Merge lzhou@bk-internal.mysql.com:/home/bk/mysql-5.1-new-ndb

into  dev3-63.(none):/home/zhl/mysql/tree-merge/5.1/mysql-5.1-new-ndb

sql/log_event.cc

@@ -5717,9 +5717,26 @@ int Rows_log_event::exec_event(st_relay_log_info *rli)
     {
       if (!need_reopen)
       {
-        slave_print_msg(ERROR_LEVEL, rli, error,
-                        "Error in %s event: when locking tables",
-                        get_type_str());
+        if (thd->query_error || thd->is_fatal_error)
+        {
+          /*
+            Error reporting borrowed from Query_log_event with many excessive
+            simplifications (we don't honour --slave-skip-errors)
+          */
+          uint actual_error= thd->net.last_errno;
+          slave_print_msg(ERROR_LEVEL, rli, actual_error,
+                          "Error '%s' in %s event: when locking tables",
+                          (actual_error ? thd->net.last_error :
+                           "unexpected success or fatal error"),
+                          get_type_str());
+          thd->is_fatal_error= 1;
+        }
+        else
+        {
+          slave_print_msg(ERROR_LEVEL, rli, error,
+                         "Error in %s event: when locking tables",
+                         get_type_str());
+        }
         rli->clear_tables_to_lock();
         DBUG_RETURN(error);
       }

sql/protocol.h

@@ -58,6 +58,8 @@ class Protocol
   String *storage_packet() { return packet; }
   inline void free() { packet->free(); }
   virtual bool write();
+  inline  bool store(int from)
+  { return store_long((longlong) from); }
   inline  bool store(uint32 from)
   { return store_long((longlong) from); }
   inline  bool store(longlong from)

sql/slave.cc

@@ -1295,7 +1295,7 @@ bool show_master_info(THD* thd, MASTER_INFO* mi)
     rpl_filter->get_wild_ignore_table(&tmp);
     protocol->store(&tmp);
 
-    protocol->store((uint32) mi->rli.last_slave_errno);
+    protocol->store(mi->rli.last_slave_errno);
     protocol->store(mi->rli.last_slave_error, &my_charset_bin);
     protocol->store((uint32) mi->rli.slave_skip_counter);
     protocol->store((ulonglong) mi->rli.group_master_log_pos);

