mirror/mariadb
1
0
Fork 0
mirror of https://github.com/MariaDB/server.git synced 2025-01-22 06:44:16 +01:00
Code Issues Projects Releases Packages Wiki Activity Actions

re-enginered ndb restore to remove new/deletes and data copy

Browse source 
BitKeeper/deleted/.del-Makefile_old~5e1138bd59f6b3aa:
  Delete: ndb/src/kernel/blocks/backup/restore/Makefile_old
This commit is contained in:
unknown 2004-06-09 19:24:36 +00:00
parent c47ec30811
commit d089a424bd
12 changed files with 1831 additions and 1133 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
ndb/src/kernel/blocks/backup/restore/Makefile.am
View file

@ -1,7 +1,7 @@
ndbtools_PROGRAMS = ndb_restore
ndb_restore_SOURCES = main.cpp Restore.cpp
ndb_restore_SOURCES = main.cpp consumer.cpp consumer_restore.cpp consumer_printer.cpp Restore.cpp
LDADD_LOC = $(top_builddir)/ndb/src/libndbclient.la

20
ndb/src/kernel/blocks/backup/restore/Makefile_old
View file

@ -1,20 +0,0 @@
include .defs.mk
TYPE := *
BIN_TARGET := restore
BIN_TARGET_LIBS :=
BIN_TARGET_ARCHIVES := NDB_API
CCFLAGS_LOC = -I.. -I$(NDB_TOP)/src/ndbapi -I$(NDB_TOP)/include/ndbapi -I$(NDB_TOP)/include/util -I$(NDB_TOP)/include/portlib -I$(NDB_TOP)/include/kernel
#ifneq ($(MYSQLCLUSTER_TOP),)
#CCFLAGS_LOC +=-I$(MYSQLCLUSTER_TOP)/include -D USE_MYSQL
#LDFLAGS_LOC += -L$(MYSQLCLUSTER_TOP)/libmysql_r/ -lmysqlclient_r
#endif
SOURCES = main.cpp Restore.cpp
include $(NDB_TOP)/Epilogue.mk

