Consistency checks for persistent cluster startup.

git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@884226 13f79535-47bb-0310-9956-ffa450edef68

8 files changed, 92 insertions, 44 deletions

diff --git a/cpp/src/qpid/cluster/Cluster.cpp b/cpp/src/qpid/cluster/Cluster.cpp
index 282b639f61..fa53fc5475 100644
--- a/cpp/src/qpid/cluster/Cluster.cpp
+++ b/cpp/src/qpid/cluster/Cluster.cpp
@@ -175,7 +175,7 @@ namespace _qmf = ::qmf::org::apache::qpid::cluster;
  * Currently use SVN revision to avoid clashes with versions from
  * different branches.
  */
-const uint32_t Cluster::CLUSTER_VERSION = 835547;
+const uint32_t Cluster::CLUSTER_VERSION = 884125;
 
 struct ClusterDispatcher : public framing::AMQP_AllOperations::ClusterHandler {
     qpid::cluster::Cluster& cluster;
@@ -202,7 +202,7 @@ struct ClusterDispatcher : public framing::AMQP_AllOperations::ClusterHandler {
         cluster.errorCheck(member, type, frameSeq, l);
     }
 
-    void shutdown() { cluster.shutdown(member, l); }
+    void shutdown(const Uuid& id) { cluster.shutdown(member, id, l); }
 
     bool invoke(AMQBody& body) { return framing::invoke(*this, body).wasHandled(); }
 };
@@ -287,7 +287,7 @@ void Cluster::initialize() {
       default:
         assert(0);
     }
-    QPID_LOG(notice, *this << (state == READY ? "joined" : "joining") << " cluster " << name << " with url=" << myUrl);
+    QPID_LOG(notice, *this << (state == READY ? " joined" : " joining") << " cluster " << name);
     broker.getKnownBrokers = boost::bind(&Cluster::getUrls, this);
     broker.setExpiryPolicy(expiryPolicy);
     dispatcher.start();
@@ -601,6 +601,7 @@ void Cluster::initMapCompleted(Lock& l) {
     // Called on completion of the initial status map. 
     if (state == INIT) {
         // We have status for all members so we can make join descisions.
+        initMap.checkConsistent();
         elders = initMap.getElders();
         QPID_LOG(debug, *this << " elders: " << elders);
         if (!elders.empty()) { // I'm not the elder, I don't handle links & replication.
@@ -611,17 +612,8 @@ void Cluster::initMapCompleted(Lock& l) {
         else {
             QPID_LOG(info, this << " active for links.");
         }
-        // Check that cluster ID matches persistent store.
-        Uuid agreedId = initMap.getClusterId();
-        if (store.hasStore()) {
-            Uuid storeId = store.getClusterId();
-            if (storeId && storeId != agreedId)
-                throw Exception(
-                    QPID_MSG("Persistent cluster-id " << storeId 
-                             << " doesn't match cluster " << agreedId));
-            store.dirty(agreedId);
-        }
-        setClusterId(agreedId, l);
+        setClusterId(initMap.getClusterId(), l);
+        if (store.hasStore()) store.dirty(clusterId);
 
         if (initMap.isUpdateNeeded())  { // Joining established cluster.
             broker.setRecovery(false); // Ditch my current store.
@@ -822,13 +814,13 @@ void Cluster::checkUpdateIn(Lock&) {
         mcast.mcastControl(ClusterReadyBody(ProtocolVersion(), myUrl.str()), self);
         state = CATCHUP;
         discarding = false;     // ok to set, we're stalled for update.
-        QPID_LOG(notice, *this << " update complete, starting catch-up, members: " << map);
+        QPID_LOG(notice, *this << " update complete, starting catch-up.");
         deliverEventQueue.start();
     }
     else if (updateRetracted) { // Update was retracted, request another update
         updateRetracted = false;
         state = JOINER;
-        QPID_LOG(notice, *this << " update retracted, sending new update request");
+        QPID_LOG(notice, *this << " update retracted, sending new update request.");
         mcast.mcastControl(ClusterUpdateRequestBody(ProtocolVersion(), myUrl.str()), self);
         deliverEventQueue.start();
     }
@@ -853,10 +845,9 @@ void Cluster::updateOutError(const std::exception& e)  {
     updateOutDone(l);
 }
 