storage/ndb/src/common/debugger/EventLogger.cpp

@@ -16,6 +16,7 @@
 #include <ndb_global.h>
 
 #include <EventLogger.hpp>
+#include <TransporterCallback.hpp>
 
 #include <NdbConfig.h>
 #include <kernel/BlockNumbers.h>
@@ -526,11 +527,100 @@ void getTextUndoLogBlocked(QQQQ) {
 		       theData[1],
 		       theData[2]);
 }
+
 void getTextTransporterError(QQQQ) {
-  BaseString::snprintf(m_text, m_text_len, 
-		       "Transporter to node %d reported error 0x%x",
-		       theData[1],
-		       theData[2]);
+  struct myTransporterError{
+    int errorNum;
+    char  errorString[256];
+  };
+  int i = 0;
+  int lenth = 0;
+  static const struct myTransporterError TransporterErrorString[]=
+  {
+    //TE_NO_ERROR = 0
+    {TE_NO_ERROR,"No error"},
+    //TE_ERROR_CLOSING_SOCKET = 0x1
+    {TE_ERROR_CLOSING_SOCKET,"Error found during closing of socket"},
+    //TE_ERROR_IN_SELECT_BEFORE_ACCEPT = 0x2
+    {TE_ERROR_IN_SELECT_BEFORE_ACCEPT,"Error found before accept. The transporter will retry"},
+    //TE_INVALID_MESSAGE_LENGTH = 0x3 | TE_DO_DISCONNECT
+    {TE_INVALID_MESSAGE_LENGTH,"Error found in message (invalid message length)"},
+    //TE_INVALID_CHECKSUM = 0x4 | TE_DO_DISCONNECT
+    {TE_INVALID_CHECKSUM,"Error found in message (checksum)"},
+    //TE_COULD_NOT_CREATE_SOCKET = 0x5
+    {TE_COULD_NOT_CREATE_SOCKET,"Error found while creating socket(can't create socket)"},
+    //TE_COULD_NOT_BIND_SOCKET = 0x6
+    {TE_COULD_NOT_BIND_SOCKET,"Error found while binding server socket"},
+    //TE_LISTEN_FAILED = 0x7
+    {TE_LISTEN_FAILED,"Error found while listening to server socket"},
+    //TE_ACCEPT_RETURN_ERROR = 0x8
+    {TE_ACCEPT_RETURN_ERROR,"Error found during accept(accept return error)"},
+    //TE_SHM_DISCONNECT = 0xb | TE_DO_DISCONNECT
+    {TE_SHM_DISCONNECT,"The remote node has disconnected"},
+    //TE_SHM_IPC_STAT = 0xc | TE_DO_DISCONNECT
+    {TE_SHM_IPC_STAT,"Unable to check shm segment"},
+    //TE_SHM_UNABLE_TO_CREATE_SEGMENT = 0xd
+    {TE_SHM_UNABLE_TO_CREATE_SEGMENT,"Unable to create shm segment"},
+    //TE_SHM_UNABLE_TO_ATTACH_SEGMENT = 0xe
+    {TE_SHM_UNABLE_TO_ATTACH_SEGMENT,"Unable to attach shm segment"},
+    //TE_SHM_UNABLE_TO_REMOVE_SEGMENT = 0xf
+    {TE_SHM_UNABLE_TO_REMOVE_SEGMENT,"Unable to remove shm segment"},
+    //TE_TOO_SMALL_SIGID = 0x10
+    {TE_TOO_SMALL_SIGID,"Sig ID too small"},
+    //TE_TOO_LARGE_SIGID = 0x11
+    {TE_TOO_LARGE_SIGID,"Sig ID too large"},
+    //TE_WAIT_STACK_FULL = 0x12 | TE_DO_DISCONNECT
+    {TE_WAIT_STACK_FULL,"Wait stack was full"},
+    //TE_RECEIVE_BUFFER_FULL = 0x13 | TE_DO_DISCONNECT
+    {TE_RECEIVE_BUFFER_FULL,"Receive buffer was full"},
+    //TE_SIGNAL_LOST_SEND_BUFFER_FULL = 0x14 | TE_DO_DISCONNECT
+    {TE_SIGNAL_LOST_SEND_BUFFER_FULL,"Send buffer was full,and trying to force send fails"},
+    //TE_SIGNAL_LOST = 0x15
+    {TE_SIGNAL_LOST,"Send failed for unknown reason(signal lost)"},
+    //TE_SEND_BUFFER_FULL = 0x16
+    {TE_SEND_BUFFER_FULL,"The send buffer was full, but sleeping for a while solved"},
+    //TE_SCI_LINK_ERROR = 0x0017
+    {TE_SCI_LINK_ERROR,"There is no link from this node to the switch"},
+    //TE_SCI_UNABLE_TO_START_SEQUENCE = 0x18 | TE_DO_DISCONNECT
+    {TE_SCI_UNABLE_TO_START_SEQUENCE,"Could not start a sequence, because system resources are exumed or no sequence has been created"},
+    //TE_SCI_UNABLE_TO_REMOVE_SEQUENCE = 0x19 | TE_DO_DISCONNECT
+    {TE_SCI_UNABLE_TO_REMOVE_SEQUENCE,"Could not remove a sequence"},
+    //TE_SCI_UNABLE_TO_CREATE_SEQUENCE = 0x1a | TE_DO_DISCONNECT
+    {TE_SCI_UNABLE_TO_CREATE_SEQUENCE,"Could not create a sequence, because system resources are exempted. Must reboot"},
+    //TE_SCI_UNRECOVERABLE_DATA_TFX_ERROR = 0x1b | TE_DO_DISCONNECT
+    {TE_SCI_UNRECOVERABLE_DATA_TFX_ERROR,"Tried to send data on redundant link but failed"},
+    //TE_SCI_CANNOT_INIT_LOCALSEGMENT = 0x1c | TE_DO_DISCONNECT
+    {TE_SCI_CANNOT_INIT_LOCALSEGMENT,"Cannot initialize local segment"},
+    //TE_SCI_CANNOT_MAP_REMOTESEGMENT = 0x1d | TE_DO_DISCONNEC
+    {TE_SCI_CANNOT_MAP_REMOTESEGMENT,"Cannot map remote segment"},
+    //TE_SCI_UNABLE_TO_UNMAP_SEGMENT = 0x1e | TE_DO_DISCONNECT
+    {TE_SCI_UNABLE_TO_UNMAP_SEGMENT,"Cannot free the resources used by this segment (step 1)"},
+    //TE_SCI_UNABLE_TO_REMOVE_SEGMENT = 0x1f  | TE_DO_DISCONNEC
+    {TE_SCI_UNABLE_TO_REMOVE_SEGMENT,"Cannot free the resources used by this segment (step 2)"},
+    //TE_SCI_UNABLE_TO_DISCONNECT_SEGMENT = 0x20 | TE_DO_DISCONNECT
+    {TE_SCI_UNABLE_TO_DISCONNECT_SEGMENT,"Cannot disconnect from a remote segment"},
+    //TE_SHM_IPC_PERMANENT = 0x21
+    {TE_SHM_IPC_PERMANENT,"Shm ipc Permanent error"},
+    //TE_SCI_UNABLE_TO_CLOSE_CHANNEL = 0x22
+    {TE_SCI_UNABLE_TO_CLOSE_CHANNEL,"Unable to close the sci channel and the resources allocated"}
+  };
+
+  lenth = sizeof(TransporterErrorString)/sizeof(struct myTransporterError);
+  for(i=0; i<lenth; i++)
+  {
+    if(theData[2] == TransporterErrorString[i].errorNum)
+    {
+      BaseString::snprintf(m_text, m_text_len,
+                          "Transporter to node %d reported error: %s",
+                          theData[1],
+                          TransporterErrorString[i].errorString);
+      break;
+    }
+  }
+  if(i == lenth)
+    BaseString::snprintf(m_text, m_text_len,   
+                        "Transporter to node %d reported error: no such error",
+                          theData[1]);
 }
 void getTextTransporterWarning(QQQQ) {
   getTextTransporterError(m_text, m_text_len, theData);