307
ndb/src/kernel/blocks/backup/restore/Restore.cpp
View file

@ -33,32 +33,32 @@ Uint32 Twiddle32(Uint32 in); // Byte shift 32-bit data
Uint64 Twiddle64(Uint64 in); // Byte shift 64-bit data
bool
BackupFile::Twiddle(AttributeS* attr, Uint32 arraySize){
BackupFile::Twiddle(const AttributeDesc* attr_desc, AttributeData* attr_data, Uint32 arraySize){
if(m_hostByteOrder)
return true;
if(arraySize == 0){
arraySize = attr->Desc->arraySize;
arraySize = attr_desc->arraySize;
}
switch(attr->Desc->size){
switch(attr_desc->size){
case 8:
return true;
case 16:
for(unsigned i = 0; i<arraySize; i++){
attr->Data.u_int16_value[i] = Twiddle16(attr->Data.u_int16_value[i]);
attr_data->u_int16_value[i] = Twiddle16(attr_data->u_int16_value[i]);
}
return true;
case 32:
for(unsigned i = 0; i<arraySize; i++){
attr->Data.u_int32_value[i] = Twiddle32(attr->Data.u_int32_value[i]);
attr_data->u_int32_value[i] = Twiddle32(attr_data->u_int32_value[i]);
}
return true;
case 64:
for(unsigned i = 0; i<arraySize; i++){
attr->Data.u_int64_value[i] = Twiddle64(attr->Data.u_int64_value[i]);
attr_data->u_int64_value[i] = Twiddle64(attr_data->u_int64_value[i]);
}
return true;
default:
@ -208,7 +208,7 @@ TableS::TableS(NdbTableImpl* tableImpl)
m_dictTable = tableImpl;
m_noOfNullable = m_nullBitmaskSize = 0;
for (Uint32 i = 0; i < tableImpl->getNoOfColumns(); i++)
for (int i = 0; i < tableImpl->getNoOfColumns(); i++)
createAttr(tableImpl->getColumn(i));
}
@ -246,56 +246,112 @@ RestoreMetaData::parseTableDescriptor(const Uint32 * data, Uint32 len)
}
// Constructor
RestoreDataIterator::RestoreDataIterator(const RestoreMetaData & md)
: m_metaData(md)
RestoreDataIterator::RestoreDataIterator(const RestoreMetaData & md, void (* _free_data_callback)())
: m_metaData(md), free_data_callback(_free_data_callback)
{
debug << "RestoreDataIterator constructor" << endl;
setDataFile(md, 0);
m_buffer_sz = 64*1024;
m_buffer = malloc(m_buffer_sz);
m_buffer_ptr = m_buffer;
m_buffer_data_left = 0;
}
RestoreDataIterator::~RestoreDataIterator(){
RestoreDataIterator::~RestoreDataIterator()
{
if (m_buffer)
free(m_buffer);
}
TupleS & TupleS::operator=(const TupleS& tuple)
{
prepareRecord(*tuple.m_currentTable);
if (allAttrData) {
allAttrData= new AttributeData[getNoOfAttributes()];
memcpy(allAttrData, tuple.allAttrData, getNoOfAttributes()*sizeof(AttributeData));
}
return *this;
};
int TupleS::getNoOfAttributes() const {
if (m_currentTable == 0)
return 0;
return m_currentTable->getNoOfAttributes();
};
const TableS * TupleS::getTable() const {
return m_currentTable;
};
const AttributeDesc * TupleS::getDesc(int i) const {
return m_currentTable->allAttributesDesc[i];
}
AttributeData * TupleS::getData(int i) const{
return &(allAttrData[i]);
};
bool
TupleS::prepareRecord(const TableS & tab){
if (allAttrData) {
delete [] allAttrData;
m_currentTable= 0;
}
allAttrData = new AttributeData[tab.getNoOfAttributes()];
if (allAttrData == 0)
return false;
m_currentTable = &tab;
for(int i = 0; i<allAttributes.size(); i++) {
if(allAttributes[i] != NULL)
delete allAttributes[i];
}
allAttributes.clear();
AttributeS * a;
for(int i = 0; i<tab.getNoOfAttributes(); i++){
a = new AttributeS;
if(a == NULL) {
ndbout_c("Restore: Failed to allocate memory");
return false;
}
a->Desc = tab[i];
allAttributes.push_back(a);
}
return true;
}
const TupleS *
RestoreDataIterator::getNextTuple(int & res) {
TupleS * tup = new TupleS();
if(tup == NULL) {
ndbout_c("Restore: Failed to allocate memory");
res = -1;
return NULL;
}
if(!tup->prepareRecord(* m_currentTable)) {
res =-1;
return NULL;
}
Uint32 RestoreDataIterator::get_buffer_ptr(void **p_buf_ptr, Uint32 size, Uint32 nmemb, FILE *stream)
{
Uint32 sz = size*nmemb;
if (sz > m_buffer_data_left) {
if (free_data_callback)
(*free_data_callback)();
memcpy(m_buffer, m_buffer_ptr, m_buffer_data_left);
size_t r = fread(((char *)m_buffer) + m_buffer_data_left, 1, m_buffer_sz - m_buffer_data_left, stream);
m_buffer_data_left += r;
m_buffer_ptr = m_buffer;
if (sz > m_buffer_data_left)
sz = size * (m_buffer_data_left / size);
}
*p_buf_ptr = m_buffer_ptr;
m_buffer_ptr = ((char*)m_buffer_ptr)+sz;
m_buffer_data_left -= sz;
return sz/size;
}
Uint32 RestoreDataIterator::fread_buffer(void *ptr, Uint32 size, Uint32 nmemb, FILE *stream)
{
void *buf_ptr;
Uint32 r = get_buffer_ptr(&buf_ptr, size, nmemb, stream);
memcpy(ptr, buf_ptr, r*size);
return r;
}
const TupleS *
RestoreDataIterator::getNextTuple(int & res)
{
Uint32 dataLength = 0;
// Read record length
if (fread(&dataLength, sizeof(dataLength), 1, m_file) != 1){
if (fread_buffer(&dataLength, sizeof(dataLength), 1, m_file) != 1){
err << "getNextTuple:Error reading length of data part" << endl;
delete tup;
res = -1;
return NULL;
} // if
@ -309,34 +365,34 @@ RestoreDataIterator::getNextTuple(int & res) {
// End of this data fragment
debug << "End of fragment" << endl;
res = 0;
delete tup;
return NULL;
} // if
tup->createDataRecord(dataLenBytes);
// Read tuple data
if (fread(tup->getDataRecord(), 1, dataLenBytes, m_file) != dataLenBytes) {
void *_buf_ptr;
if (get_buffer_ptr(&_buf_ptr, 1, dataLenBytes, m_file) != dataLenBytes) {
err << "getNextTuple:Read error: " << endl;
delete tup;
res = -1;
return NULL;
}
Uint32 * ptr = tup->getDataRecord();
Uint32 *buf_ptr = (Uint32*)_buf_ptr, *ptr = buf_ptr;
ptr += m_currentTable->m_nullBitmaskSize;
for(int i = 0; i < m_currentTable->m_fixedKeys.size(); i++){
assert(ptr < tup->getDataRecord() + dataLength);
assert(ptr < buf_ptr + dataLength);
const Uint32 attrId = m_currentTable->m_fixedKeys[i]->attrId;
AttributeS * attr = tup->allAttributes[attrId];
const Uint32 sz = attr->Desc->getSizeInWords();
AttributeData * attr_data = m_tuple.getData(attrId);
const AttributeDesc * attr_desc = m_tuple.getDesc(attrId);
attr->Data.null = false;
attr->Data.void_value = ptr;
const Uint32 sz = attr_desc->getSizeInWords();
if(!Twiddle(attr))
attr_data->null = false;
attr_data->void_value = ptr;
if(!Twiddle(attr_desc, attr_data))
{
res = -1;
return NULL;
@ -345,17 +401,19 @@ RestoreDataIterator::getNextTuple(int & res) {