-void Cluster ::shutdown(const MemberId& , Lock& l) {
+void Cluster ::shutdown(const MemberId& , const Uuid& id, Lock& l) {
     QPID_LOG(notice, *this << " cluster shut down by administrator.");
-    // FIXME aconway 2009-11-20: incorrect! Need to pass UUID on shutdown command.
-    if (store.hasStore()) store.clean(Uuid(true));
+    if (store.hasStore()) store.clean(Uuid(id));
     leave(l);
 }
 
@@ -885,13 +876,13 @@ Manageable::status_t Cluster::ManagementMethod (uint32_t methodId, Args& args, s
 }
 
 void Cluster::stopClusterNode(Lock& l) {
-    QPID_LOG(notice, *this << " stopped by admin");
+    QPID_LOG(notice, *this << " cluster member stopped by administrator.");
     leave(l);
 }
 
 void Cluster::stopFullCluster(Lock& ) {
     QPID_LOG(notice, *this << " shutting down cluster " << name);
-    mcast.mcastControl(ClusterShutdownBody(), self);
+    mcast.mcastControl(ClusterShutdownBody(ProtocolVersion(), Uuid(true)), self);
 }
 
 void Cluster::memberUpdate(Lock& l) {
diff --git a/cpp/src/qpid/cluster/Cluster.h b/cpp/src/qpid/cluster/Cluster.h
index 0f931bbe29..7872588307 100644
--- a/cpp/src/qpid/cluster/Cluster.h
+++ b/cpp/src/qpid/cluster/Cluster.h
@@ -160,7 +160,7 @@ class Cluster : private Cpg::Handler, public management::Manageable {
     void messageExpired(const MemberId&, uint64_t, Lock& l);
     void errorCheck(const MemberId&, uint8_t type, SequenceNumber frameSeq, Lock&);
 
-    void shutdown(const MemberId&, Lock&);
+    void shutdown(const MemberId&, const framing::Uuid& shutdownId, Lock&);
 
     // Helper functions
     ConnectionPtr getConnection(const EventFrame&, Lock&);
diff --git a/cpp/src/qpid/cluster/InitialStatusMap.cpp b/cpp/src/qpid/cluster/InitialStatusMap.cpp
index 51d6140008..a5618db3e6 100644
--- a/cpp/src/qpid/cluster/InitialStatusMap.cpp
+++ b/cpp/src/qpid/cluster/InitialStatusMap.cpp
@@ -29,6 +29,7 @@ namespace cluster {
 using namespace std;
 using namespace boost;
 using namespace framing::cluster;
+using namespace framing;
 
 InitialStatusMap::InitialStatusMap(const MemberId& self_, size_t size_)
     : self(self_), completed(), resendNeeded(), size(size_)
@@ -106,7 +107,6 @@ bool InitialStatusMap::hasStore(const Map::value_type& v) {
 }
 
 bool InitialStatusMap::isUpdateNeeded() {
-    // FIXME aconway 2009-11-20: consistency checks isComplete or here?
     assert(isComplete());
     // We need an update if there are any active members.
     if (find_if(map.begin(), map.end(), &isActive) != map.end()) return true;
@@ -145,7 +145,43 @@ framing::Uuid InitialStatusMap::getClusterId() {
     if (i != map.end())
         return i->second->getClusterId(); // An active member
     else
-        return map.begin()->second->getClusterId();
+        return map.begin()->second->getClusterId(); // Youngest newcomer in node-id order
 }
 