storage/ndb/src/kernel/blocks/backup/Backup.cpp

@@ -2405,6 +2405,18 @@ Backup::defineBackupRef(Signal* signal, BackupRecordPtr ptr, Uint32 errCode)
   if(ptr.p->is_lcp()) 
   {
     jam();
+     if (ptr.p->ctlFilePtr == RNIL) {
+       ptr.p->m_gsn = GSN_DEFINE_BACKUP_REF;
+       ndbrequire(ptr.p->errorCode != 0);
+       DefineBackupRef* ref = (DefineBackupRef*)signal->getDataPtrSend();
+       ref->backupId = ptr.p->backupId;
+       ref->backupPtr = ptr.i;
+       ref->errorCode = ptr.p->errorCode;
+       ref->nodeId = getOwnNodeId();
+       sendSignal(ptr.p->masterRef, GSN_DEFINE_BACKUP_REF, signal,
+                  DefineBackupRef::SignalLength, JBB);
+       return;
+     }
 
     BackupFilePtr filePtr LINT_SET_PTR;
     ptr.p->files.getPtr(filePtr, ptr.p->ctlFilePtr);

storage/ndb/src/kernel/blocks/dblqh/DblqhMain.cpp

@@ -582,6 +582,26 @@ Dblqh::execDEFINE_BACKUP_REF(Signal* signal)
 {
   jamEntry();
   m_backup_ptr = RNIL;
+  DefineBackupRef* ref = (DefineBackupRef*)signal->getDataPtrSend();
+  int err_code = 0;
+  char * extra_msg = NULL;
+
+  switch(ref->errorCode){
+    case DefineBackupRef::Undefined:
+    case DefineBackupRef::FailedToSetupFsBuffers:
+    case DefineBackupRef::FailedToAllocateBuffers: 
+    case DefineBackupRef::FailedToAllocateTables: 
+    case DefineBackupRef::FailedAllocateTableMem: 
+    case DefineBackupRef::FailedToAllocateFileRecord:
+    case DefineBackupRef::FailedToAllocateAttributeRecord:
+    case DefineBackupRef::FailedInsertFileHeader: 
+    case DefineBackupRef::FailedInsertTableList: 
+      jam();
+      err_code = NDBD_EXIT_INVALID_CONFIG;
+      extra_msg = "Probably Backup parameters configuration error, Please consult the manual";
+      progError(__LINE__, err_code, extra_msg);
+  }
+
   sendsttorryLab(signal);
 }
 

storage/ndb/src/mgmclient/CommandInterpreter.cpp

@@ -2057,6 +2057,18 @@ CommandInterpreter::executeStatus(int processId,
     ndbout << processId << ": Node not found" << endl;
     return -1;
   }
+  if (cl->node_states[i].node_type != NDB_MGM_NODE_TYPE_NDB){
+    if (cl->node_states[i].version != 0){
+      ndbout << "Node "<< cl->node_states[i].node_id <<": connected" ;
+      ndbout_c(" (Version %d.%d.%d)",
+             getMajor(version) ,
+             getMinor(version),
+             getBuild(version));
+
+    }else
+     ndbout << "Node "<< cl->node_states[i].node_id <<": not connected" << endl;
+    return 0;
+  } 
   status = cl->node_states[i].node_status;
   startPhase = cl->node_states[i].start_phase;
   version = cl->node_states[i].version;