}
for(int i = 0; i<m_currentTable->m_fixedAttribs.size(); i++){
assert(ptr < tup->getDataRecord() + dataLength);
assert(ptr < buf_ptr + dataLength);
const Uint32 attrId = m_currentTable->m_fixedAttribs[i]->attrId;
AttributeS * attr = tup->allAttributes[attrId];
const Uint32 sz = attr->Desc->getSizeInWords();
AttributeData * attr_data = m_tuple.getData(attrId);
const AttributeDesc * attr_desc = m_tuple.getDesc(attrId);
attr->Data.null = false;
attr->Data.void_value = ptr;
const Uint32 sz = attr_desc->getSizeInWords();
if(!Twiddle(attr))
attr_data->null = false;
attr_data->void_value = ptr;
if(!Twiddle(attr_desc, attr_data))
{
res = -1;
return NULL;
@ -366,19 +424,21 @@ RestoreDataIterator::getNextTuple(int & res) {
for(int i = 0; i<m_currentTable->m_variableAttribs.size(); i++){
const Uint32 attrId = m_currentTable->m_variableAttribs[i]->attrId;
AttributeS * attr = tup->allAttributes[attrId];
AttributeData * attr_data = m_tuple.getData(attrId);
const AttributeDesc * attr_desc = m_tuple.getDesc(attrId);
if(attr->Desc->m_column->getNullable()){
const Uint32 ind = attr->Desc->m_nullBitIndex;
if(attr_desc->m_column->getNullable()){
const Uint32 ind = attr_desc->m_nullBitIndex;
if(BitmaskImpl::get(m_currentTable->m_nullBitmaskSize,
tup->getDataRecord(),ind)){
attr->Data.null = true;
attr->Data.void_value = NULL;
buf_ptr,ind)){
attr_data->null = true;
attr_data->void_value = NULL;
continue;
}
}
assert(ptr < tup->getDataRecord() + dataLength);
assert(ptr < buf_ptr + dataLength);
typedef BackupFormat::DataFile::VariableData VarData;
VarData * data = (VarData *)ptr;
@ -386,15 +446,15 @@ RestoreDataIterator::getNextTuple(int & res) {
Uint32 id = ntohl(data->Id);
assert(id == attrId);
attr->Data.null = false;
attr->Data.void_value = &data->Data[0];
attr_data->null = false;
attr_data->void_value = &data->Data[0];
/**
* Compute array size
*/
const Uint32 arraySize = (4 * sz) / (attr->Desc->size / 8);
assert(arraySize >= attr->Desc->arraySize);
if(!Twiddle(attr, attr->Desc->arraySize))
const Uint32 arraySize = (4 * sz) / (attr_desc->size / 8);
assert(arraySize >= attr_desc->arraySize);
if(!Twiddle(attr_desc, attr_data, attr_desc->arraySize))
{
res = -1;
return NULL;
@ -405,7 +465,7 @@ RestoreDataIterator::getNextTuple(int & res) {
m_count ++;
res = 0;
return tup;
return &m_tuple;
} // RestoreDataIterator::getNextTuple
BackupFile::BackupFile(){
@ -558,7 +618,7 @@ RestoreDataIterator::readFragmentHeader(int & ret)
debug << "RestoreDataIterator::getNextFragment" << endl;
if (fread(&Header, sizeof(Header), 1, m_file) != 1){
if (fread_buffer(&Header, sizeof(Header), 1, m_file) != 1){
ret = 0;
return false;
} // if
@ -581,6 +641,12 @@ RestoreDataIterator::readFragmentHeader(int & ret)
return false;
}
if(!m_tuple.prepareRecord(*m_currentTable))
{
ret =-1;
return false;
}
info << "_____________________________________________________" << endl
<< "Restoring data in table: " << m_currentTable->getTableName()
<< "(" << Header.TableId << ") fragment "
@ -588,6 +654,7 @@ RestoreDataIterator::readFragmentHeader(int & ret)
m_count = 0;
ret = 0;
return true;
} // RestoreDataIterator::getNextFragment
@ -596,7 +663,7 @@ bool
RestoreDataIterator::validateFragmentFooter() {
BackupFormat::DataFile::FragmentFooter footer;
if (fread(&footer, sizeof(footer), 1, m_file) != 1){
if (fread_buffer(&footer, sizeof(footer), 1, m_file) != 1){
err << "getFragmentFooter:Error reading fragment footer" << endl;
return false;
}
@ -787,14 +854,14 @@ RestoreLogIterator::getNextLogEntry(int & res) {
const Uint32 sz = ah->getDataSize();
if(sz == 0){
attr->Data.null = true;
attr->Data.void_value = NULL;
attr->Data->null = true;
attr->Data->void_value = NULL;
} else {
attr->Data.null = false;
attr->Data.void_value = ah->getDataPtr();
attr->Data->null = false;
attr->Data->void_value = ah->getDataPtr();
}
Twiddle(attr);
Twiddle(attr->Desc, attr->Data);
m_logEntry.m_values.push_back(attr);
ah = ah->getNext();
@ -804,3 +871,85 @@ RestoreLogIterator::getNextLogEntry(int & res) {
res = 0;
return &m_logEntry;
}
NdbOut &
operator<<(NdbOut& ndbout, const AttributeS& attr){
const AttributeData & data = *(attr.Data);
const AttributeDesc & desc = *(attr.Desc);
if (data.null)
{
ndbout << "<NULL>";
return ndbout;
}
NdbRecAttr tmprec;
tmprec.setup(desc.m_column, (char *)data.void_value);
ndbout << tmprec;
return ndbout;
}
// Print tuple data
NdbOut&
operator<<(NdbOut& ndbout, const TupleS& tuple)
{
ndbout << tuple.getTable()->getTableName() << "; ";
for (int i = 0; i < tuple.getNoOfAttributes(); i++)
{
AttributeData * attr_data = tuple.getData(i);
const AttributeDesc * attr_desc = tuple.getDesc(i);
const AttributeS attr = {attr_desc, attr_data};
debug << i << " " << attr_desc->m_column->getName();
ndbout << attr;
if (i != (tuple.getNoOfAttributes() - 1))
ndbout << delimiter << " ";
} // for
return ndbout;
}
// Print tuple data
NdbOut&
operator<<(NdbOut& ndbout, const LogEntry& logE)
{
switch(logE.m_type)
{
case LogEntry::LE_INSERT:
ndbout << "INSERT " << logE.m_table->getTableName() << " ";
break;
case LogEntry::LE_DELETE:
ndbout << "DELETE " << logE.m_table->getTableName() << " ";
break;
case LogEntry::LE_UPDATE:
ndbout << "UPDATE " << logE.m_table->getTableName() << " ";
break;
default:
ndbout << "Unknown log entry type (not insert, delete or update)" ;
}
for (int i = 0; i < logE.m_values.size();i++)
{
const AttributeS * attr = logE.m_values[i];
ndbout << attr->Desc->m_column->getName() << "=";
ndbout << (* attr);
if (i < (logE.m_values.size() - 1))
ndbout << ", ";
}
return ndbout;
}
NdbOut &
operator<<(NdbOut& ndbout, const TableS & table){
ndbout << endl << "Table: " << table.getTableName() << endl;
for (int j = 0; j < table.getNoOfAttributes(); j++)
{
const AttributeDesc * desc = table[j];
ndbout << desc->m_column->getName() << ": " << desc->m_column->getType();
ndbout << " key: " << desc->m_column->getPrimaryKey();
ndbout << " array: " << desc->arraySize;
ndbout << " size: " << desc->size << endl;
} // for
return ndbout;
}