+void InitialStatusMap::checkConsistent() {
+    assert(isComplete());
+    bool persistent = (map.begin()->second->getStoreState() != STORE_STATE_NO_STORE);
+    Uuid clusterId;
+    for (Map::iterator i = map.begin(); i != map.end(); ++i) {
+        // Must not mix transient and persistent members.
+        if (persistent != (i->second->getStoreState() != STORE_STATE_NO_STORE))
+            throw Exception("Mixing transient and persistent brokers in a cluster");
+        // Members with non-empty stores must have same cluster-id
+        switch (i->second->getStoreState()) {
+          case STORE_STATE_NO_STORE:
+          case STORE_STATE_EMPTY_STORE:
+            break;               
+          case STORE_STATE_DIRTY_STORE:
+          case STORE_STATE_CLEAN_STORE:
+            if (!clusterId) clusterId = i->second->getClusterId();
+            assert(clusterId);
+            if (clusterId != i->second->getClusterId())
+                throw Exception("Cluster-id mismatch, brokers belonged to different clusters.");
+        }
+    }
+    // If this is a newly forming cluster, clean stores must have same shutdown-id
+    if (find_if(map.begin(), map.end(), &isActive) == map.end()) {
+        Uuid shutdownId;
+        for (Map::iterator i = map.begin(); i != map.end(); ++i) {
+            if (i->second->getStoreState() == STORE_STATE_CLEAN_STORE) {
+                if (!shutdownId) shutdownId = i->second->getShutdownId();
+                assert(shutdownId);
+                if (shutdownId != i->second->getShutdownId())
+                    throw Exception("Shutdown-id mismatch, brokers were not shut down together.");
+            }
+        }
+    }
+}
+
+
 }} // namespace qpid::cluster
diff --git a/cpp/src/qpid/cluster/InitialStatusMap.h b/cpp/src/qpid/cluster/InitialStatusMap.h
index 72963ea2bb..40fd9ee49d 100644
--- a/cpp/src/qpid/cluster/InitialStatusMap.h
+++ b/cpp/src/qpid/cluster/InitialStatusMap.h
@@ -56,13 +56,14 @@ class InitialStatusMap
     bool isUpdateNeeded();
     /**@pre isComplete(). @return Cluster-wide cluster ID. */
     framing::Uuid getClusterId();
+    /**@pre isComplete(). @throw Exception if there are any inconsistencies. */
+    void checkConsistent();
 
   private:
     typedef std::map<MemberId, boost::optional<Status> > Map;
     static bool notInitialized(const Map::value_type&);
     static bool isActive(const Map::value_type&);
     static bool hasStore(const Map::value_type&);
-    void check();
     Map map;
     MemberSet firstConfig;
     MemberId self;
diff --git a/cpp/src/qpid/cluster/StoreStatus.cpp b/cpp/src/qpid/cluster/StoreStatus.cpp
index 3602ec9188..a7da3baa50 100644
--- a/cpp/src/qpid/cluster/StoreStatus.cpp
+++ b/cpp/src/qpid/cluster/StoreStatus.cpp
@@ -85,8 +85,6 @@ void StoreStatus::dirty(const Uuid& clusterId_) {
 }
 
 void StoreStatus::clean(const Uuid& shutdownId_) {
-    assert(clusterId);              // FIXME aconway 2009-11-20: throw exception
-    assert(shutdownId_);
     state = STORE_STATE_CLEAN_STORE;
     shutdownId = shutdownId_;
     save();
diff --git a/cpp/src/qpid/cluster/StoreStatus.h b/cpp/src/qpid/cluster/StoreStatus.h
index ead30b8fb8..539f46c10b 100644
--- a/cpp/src/qpid/cluster/StoreStatus.h
+++ b/cpp/src/qpid/cluster/StoreStatus.h
@@ -50,7 +50,6 @@ class StoreStatus
     void save();
 
     bool hasStore() { return state != framing::cluster::STORE_STATE_NO_STORE; }
-    bool isEmpty() { return state != framing::cluster::STORE_STATE_EMPTY_STORE; }
 
   private:
     framing::cluster::StoreState state;
diff --git a/cpp/src/tests/cluster_tests.py b/cpp/src/tests/cluster_tests.py
index 1437c9e20a..78967196a9 100755
--- a/cpp/src/tests/cluster_tests.py
+++ b/cpp/src/tests/cluster_tests.py
@@ -122,9 +122,9 @@ class StoreTests(BrokerTest):
     def test_persistent_restart(self):
         """Verify persistent cluster shutdown/restart scenarios"""
         cluster = self.cluster(0, args=self.args() + ["--cluster-size=3"])
-        a = cluster.start("a", expect=EXPECT_EXIT_OK, wait_for_start=False)
-        b = cluster.start("b", expect=EXPECT_EXIT_OK, wait_for_start=False)
-        c = cluster.start("c", expect=EXPECT_EXIT_FAIL, wait_for_start=True)
+        a = cluster.start("a", expect=EXPECT_EXIT_OK, wait=False)
+        b = cluster.start("b", expect=EXPECT_EXIT_OK, wait=False)
+        c = cluster.start("c", expect=EXPECT_EXIT_FAIL, wait=True)
         a.send_message("q", Message("1", durable=True))
         # Kill & restart one member.
         c.kill()
@@ -135,30 +135,30 @@ class StoreTests(BrokerTest):
         # Shut down the entire cluster cleanly and bring it back up
         a.send_message("q", Message("3", durable=True))
         qpid_cluster.main(["qpid-cluster", "-kf", a.host_port()])      
-        a = cluster.start("a", wait_for_start=False)
-        b = cluster.start("b", wait_for_start=False)
-        c = cluster.start("c", wait_for_start=True)
+        a = cluster.start("a", wait=False)
+        b = cluster.start("b", wait=False)
+        c = cluster.start("c", wait=True)
         self.assertEqual(a.get_message("q").content, "3")
 