storage/ndb/src/mgmsrv/ConfigInfo.cpp

@@ -3615,6 +3615,7 @@ check_node_vs_replicas(Vector<ConfigInfo::ConfigRuleSection>&sections,
   Uint32 db_nodes= 0;
   Uint32 replicas= 0;
   Uint32 db_host_count= 0;
+  bool  with_arbitration_rank= false;
   ctx.m_userProperties.get(DB_TOKEN, &db_nodes);
   ctx.m_userProperties.get("NoOfReplicas", &replicas);
   if((db_nodes % replicas) != 0){
@@ -3650,83 +3651,90 @@ check_node_vs_replicas(Vector<ConfigInfo::ConfigRuleSection>&sections,
       tmp->get("HostName", &host);
 
       if (strcmp(type,DB_TOKEN) == 0)
-      {
-	{
-	  Uint32 ii;
-	  if (!p_db_hosts.get(host,&ii))
-	    db_host_count++;
-	  p_db_hosts.put(host,i);
-	  if (p_arbitrators.get(host,&ii))
-	  {
-	    arbitration_warning.appfmt(arbit_warn_fmt, ii, i, host);
-	    p_arbitrators.remove(host); // only one warning per db node
-	  }
-	}
-	{
-	  unsigned j;
-	  BaseString str, str2;
-	  str.assfmt("#group%d_",group);
-	  p_db_hosts.put(str.c_str(),i_group,host);
-	  str2.assfmt("##group%d_",group);
-	  p_db_hosts.put(str2.c_str(),i_group,i);
-	  for (j= 0; j < i_group; j++)
-	  {
-	    const char *other_host;
-	    p_db_hosts.get(str.c_str(),j,&other_host);
-	    if (strcmp(host,other_host) == 0) {
-	      unsigned int other_i, c= 0;
-	      p_db_hosts.get(str2.c_str(),j,&other_i);
-	      p_db_hosts.get(str.c_str(),&c);
-	      if (c == 0) // first warning in this node group
-		node_group_warning.appfmt("  Node group %d", group);
-	      c|= 1 << j;
-	      p_db_hosts.put(str.c_str(),c);
-
-	      node_group_warning.appfmt(",\n    db node with id %d and id %d "
-					"on same host %s", other_i, i, host);
-	    }
-	  }
-	  i_group++;
-	  DBUG_ASSERT(i_group <= replicas);
-	  if (i_group == replicas)
-	  {
-	    unsigned c= 0;
-	    p_db_hosts.get(str.c_str(),&c);
-	    if (c+1 == (1u << (replicas-1))) // all nodes on same machine
-	      node_group_warning.append(".\n    Host failure will "
-					"cause complete cluster shutdown.");
-	    else if (c > 0)
-	      node_group_warning.append(".\n    Host failure may "
-					"cause complete cluster shutdown.");
-	    group++;
-	    i_group= 0;
-	  }
-	}
+      { 
+        { 
+          Uint32 ii; 
+          if (!p_db_hosts.get(host,&ii)) 
+            db_host_count++; 
+          p_db_hosts.put(host,i); 
+          if (p_arbitrators.get(host,&ii)) 
+          { 
+            arbitration_warning.appfmt(arbit_warn_fmt, ii, i, host); 
+            p_arbitrators.remove(host); // only one warning per db node 
+          } 
+        } 
+        { 
+          unsigned j; 
+          BaseString str, str2; 
+          str.assfmt("#group%d_",group); 
+          p_db_hosts.put(str.c_str(),i_group,host); 
+          str2.assfmt("##group%d_",group); 
+          p_db_hosts.put(str2.c_str(),i_group,i); 
+          for (j= 0; j < i_group; j++) 
+          { 
+            const char *other_host; 
+            p_db_hosts.get(str.c_str(),j,&other_host); 
+            if (strcmp(host,other_host) == 0) { 
+              unsigned int other_i, c= 0; 
+              p_db_hosts.get(str2.c_str(),j,&other_i); 
+              p_db_hosts.get(str.c_str(),&c); 
+              if (c == 0) // first warning in this node group 
+                node_group_warning.appfmt("  Node group %d", group); 
+              c|= 1 << j; 
+              p_db_hosts.put(str.c_str(),c); 
+              node_group_warning.appfmt(",\n    db node with id %d and id %d " 
+              "on same host %s", other_i, i, host); 
+            } 
+          } 
+          i_group++; 
+          DBUG_ASSERT(i_group <= replicas); 
+          if (i_group == replicas) 
+          { 
+            unsigned c= 0; 
+            p_db_hosts.get(str.c_str(),&c); 
+            if (c+1 == (1u << (replicas-1))) // all nodes on same machine 
+              node_group_warning.append(".\n    Host failure will " 
+              "cause complete cluster shutdown."); 
+            else if (c > 0) 
+              node_group_warning.append(".\n    Host failure may " 
+              "cause complete cluster shutdown."); 
+            group++; 
+            i_group= 0; 
+          } 
+        }
       }
       else if (strcmp(type,API_TOKEN) == 0 ||
 	       strcmp(type,MGM_TOKEN) == 0)
-      {
-	Uint32 rank;
-	if(tmp->get("ArbitrationRank", &rank) && rank > 0)
-	{
-	  if(host && host[0] != 0)
-	  {
-	    Uint32 ii;
-	    p_arbitrators.put(host,i);
-	    if (p_db_hosts.get(host,&ii))
-	    {
-	      arbitration_warning.appfmt(arbit_warn_fmt, i, ii, host);
-	    }
-	  }
-	  else
-	  {
-	    arbitration_warning.appfmt(arbit_warn_fmt2, i);
-	  }
-	}
+      { 
+        Uint32 rank; 
+        if(tmp->get("ArbitrationRank", &rank) && rank > 0) 
+        { 
+          with_arbitration_rank = true;  //check whether MGM or API node configured with rank >0 
+          if(host && host[0] != 0) 
+          { 
+            Uint32 ii; 
+            p_arbitrators.put(host,i); 
+            if (p_db_hosts.get(host,&ii)) 
+            { 
+              arbitration_warning.appfmt(arbit_warn_fmt, i, ii, host); 
+            } 
+          } 
+          else 
+          { 
+            arbitration_warning.appfmt(arbit_warn_fmt2, i); 
+          } 
+        }
       }
     }
     if (db_host_count > 1 && node_group_warning.length() > 0)
       ndbout_c("Cluster configuration warning:\n%s",node_group_warning.c_str());
+    if (!with_arbitration_rank) 
+    {
+      ndbout_c("Cluster configuration warning:" 
+         "\n  Neither %s nor %s nodes are configured with arbitrator,"
+         "\n  may cause complete cluster shutdown in case of host failure.", 
+         MGM_TOKEN, API_TOKEN);
+    }
     if (db_host_count > 1 && arbitration_warning.length() > 0)
       ndbout_c("Cluster configuration warning:%s%s",arbitration_warning.c_str(),
 	       "\n  Running arbitrator on the same host as a database node may"