52
ndb/src/kernel/blocks/backup/restore/Restore.hpp
View file

@ -18,6 +18,7 @@
#define RESTORE_H
#include <ndb_global.h>
#include <NdbOut.hpp>
#include <BackupFormat.hpp>
#include <NdbApi.hpp>
#include "myVector.hpp"
@ -25,6 +26,8 @@
#include <ndb_version.h>
#include <version.h>
static const char * delimiter = ";"; // Delimiter in file dump
const int FileNameLenC = 256;
const int TableNameLenC = 256;
const int AttrNameLenC = 256;
@ -82,26 +85,30 @@ public:
struct AttributeS {
const AttributeDesc * Desc;
AttributeData Data;
AttributeData * Data;
};
class TupleS {
private:
friend class RestoreDataIterator;
const TableS * m_currentTable;
myVector<AttributeS*> allAttributes;
Uint32 * dataRecord;
const TableS *m_currentTable;
AttributeData *allAttrData;
bool prepareRecord(const TableS &);
public:
TupleS() {dataRecord = NULL;};
~TupleS() {if(dataRecord != NULL) delete [] dataRecord;};
int getNoOfAttributes() const { return allAttributes.size(); };
const TableS * getTable() const { return m_currentTable;};
const AttributeS * operator[](int i) const { return allAttributes[i];};
Uint32 * getDataRecord() { return dataRecord;};
void createDataRecord(Uint32 bytes) { dataRecord = new Uint32[bytes];};
TupleS() {};
~TupleS()
{
if (allAttrData)
delete [] allAttrData;
};
TupleS(const TupleS& tuple); // disable copy constructor
TupleS & operator=(const TupleS& tuple);
int getNoOfAttributes() const;
const TableS * getTable() const;
const AttributeDesc * getDesc(int i) const;
AttributeData * getData(int i) const;
}; // class TupleS
class TableS {
@ -206,7 +213,7 @@ public:
const char * getFilename() const { return m_fileName;}
Uint32 getNodeId() const { return m_nodeId;}
const BackupFormat::FileHeader & getFileHeader() const { return m_fileHeader;}
bool Twiddle(AttributeS * attr, Uint32 arraySize = 0);
bool Twiddle(const AttributeDesc * attr_desc, AttributeData * attr_data, Uint32 arraySize = 0);
};
class RestoreMetaData : public BackupFile {
@ -243,20 +250,28 @@ public:
class RestoreDataIterator : public BackupFile {
const RestoreMetaData & m_metaData;
Uint32 m_count;
TupleS m_tuple;
const TableS* m_currentTable;
TupleS m_tuple;
void * m_buffer;
void * m_buffer_ptr;
Uint32 m_buffer_sz;
Uint32 m_buffer_data_left;
void (* free_data_callback)();
public:
// Constructor
RestoreDataIterator(const RestoreMetaData &);
RestoreDataIterator(const RestoreMetaData &, void (* free_data_callback)());
~RestoreDataIterator();
// Read data file fragment header
bool readFragmentHeader(int & res);
bool validateFragmentFooter();
Uint32 get_buffer_ptr(void **p_buf_ptr, Uint32 size, Uint32 nmemb, FILE *stream);
Uint32 fread_buffer(void *ptr, Uint32 size, Uint32 nmemb, FILE *stream);
const TupleS *getNextTuple(int & res);
};
@ -286,6 +301,11 @@ public:
const LogEntry * getNextLogEntry(int & res);
};
NdbOut& operator<<(NdbOut& ndbout, const TableS&);
NdbOut& operator<<(NdbOut& ndbout, const TupleS&);
NdbOut& operator<<(NdbOut& ndbout, const LogEntry&);
NdbOut& operator<<(NdbOut& ndbout, const RestoreMetaData&);
#endif