     def test_persistent_partial_failure(self):
         # Kill 2 members, shut down the last cleanly then restart
         # Ensure we use the clean database
         cluster = self.cluster(0, args=self.args() + ["--cluster-size=3"])
-        a = cluster.start("a", expect=EXPECT_EXIT_FAIL, wait_for_start=False)
-        b = cluster.start("b", expect=EXPECT_EXIT_FAIL, wait_for_start=False)
-        c = cluster.start("c", expect=EXPECT_EXIT_OK, wait_for_start=True)
+        a = cluster.start("a", expect=EXPECT_EXIT_FAIL, wait=False)
+        b = cluster.start("b", expect=EXPECT_EXIT_FAIL, wait=False)
+        c = cluster.start("c", expect=EXPECT_EXIT_OK, wait=True)
         a.send_message("q", Message("4", durable=True))
         a.kill()
         b.kill()
         self.assertEqual(c.get_message("q").content, "4")
         c.send_message("q", Message("clean", durable=True))
         qpid_cluster.main(["qpid-cluster", "-kf", c.host_port()])              
-        a = cluster.start("a", wait_for_start=False)
-        b = cluster.start("b", wait_for_start=False)
-        c = cluster.start("c", wait_for_start=True)
+        a = cluster.start("a", wait=False)
+        b = cluster.start("b", wait=False)
+        c = cluster.start("c", wait=True)
         self.assertEqual(a.get_message("q").content, "clean")
         
-    def test_wrong_store_uuid(self):
+    def test_wrong_cluster_id(self):
         # Start a cluster1 broker, then try to restart in cluster2
         cluster1 = self.cluster(0, args=self.args())
         a = cluster1.start("a", expect=EXPECT_EXIT_OK)
@@ -168,4 +168,25 @@ class StoreTests(BrokerTest):
             a = cluster2.start("a", expect=EXPECT_EXIT_FAIL)
             self.fail("Expected exception")
         except: pass
-            
+
+    def test_wrong_shutdown_id(self):
+        # Start 2 members and shut down.
+        cluster = self.cluster(0, args=self.args()+["--cluster-size=2"])
+        a = cluster.start("a", expect=EXPECT_EXIT_OK, wait=False)
+        b = cluster.start("b", expect=EXPECT_EXIT_OK, wait=False)
+        self.assertEqual(0, qpid_cluster.main(["qpid_cluster", "-kf", a.host_port()]))
+        self.assertEqual(a.wait(), 0)
+        self.assertEqual(b.wait(), 0)
+
+        # Restart with a different member and shut down.
+        a = cluster.start("a", expect=EXPECT_EXIT_OK, wait=False)
+        c = cluster.start("c", expect=EXPECT_EXIT_OK, wait=False)
+        self.assertEqual(0, qpid_cluster.main(["qpid_cluster", "-kf", a.host_port()]))
+        self.assertEqual(a.wait(), 0)
+        self.assertEqual(c.wait(), 0)
+
+        # Mix members from both shutdown events, they should fail
+        a = cluster.start("a", expect=EXPECT_EXIT_FAIL, wait=False)
+        b = cluster.start("b", expect=EXPECT_EXIT_FAIL, wait=False)
+
+
diff --git a/cpp/xml/cluster.xml b/cpp/xml/cluster.xml
index 49a1ea3638..c0b7127f68 100644
--- a/cpp/xml/cluster.xml
+++ b/cpp/xml/cluster.xml
@@ -92,9 +92,11 @@
       <field name="type" type="error-type"/>
       <field name="frame-seq" type="sequence-no"/>
     </control>
-    
 
-    <control name="shutdown" code="0x20" label="Shut down entire cluster"/>
+    <!-- Shut down the entire cluster -->
+    <control name="shutdown" code="0x20">
+      <field name="shutdown-id" type="uuid"/>
+    </control>
 
   </class>