storage/ndb/src/mgmsrv/main.cpp

@@ -131,8 +131,6 @@ bool g_StopServer;
 bool g_RestartServer;
 extern EventLogger g_eventLogger;
 
-extern int global_mgmt_server_check;
-
 enum ndb_mgmd_options {
   OPT_INTERACTIVE = NDB_STD_OPTIONS_LAST,
   OPT_NO_NODEID_CHECKS,
@@ -207,8 +205,6 @@ int main(int argc, char** argv)
 start:
   glob= new MgmGlobals;
 
-  global_mgmt_server_check = 1;
-
   if (opt_interactive ||
       opt_non_interactive ||
       g_print_full_config) {

storage/ndb/src/ndbapi/ClusterMgr.cpp

@@ -360,8 +360,6 @@ ClusterMgr::execAPI_REGREQ(const Uint32 * theData){
   theFacade.sendSignalUnCond(&signal, nodeId);
 }
 
-int global_mgmt_server_check = 0; // set to one in mgmtsrvr main;
-
 void
 ClusterMgr::execAPI_REGCONF(const Uint32 * theData){
   const ApiRegConf * const apiRegConf = (ApiRegConf *)&theData[0];
@@ -379,7 +377,7 @@ ClusterMgr::execAPI_REGCONF(const Uint32 * theData){
 
   if(node.m_info.m_version != apiRegConf->version){
     node.m_info.m_version = apiRegConf->version;
-    if (global_mgmt_server_check == 1)
+    if(theNodes[theFacade.ownId()].m_info.m_type == NodeInfo::MGM)
       node.compatible = ndbCompatible_mgmt_ndb(NDB_VERSION,
 					       node.m_info.m_version);
     else