107
ndb/src/kernel/blocks/backup/restore/consumer.cpp Normal file
View file

@ -0,0 +1,107 @@
/* Copyright (C) 2003 MySQL AB
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
#include "consumer.hpp"
#ifdef USE_MYSQL
int
BackupConsumer::create_table_string(const TableS & table,
char * tableName,
char *buf){
int pos = 0;
int pos2 = 0;
char buf2[2048];
pos += sprintf(buf+pos, "%s%s", "CREATE TABLE ", tableName);
pos += sprintf(buf+pos, "%s", "(");
pos2 += sprintf(buf2+pos2, "%s", " primary key(");
for (int j = 0; j < table.getNoOfAttributes(); j++)
{
const AttributeDesc * desc = table[j];
// ndbout << desc->name << ": ";
pos += sprintf(buf+pos, "%s%s", desc->m_column->getName()," ");
switch(desc->m_column->getType()){
case NdbDictionary::Column::Int:
pos += sprintf(buf+pos, "%s", "int");
break;
case NdbDictionary::Column::Unsigned:
pos += sprintf(buf+pos, "%s", "int unsigned");
break;
case NdbDictionary::Column::Float:
pos += sprintf(buf+pos, "%s", "float");
break;
case NdbDictionary::Column::Decimal:
pos += sprintf(buf+pos, "%s", "decimal");
break;
case NdbDictionary::Column::Char:
pos += sprintf(buf+pos, "%s", "char");
break;
case NdbDictionary::Column::Varchar:
pos += sprintf(buf+pos, "%s", "varchar");
break;
case NdbDictionary::Column::Binary:
pos += sprintf(buf+pos, "%s", "binary");
break;
case NdbDictionary::Column::Varbinary:
pos += sprintf(buf+pos, "%s", "varchar binary");
break;
case NdbDictionary::Column::Bigint:
pos += sprintf(buf+pos, "%s", "bigint");
break;
case NdbDictionary::Column::Bigunsigned:
pos += sprintf(buf+pos, "%s", "bigint unsigned");
break;
case NdbDictionary::Column::Double:
pos += sprintf(buf+pos, "%s", "double");
break;
case NdbDictionary::Column::Datetime:
pos += sprintf(buf+pos, "%s", "datetime");
break;
case NdbDictionary::Column::Timespec:
pos += sprintf(buf+pos, "%s", "time");
break;
case NdbDictionary::Column::Undefined:
// pos += sprintf(buf+pos, "%s", "varchar binary");
return -1;
break;
default:
//pos += sprintf(buf+pos, "%s", "varchar binary");
return -1;
}
if (desc->arraySize > 1) {
int attrSize = desc->arraySize;
pos += sprintf(buf+pos, "%s%u%s",
"(",
attrSize,
")");
}
if (desc->m_column->getPrimaryKey()) {
pos += sprintf(buf+pos, "%s", " not null");
pos2 += sprintf(buf2+pos2, "%s%s", desc->m_column->getName(), ",");
}
pos += sprintf(buf+pos, "%s", ",");
} // for
pos2--; // remove trailing comma
pos2 += sprintf(buf2+pos2, "%s", ")");
// pos--; // remove trailing comma
pos += sprintf(buf+pos, "%s", buf2);
pos += sprintf(buf+pos, "%s", ") type=ndbcluster");
return 0;
}
#endif // USE_MYSQL

40
ndb/src/kernel/blocks/backup/restore/consumer.hpp Normal file
View file

@ -0,0 +1,40 @@
/* Copyright (C) 2003 MySQL AB
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
#ifndef CONSUMER_HPP
#define CONSUMER_HPP
#include "Restore.hpp"
class BackupConsumer {
public:
virtual bool init() { return true;}
virtual bool table(const TableS &){return true;}
#ifdef USE_MYSQL
virtual bool table(const TableS &, MYSQL* mysqlp) {return true;};
#endif
virtual void tuple(const TupleS &){}
virtual void tuple_free(){}
virtual void endOfTuples(){}
virtual void logEntry(const LogEntry &){}
virtual void endOfLogEntrys(){}
protected:
#ifdef USE_MYSQL
int create_table_string(const TableS & table, char * ,char *);
#endif
};
#endif

96
ndb/src/kernel/blocks/backup/restore/consumer_printer.cpp Normal file
View file

@ -0,0 +1,96 @@
/* Copyright (C) 2003 MySQL AB
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
#include "consumer_printer.hpp"
bool
BackupPrinter::table(const TableS & tab)
{
if (m_print || m_print_meta)
{
m_ndbout << tab;
ndbout_c("Successfully printed table: %s", tab.m_dictTable->getName());
}
return true;
}
#ifdef USE_MYSQL
bool
BackupPrinter::table(const TableS & tab, MYSQL * mysql)
{
if (m_print || m_print_meta)
{
char tmpTabName[MAX_TAB_NAME_SIZE*2];
sprintf(tmpTabName, "%s", tab.getTableName());
char * database = strtok(tmpTabName, "/");
char * schema = strtok( NULL , "/");
char * tableName = strtok( NULL , "/");
/**
* this means that the user did not specify schema
* and it is a v2x backup
*/
if(database == NULL)
return false;
if(schema == NULL)
return false;
if(tableName==NULL)
tableName = schema;
char stmtCreateDB[255];
sprintf(stmtCreateDB,"CREATE DATABASE %s", database);
ndbout_c("%s", stmtCreateDB);
char buf [2048];
create_table_string(tab, tableName, buf);
ndbout_c("%s", buf);
ndbout_c("Successfully printed table: %s", tab.m_dictTable->getName());
}
return true;
}
#endif
void
BackupPrinter::tuple(const TupleS & tup)
{
m_dataCount++;
if (m_print || m_print_data)
m_ndbout << tup << endl;
}
void
BackupPrinter::logEntry(const LogEntry & logE)
{
if (m_print || m_print_log)
m_ndbout << logE << endl;
m_logCount++;
}
void
BackupPrinter::endOfLogEntrys()
{
if (m_print || m_print_log)
{
ndbout << "Printed " << m_dataCount << " tuples and "
<< m_logCount << " log entries"
<< " to stdout." << endl;
}
}

50
ndb/src/kernel/blocks/backup/restore/consumer_printer.hpp Normal file
View file

