Groups["Orderer"].Values["BatchSize"]
+			batchSizeValue := &ordererProtos.BatchSize{}
+			Expect(proto.Unmarshal(batchSizeConfigValue.Value, batchSizeValue)).To(Succeed())
+			batchSizeValue.AbsoluteMaxBytes = uint32(newAbsoluteMaxBytes)
+			updatedConfig.ChannelGroup.Groups["Orderer"].Values["BatchSize"] = &common.ConfigValue{
+				ModPolicy: "Admins",
+				Value:     protoutil.MarshalOrPanic(batchSizeValue),
+			}
+
+			tempDir, err := os.MkdirTemp(network.RootDir, "updateConfig")
+			Expect(err).NotTo(HaveOccurred())
+			updateFile := filepath.Join(tempDir, "update.pb")
+			defer os.RemoveAll(tempDir)
+
+			nwo.ComputeUpdateOrdererConfig(updateFile, network, channel, config, updatedConfig, peer, orderer1)
+
+			fileData, err := os.ReadFile(updateFile)
+			Expect(err).NotTo(HaveOccurred())
+
+			ctxEnv, err := protoutil.UnmarshalEnvelope(fileData)
+			Expect(err).NotTo(HaveOccurred())
+
+			By("Begin broadcast")
+			var wg sync.WaitGroup
+			for i := range network.Orderers {
+				wg.Add(1)
+				go func(ccid int) {
+					defer wg.Done()
+					resp, err := ordererclient.Broadcast(network, network.Orderers[ccid], ctxEnv)
+					Expect(err).NotTo(HaveOccurred())
+					Expect(resp.Status).To(Equal(common.Status_SUCCESS))
+				}(i)
+			}
+			wg.Wait()
+			By("End broadcast")
+
+			// move the logger cursor to the beginning of the broadcast
+			Eventually(ordererRunners[0].Err(), network.EventuallyTimeout, time.Second).Should(gbytes.Say("verifyConfigUpdateMsg"))
+			Eventually(ordererRunners[1].Err(), network.EventuallyTimeout, time.Second).Should(gbytes.Say("verifyConfigUpdateMsg"))
+			Eventually(ordererRunners[2].Err(), network.EventuallyTimeout, time.Second).Should(gbytes.Say("verifyConfigUpdateMsg"))
+			Eventually(ordererRunners[3].Err(), network.EventuallyTimeout, time.Second).Should(gbytes.Say("verifyConfigUpdateMsg"))
+
+			// the channel update transaction will have to be deleted from each orderer
+			Eventually(ordererRunners[0].Err(), network.EventuallyTimeout, time.Second).Should(gbytes.Say("deleteRequest"))
+			Eventually(ordererRunners[1].Err(), network.EventuallyTimeout, time.Second).Should(gbytes.Say("deleteRequest"))
+			Eventually(ordererRunners[2].Err(), network.EventuallyTimeout, time.Second).Should(gbytes.Say("deleteRequest"))
+			Eventually(ordererRunners[3].Err(), network.EventuallyTimeout, time.Second).Should(gbytes.Say("deleteRequest"))
+		})
 	})
 })
 

orderer/consensus/smartbft/chain.go

@@ -312,6 +312,13 @@ func (c *BFTChain) pruneCommittedRequests(block *cb.Block) {
 	wg.Wait()
 }
 
+func (c *BFTChain) pruneBadRequests() {
+	c.consensus.Pool.Prune(func(req []byte) error {
+		_, err := c.consensus.Verifier.VerifyRequest(req)
+		return err
+	})
+}
+
 func (c *BFTChain) submit(env *cb.Envelope, configSeq uint64) error {
 	reqBytes, err := proto.Marshal(env)
 	if err != nil {
@@ -566,6 +573,7 @@ func (c *BFTChain) updateRuntimeConfig(block *cb.Block) types.Reconfig {
 	if protoutil.IsConfigBlock(block) {
 		c.Comm.Configure(c.Channel, newRTC.RemoteNodes)
 		c.clusterService.ConfigureNodeCerts(c.Channel, newRTC.consenters)
+		c.pruneBadRequests()
 	}
 
 	membershipDidNotChange := reflect.DeepEqual(newRTC.Nodes, prevRTC.Nodes)