@ -0,0 +1,50 @@
/* Copyright (C) 2003 MySQL AB
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
#ifndef CONSUMER_PRINTER_HPP
#define CONSUMER_PRINTER_HPP
#include "consumer.hpp"
class BackupPrinter : public BackupConsumer
{
NdbOut & m_ndbout;
public:
BackupPrinter(NdbOut & out = ndbout) : m_ndbout(out)
{
m_print = false;
m_print_log = false;
m_print_data = false;
m_print_meta = false;
}
virtual bool table(const TableS &);
#ifdef USE_MYSQL
virtual bool table(const TableS &, MYSQL* mysqlp);
#endif
virtual void tuple(const TupleS &);
virtual void logEntry(const LogEntry &);
virtual void endOfTuples() {};
virtual void endOfLogEntrys();
bool m_print;
bool m_print_log;
bool m_print_data;
bool m_print_meta;
Uint32 m_logCount;
Uint32 m_dataCount;
};
#endif

508
ndb/src/kernel/blocks/backup/restore/consumer_restore.cpp Normal file
View file

@ -0,0 +1,508 @@
/* Copyright (C) 2003 MySQL AB
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
#include "consumer_restore.hpp"
#include <NdbSleep.h>
extern FilteredNdbOut err;
extern FilteredNdbOut info;
extern FilteredNdbOut debug;
static void callback(int, NdbConnection*, void*);
bool
BackupRestore::init()
{
release();
if (!m_restore && !m_restore_meta)
return true;
m_ndb = new Ndb();
if (m_ndb == NULL)
return false;
// Turn off table name completion
m_ndb->useFullyQualifiedNames(false);
m_ndb->init(1024);
if (m_ndb->waitUntilReady(30) != 0)
{
err << "Failed to connect to ndb!!" << endl;
return false;
}
info << "Connected to ndb!!" << endl;
m_callback = new restore_callback_t[m_parallelism];
if (m_callback == 0)
{
err << "Failed to allocate callback structs" << endl;
return false;
}
m_tuples = new TupleS[m_parallelism];
if (m_tuples == 0)
{
err << "Failed to allocate tuples" << endl;
return false;
}
m_free_callback= m_callback;
for (Uint32 i= 0; i < m_parallelism; i++) {
m_callback[i].restore= this;
m_callback[i].connection= 0;
m_callback[i].tup= &m_tuples[i];
if (i > 0)
m_callback[i-1].next= &(m_callback[i]);
}
m_callback[m_parallelism-1].next = 0;
return true;
}
void BackupRestore::release()
{
if (m_ndb)
{
delete m_ndb;
m_ndb= 0;
}
if (m_callback)
{
delete [] m_callback;
m_callback= 0;
}
if (m_tuples)
{
delete [] m_tuples;
m_tuples= 0;
}
}
BackupRestore::~BackupRestore()
{
release();
}
bool
BackupRestore::table(const TableS & table){
if (!m_restore_meta)
{
return true;
}
NdbDictionary::Dictionary* dict = m_ndb->getDictionary();
if (dict->createTable(*table.m_dictTable) == -1)
{
err << "Create table " << table.getTableName() << " failed: "
<< dict->getNdbError() << endl;
return false;
}
info << "Successfully restored table " << table.getTableName()<< endl ;
return true;
}
void BackupRestore::tuple(const TupleS & tup)
{
if (!m_restore)
return;
restore_callback_t * cb = m_free_callback;
if (cb == 0)
assert(false);
m_free_callback = cb->next;
cb->retries = 0;
*(cb->tup) = tup; // must do copy!
tuple_a(cb);
if (m_free_callback == 0)
{
// send-poll all transactions
// close transaction is done in callback
m_ndb->sendPollNdb(3000, 1);
}
}
void BackupRestore::tuple_a(restore_callback_t *cb)
{
while (cb->retries < 10)
{
/**
* start transactions
*/
cb->connection = m_ndb->startTransaction();
if (cb->connection == NULL)
{
/*
if (errorHandler(cb))
{
continue;
}
*/
exitHandler();
} // if
const TupleS &tup = *(cb->tup);
const TableS * table = tup.getTable();
NdbOperation * op = cb->connection->getNdbOperation(table->getTableName());
if (op == NULL)
{
if (errorHandler(cb))
continue;
exitHandler();
} // if
if (op->writeTuple() == -1)
{
if (errorHandler(cb))
continue;
exitHandler();
} // if
int ret = 0;
for (int j = 0; j < 2; j++)
{
for (int i = 0; i < tup.getNoOfAttributes(); i++)
{
const AttributeDesc * attr_desc = tup.getDesc(i);
const AttributeData * attr_data = tup.getData(i);
int size = attr_desc->size;
int arraySize = attr_desc->arraySize;
char * dataPtr = attr_data->string_value;
Uint32 length = (size * arraySize) / 8;
if (attr_desc->m_column->getPrimaryKey())
{
if (j == 1) continue;
ret = op->equal(i, dataPtr, length);
}
else
{
if (j == 0) continue;
if (attr_data->null)
ret = op->setValue(i, NULL, 0);
else
ret = op->setValue(i, dataPtr, length);
}
if (ret < 0) {
ndbout_c("Column: %d type %d",i,
attr_desc->m_column->getType());
break;
}
}
if (ret < 0)
break;
}
if (ret < 0)
{
if (errorHandler(cb))
continue;
exitHandler();
}
// Prepare transaction (the transaction is NOT yet sent to NDB)
cb->connection->executeAsynchPrepare(Commit, &callback, cb);
m_transactions++;
return;
}
err << "Unable to recover from errors. Exiting..." << endl;
exitHandler();
}
void BackupRestore::cback(int result, restore_callback_t *cb)
{
m_transactions--;
if (result < 0)
{
/**
* Error. temporary or permanent?
*/
if (errorHandler(cb))
tuple_a(cb); // retry
else
{
err << "Restore: Failed to restore data due to a unrecoverable error. Exiting..." << endl;
exitHandler();
}
}
else
{
/**
* OK! close transaction
*/
m_ndb->closeTransaction(cb->connection);
cb->connection= 0;
cb->next= m_free_callback;
m_free_callback= cb;
m_dataCount++;
}
}
/**
* returns true if is recoverable,
* Error handling based on hugo
* false if it is an error that generates an abort.
*/
bool BackupRestore::errorHandler(restore_callback_t *cb)
{
NdbError error= cb->connection->getNdbError();
m_ndb->closeTransaction(cb->connection);
cb->connection= 0;
cb->retries++;
switch(error.status)
{
case NdbError::Success:
return false;
// ERROR!
break;
case NdbError::TemporaryError:
NdbSleep_MilliSleep(10);
return true;
// RETRY
break;
case NdbError::UnknownResult:
err << error << endl;
return false;
// ERROR!
break;
default:
case NdbError::PermanentError:
switch (error.code)
{
case 499:
case 250:
NdbSleep_MilliSleep(10);
return true; //temp errors?
default:
break;
}
//ERROR
err << error << endl;
return false;
break;
}
return false;
}
void BackupRestore::exitHandler()
{
release();
exit(-1);
}
void
BackupRestore::tuple_free()
{
if (!m_restore)
return;
// Send all transactions to NDB
if (m_transactions > 0)
m_ndb->sendPreparedTransactions(0);
// Poll all transactions
while (m_transactions > 0)
m_ndb->pollNdb(3000, m_transactions);
}
void
BackupRestore::endOfTuples()
{
tuple_free();
}
void
BackupRestore::logEntry(const LogEntry & tup)
{
if (!m_restore)
return;
NdbConnection * trans = m_ndb->startTransaction();
if (trans == NULL)
{
// Deep shit, TODO: handle the error
err << "Cannot start transaction" << endl;
exit(-1);
} // if
const TableS * table = tup.m_table;
NdbOperation * op = trans->getNdbOperation(table->getTableName());
if (op == NULL)
{
err << "Cannot get operation: " << trans->getNdbError() << endl;
exit(-1);
} // if
int check = 0;
switch(tup.m_type)
{
case LogEntry::LE_INSERT:
check = op->insertTuple();
break;
case LogEntry::LE_UPDATE:
check = op->updateTuple();
break;
case LogEntry::LE_DELETE:
check = op->deleteTuple();
break;
default:
err << "Log entry has wrong operation type."
<< " Exiting...";
exit(-1);
}
for (int i = 0; i < tup.m_values.size(); i++)
{
const AttributeS * attr = tup.m_values[i];
int size = attr->Desc->size;
int arraySize = attr->Desc->arraySize;
const char * dataPtr = attr->Data->string_value;
const Uint32 length = (size / 8) * arraySize;
if (attr->Desc->m_column->getPrimaryKey())
op->equal(attr->Desc->attrId, dataPtr, length);
else
op->setValue(attr->Desc->attrId, dataPtr, length);
}
#if 1
trans->execute(Commit);
#else
const int ret = trans->execute(Commit);
// Both insert update and delete can fail during log running
// and it's ok
if (ret != 0)
{
err << "execute failed: " << trans->getNdbError() << endl;
exit(-1);
}
#endif
m_ndb->closeTransaction(trans);
m_logCount++;
}
void
BackupRestore::endOfLogEntrys()
{
if (m_restore)
{
info << "Restored " << m_dataCount << " tuples and "
<< m_logCount << " log entries" << endl;
}
}
/*
* callback : This is called when the transaction is polled
*
* (This function must have three arguments:
* - The result of the transaction,
* - The NdbConnection object, and
* - A pointer to an arbitrary object.)
*/
static void
callback(int result, NdbConnection* trans, void* aObject)
{
restore_callback_t *cb = (restore_callback_t *)aObject;
(cb->restore)->cback(result, cb);
}
#if 0 // old tuple impl
void
BackupRestore::tuple(const TupleS & tup)
{
if (!m_restore)
return;
while (1)
{
NdbConnection * trans = m_ndb->startTransaction();
if (trans == NULL)
{
// Deep shit, TODO: handle the error
ndbout << "Cannot start transaction" << endl;
exit(-1);
} // if
const TableS * table = tup.getTable();
NdbOperation * op = trans->getNdbOperation(table->getTableName());
if (op == NULL)
{
ndbout << "Cannot get operation: ";
ndbout << trans->getNdbError() << endl;
exit(-1);
} // if
// TODO: check return value and handle error
if (op->writeTuple() == -1)
{
ndbout << "writeTuple call failed: ";
ndbout << trans->getNdbError() << endl;
exit(-1);
} // if
for (int i = 0; i < tup.getNoOfAttributes(); i++)
{
const AttributeS * attr = tup[i];
int size = attr->Desc->size;
int arraySize = attr->Desc->arraySize;
const char * dataPtr = attr->Data->string_value;
const Uint32 length = (size * arraySize) / 8;
if (attr->Desc->m_column->getPrimaryKey())
op->equal(i, dataPtr, length);
}
for (int i = 0; i < tup.getNoOfAttributes(); i++)
{
const AttributeS * attr = tup[i];
int size = attr->Desc->size;
int arraySize = attr->Desc->arraySize;
const char * dataPtr = attr->Data->string_value;
const Uint32 length = (size * arraySize) / 8;
if (!attr->Desc->m_column->getPrimaryKey())
if (attr->Data->null)
op->setValue(i, NULL, 0);
else
op->setValue(i, dataPtr, length);
}
int ret = trans->execute(Commit);
if (ret != 0)
{
ndbout << "execute failed: ";
ndbout << trans->getNdbError() << endl;
exit(-1);
}
m_ndb->closeTransaction(trans);
if (ret == 0)
break;
}
m_dataCount++;
}
#endif

79
ndb/src/kernel/blocks/backup/restore/consumer_restore.hpp Normal file
View file

@ -0,0 +1,79 @@
/* Copyright (C) 2003 MySQL AB
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
#ifndef CONSUMER_RESTORE_HPP
#define CONSUMER_RESTORE_HPP
#include "consumer.hpp"
struct restore_callback_t {
class BackupRestore *restore;
class TupleS *tup;
class NdbConnection *connection;
int retries;
restore_callback_t *next;
};
class BackupRestore : public BackupConsumer
{
public:
BackupRestore(Uint32 parallelism=1)
{
m_ndb = 0;
m_logCount = m_dataCount = 0;
m_restore = false;
m_restore_meta = false;
m_parallelism = parallelism;
m_callback = 0;
m_tuples = 0;
m_free_callback = 0;
m_transactions = 0;
}
virtual ~BackupRestore();
virtual bool init();
virtual void release();
virtual bool table(const TableS &);
#ifdef USE_MYSQL
virtual bool table(const TableS &, MYSQL* mysqlp);
#endif
virtual void tuple(const TupleS &);
virtual void tuple_free();
virtual void tuple_a(restore_callback_t *cb);
virtual void cback(int result, restore_callback_t *cb);
virtual bool errorHandler(restore_callback_t *cb);
virtual void exitHandler();
virtual void endOfTuples();
virtual void logEntry(const LogEntry &);
virtual void endOfLogEntrys();
void connectToMysql();
Ndb * m_ndb;
bool m_restore;
bool m_restore_meta;
Uint32 m_logCount;
Uint32 m_dataCount;
Uint32 m_parallelism;
Uint32 m_transactions;
TupleS *m_tuples;
restore_callback_t *m_callback;
restore_callback_t *m_free_callback;
};
#endif

652
ndb/src/kernel/blocks/backup/restore/consumer_restorem.cpp Normal file
View file

@ -0,0 +1,652 @@
/* Copyright (C) 2003 MySQL AB
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
#include "consumer_restore.hpp"
#include <NdbSleep.h>
extern FilteredNdbOut err;
extern FilteredNdbOut info;
extern FilteredNdbOut debug;
static bool asynchErrorHandler(NdbConnection * trans, Ndb * ndb);
static void callback(int result, NdbConnection* trans, void* aObject);
bool
BackupRestore::init()
{
if (!m_restore && !m_restore_meta)
return true;
m_ndb = new Ndb();
if (m_ndb == NULL)
return false;
// Turn off table name completion
m_ndb->useFullyQualifiedNames(false);
m_ndb->init(1024);
if (m_ndb->waitUntilReady(30) != 0)
{
ndbout << "Failed to connect to ndb!!" << endl;
return false;
}
ndbout << "Connected to ndb!!" << endl;
#if USE_MYSQL
if(use_mysql)
{
if ( mysql_thread_safe() == 0 )
{
ndbout << "Not thread safe mysql library..." << endl;
exit(-1);
}
ndbout << "Connecting to MySQL..." <<endl;
/**
* nwe param:
* port
* host
* user
*/
bool returnValue = true;
mysql_init(&mysql);
{
int portNo = 3306;
if ( mysql_real_connect(&mysql,
ga_host,
ga_user,
ga_password,
ga_database,
ga_port,
:: ga_socket,
0) == NULL )
{
ndbout_c("Connect failed: %s", mysql_error(&mysql));
returnValue = false;
}
ndbout << "Connected to MySQL!!!" <<endl;
}
/* if(returnValue){
mysql_set_server_option(&mysql, MYSQL_OPTION_MULTI_STATEMENTS_ON);
}
*/
return returnValue;
}
#endif
if (m_callback) {
delete [] m_callback;
m_callback = 0;
}
m_callback = new restore_callback_t[m_parallelism];
if (m_callback == 0)
{
ndbout << "Failed to allocate callback structs" << endl;
return false;
}
m_free_callback = m_callback;
for (int i= 0; i < m_parallelism; i++) {
m_callback[i].restore = this;
m_callback[i].connection = 0;
m_callback[i].retries = 0;
if (i > 0)
m_callback[i-1].next = &(m_callback[i]);
}
m_callback[m_parallelism-1].next = 0;
return true;
}
BackupRestore::~BackupRestore()
{
if (m_ndb != 0)
delete m_ndb;
if (m_callback)
delete [] m_callback;
}
#ifdef USE_MYSQL
bool
BackupRestore::table(const TableS & table, MYSQL * mysqlp){
if (!m_restore_meta)
{
return true;
}
char tmpTabName[MAX_TAB_NAME_SIZE*2];
sprintf(tmpTabName, "%s", table.getTableName());
char * database = strtok(tmpTabName, "/");
char * schema = strtok( NULL , "/");
char * tableName = strtok( NULL , "/");
/**
* this means that the user did not specify schema
* and it is a v2x backup
*/
if(database == NULL)
return false;
if(schema == NULL)
return false;
if(tableName==NULL)
tableName = schema;
char stmtCreateDB[255];
sprintf(stmtCreateDB,"CREATE DATABASE %s", database);
/*ignore return value. mysql_select_db will trap errors anyways*/
if (mysql_query(mysqlp,stmtCreateDB) == 0)
{
//ndbout_c("%s", stmtCreateDB);
}
if (mysql_select_db(&mysql, database) != 0)
{
ndbout_c("Error: %s", mysql_error(&mysql));
return false;
}
char buf [2048];
/**
* create table ddl
*/
if (create_table_string(table, tableName, buf))
{
ndbout_c("Unable to create a table definition since the "
"backup contains undefined types");
return false;
}
//ndbout_c("%s", buf);
if (mysql_query(mysqlp,buf) != 0)
{
ndbout_c("Error: %s", mysql_error(&mysql));
return false;
} else
{
ndbout_c("Successfully restored table %s into database %s", tableName, database);
}
return true;
}
#endif
bool
BackupRestore::table(const TableS & table){
if (!m_restore_meta)
{
return true;
}
NdbDictionary::Dictionary* dict = m_ndb->getDictionary();
if (dict->createTable(*table.m_dictTable) == -1)
{
err << "Create table " << table.getTableName() << " failed: "
<< dict->getNdbError() << endl;
return false;
}
info << "Successfully restored table " << table.getTableName()<< endl ;
return true;
}
void BackupRestore::tuple(const TupleS & tup)
{
if (!m_restore)
{
delete &tup;
return;
}
restore_callback_t * cb = m_free_callback;
if (cb)
{
m_free_callback = cb->next;
cb->retries = 0;
cb->tup = &tup;
tuple_a(cb);
}
if (m_free_callback == 0)
{
// send-poll all transactions
// close transaction is done in callback
m_ndb->sendPollNdb(3000, 1);
}
}
void BackupRestore::tuple_a(restore_callback_t *cb)
{
while (cb->retries < 10)
{
/**
* start transactions
*/
cb->connection = m_ndb->startTransaction();
if (cb->connection == NULL)
{
/*
if (asynchErrorHandler(cb->connection, m_ndb))
{
cb->retries++;
continue;
}
*/
asynchExitHandler();
} // if
const TupleS &tup = *(cb->tup);
const TableS * table = tup.getTable();
NdbOperation * op = cb->connection->getNdbOperation(table->getTableName());
if (op == NULL)
{
if (asynchErrorHandler(cb->connection, m_ndb))
{
cb->retries++;
continue;
}
asynchExitHandler();
} // if
if (op->writeTuple() == -1)
{
if (asynchErrorHandler(cb->connection, m_ndb))
{
cb->retries++;
continue;
}
asynchExitHandler();
} // if
Uint32 ret = 0;
for (int i = 0; i < tup.getNoOfAttributes(); i++)
{
const AttributeS * attr = tup[i];
int size = attr->Desc->size;
int arraySize = attr->Desc->arraySize;
char * dataPtr = attr->Data.string_value;
Uint32 length = (size * arraySize) / 8;
if (attr->Desc->m_column->getPrimaryKey())
{
ret = op->equal(i, dataPtr, length);
}
else
{
if (attr->Data.null)
ret = op->setValue(i, NULL, 0);
else
ret = op->setValue(i, dataPtr, length);
}
if (ret<0)
{
ndbout_c("Column: %d type %d",i,
tup.getTable()->m_dictTable->getColumn(i)->getType());
if (asynchErrorHandler(cb->connection, m_ndb))
{
cb->retries++;
break;
}
asynchExitHandler();
}
}
if (ret < 0)
continue;
// Prepare transaction (the transaction is NOT yet sent to NDB)
cb->connection->executeAsynchPrepare(Commit, &callback, cb);
m_transactions++;
}
ndbout_c("Unable to recover from errors. Exiting...");
asynchExitHandler();
}
void BackupRestore::cback(int result, restore_callback_t *cb)
{
if (result<0)
{
/**
* Error. temporary or permanent?
*/
if (asynchErrorHandler(cb->connection, m_ndb))
{
cb->retries++;
tuple_a(cb);
}
else
{
ndbout_c("Restore: Failed to restore data "
"due to a unrecoverable error. Exiting...");
delete m_ndb;
delete cb->tup;
exit(-1);
}
}
else
{
/**
* OK! close transaction
*/
m_ndb->closeTransaction(cb->connection);
delete cb->tup;
m_transactions--;
}
}
void BackupRestore::asynchExitHandler()
{
if (m_ndb != NULL)
delete m_ndb;
exit(-1);
}
#if 0 // old tuple impl
void
BackupRestore::tuple(const TupleS & tup)
{
if (!m_restore)
return;
while (1)
{
NdbConnection * trans = m_ndb->startTransaction();
if (trans == NULL)
{
// Deep shit, TODO: handle the error
ndbout << "Cannot start transaction" << endl;
exit(-1);
} // if
const TableS * table = tup.getTable();
NdbOperation * op = trans->getNdbOperation(table->getTableName());
if (op == NULL)
{
ndbout << "Cannot get operation: ";
ndbout << trans->getNdbError() << endl;
exit(-1);
} // if
// TODO: check return value and handle error
if (op->writeTuple() == -1)
{
ndbout << "writeTuple call failed: ";
ndbout << trans->getNdbError() << endl;
exit(-1);
} // if
for (int i = 0; i < tup.getNoOfAttributes(); i++)
{
const AttributeS * attr = tup[i];
int size = attr->Desc->size;
int arraySize = attr->Desc->arraySize;
const char * dataPtr = attr->Data.string_value;
const Uint32 length = (size * arraySize) / 8;
if (attr->Desc->m_column->getPrimaryKey())
op->equal(i, dataPtr, length);
}
for (int i = 0; i < tup.getNoOfAttributes(); i++)
{
const AttributeS * attr = tup[i];
int size = attr->Desc->size;
int arraySize = attr->Desc->arraySize;
const char * dataPtr = attr->Data.string_value;
const Uint32 length = (size * arraySize) / 8;
if (!attr->Desc->m_column->getPrimaryKey())
if (attr->Data.null)
op->setValue(i, NULL, 0);
else
op->setValue(i, dataPtr, length);
}
int ret = trans->execute(Commit);
if (ret != 0)
{
ndbout << "execute failed: ";
ndbout << trans->getNdbError() << endl;
exit(-1);
}
m_ndb->closeTransaction(trans);
if (ret == 0)
break;
}
m_dataCount++;
}
#endif
void
BackupRestore::endOfTuples()
{
if (!m_restore)
return;
// Send all transactions to NDB
m_ndb->sendPreparedTransactions(0);
// Poll all transactions
m_ndb->pollNdb(3000, m_transactions);
// Close all transactions
// for (int i = 0; i < nPreparedTransactions; i++)
// m_ndb->closeTransaction(asynchTrans[i]);
}
void
BackupRestore::logEntry(const LogEntry & tup)
{
if (!m_restore)
return;
NdbConnection * trans = m_ndb->startTransaction();
if (trans == NULL)
{
// Deep shit, TODO: handle the error
ndbout << "Cannot start transaction" << endl;
exit(-1);
} // if
const TableS * table = tup.m_table;
NdbOperation * op = trans->getNdbOperation(table->getTableName());
if (op == NULL)
{
ndbout << "Cannot get operation: ";
ndbout << trans->getNdbError() << endl;
exit(-1);
} // if
int check = 0;
switch(tup.m_type)
{
case LogEntry::LE_INSERT:
check = op->insertTuple();
break;
case LogEntry::LE_UPDATE:
check = op->updateTuple();
break;
case LogEntry::LE_DELETE:
check = op->deleteTuple();
break;
default:
ndbout << "Log entry has wrong operation type."
<< " Exiting...";
exit(-1);
}
for (int i = 0; i < tup.m_values.size(); i++)
{
const AttributeS * attr = tup.m_values[i];
int size = attr->Desc->size;
int arraySize = attr->Desc->arraySize;
const char * dataPtr = attr->Data.string_value;
const Uint32 length = (size / 8) * arraySize;
if (attr->Desc->m_column->getPrimaryKey())
op->equal(attr->Desc->attrId, dataPtr, length);
else
op->setValue(attr->Desc->attrId, dataPtr, length);
}
#if 1
trans->execute(Commit);
#else
const int ret = trans->execute(Commit);
// Both insert update and delete can fail during log running
// and it's ok
if (ret != 0)
{
ndbout << "execute failed: ";
ndbout << trans->getNdbError() << endl;
exit(-1);
}
#endif
m_ndb->closeTransaction(trans);
m_logCount++;
}
void
BackupRestore::endOfLogEntrys()
{
if (m_restore)
{
ndbout << "Restored " << m_dataCount << " tuples and "
<< m_logCount << " log entries" << endl;
}
}
#if 0
/*****************************************
*
* Callback function for asynchronous transactions
*
* Idea for error handling: Transaction objects have to be stored globally when
* they are prepared.
* In the callback function if the transaction:
* succeeded: delete the object from global storage
* failed but can be retried: execute the object that is in global storage
* failed but fatal: delete the object from global storage
*
******************************************/
static void restoreCallback(int result, // Result for transaction
NdbConnection *object, // Transaction object
void *anything) // Not used
{
static Uint32 counter = 0;
debug << "restoreCallback function called " << counter << " time(s)" << endl;
++counter;
if (result == -1)
{
ndbout << " restoreCallback (" << counter;
if ((counter % 10) == 1)
{
ndbout << "st";
} // if
else if ((counter % 10) == 2)
{
ndbout << "nd";
} // else if
else if ((counter % 10 ) ==3)
{
ndbout << "rd";
} // else if
else
{
ndbout << "th";
} // else
err << " time: error detected " << object->getNdbError() << endl;
} // if
} // restoreCallback
#endif
/*
* callback : This is called when the transaction is polled
*
* (This function must have three arguments:
* - The result of the transaction,
* - The NdbConnection object, and
* - A pointer to an arbitrary object.)
*/
static void
callback(int result, NdbConnection* trans, void* aObject)
{
restore_callback_t *cb = (restore_callback_t *)aObject;
(cb->restore)->cback(result, cb);
}
/**
* returns true if is recoverable,
* Error handling based on hugo
* false if it is an error that generates an abort.
*/
static
bool asynchErrorHandler(NdbConnection * trans, Ndb* ndb)
{
NdbError error = trans->getNdbError();
ndb->closeTransaction(trans);
switch(error.status)
{
case NdbError::Success:
return false;
// ERROR!
break;
case NdbError::TemporaryError:
NdbSleep_MilliSleep(10);
return true;
// RETRY
break;
case NdbError::UnknownResult:
ndbout << error << endl;
return false;
// ERROR!
break;
default:
case NdbError::PermanentError:
switch (error.code)
{
case 499:
case 250:
NdbSleep_MilliSleep(10);
return true; //temp errors?
default:
break;
}
//ERROR
ndbout << error << endl;
return false;
break;
}
return false;
}

1051
ndb/src/kernel/blocks/backup/restore/main.cpp
View file

File diff suppressed because it is too large Load diff