Map<Integer, AtomicInteger> invalidMutations = new HashMap<Integer, AtomicInteger>();
for (File file : clogs)
{
int bufferSize = (int) Math.min(Math.max(file.length(), 1), 32 * 1024 * 1024);
BufferedRandomAccessFile reader = new BufferedRandomAccessFile(new File(file.getAbsolutePath()), "r", bufferSize, true);
try
{
CommitLogHeader clHeader = null;
int replayPosition = 0;
String headerPath = CommitLogHeader.getHeaderPathFromSegmentPath(file.getAbsolutePath());
try
{
clHeader = CommitLogHeader.readCommitLogHeader(headerPath);
replayPosition = clHeader.getReplayPosition();
}
catch (IOException ioe)
{
logger.info(headerPath + " incomplete, missing or corrupt. Everything is ok, don't panic. CommitLog will be replayed from the beginning");
logger.debug("exception was", ioe);
}
if (replayPosition < 0)
{
logger.debug("skipping replay of fully-flushed {}", file);
continue;
}
reader.seek(replayPosition);
if (logger.isDebugEnabled())
logger.debug("Replaying " + file + " starting at " + reader.getFilePointer());
/* read the logs populate RowMutation and apply */
while (!reader.isEOF())
{
if (logger.isDebugEnabled())
logger.debug("Reading mutation at " + reader.getFilePointer());
long claimedCRC32;
Checksum checksum = new CRC32();
int serializedSize;
try
{
// any of the reads may hit EOF
serializedSize = reader.readInt();
// RowMutation must be at LEAST 10 bytes:
// 3 each for a non-empty Table and Key (including the 2-byte length from
// writeUTF/writeWithShortLength) and 4 bytes for column count.
// This prevents CRC by being fooled by special-case garbage in the file; see CASSANDRA-2128
if (serializedSize < 10)
break;
long claimedSizeChecksum = reader.readLong();
checksum.update(serializedSize);
if (checksum.getValue() != claimedSizeChecksum)
break; // entry wasn't synced correctly/fully. that's ok.
if (serializedSize > bytes.length)
bytes = new byte[(int) (1.2 * serializedSize)];
reader.readFully(bytes, 0, serializedSize);
claimedCRC32 = reader.readLong();
}
catch(EOFException eof)
{
break; // last CL entry didn't get completely written. that's ok.
}
checksum.update(bytes, 0, serializedSize);
if (claimedCRC32 != checksum.getValue())
{
// this entry must not have been fsynced. probably the rest is bad too,
// but just in case there is no harm in trying them (since we still read on an entry boundary)
continue;
}
/* deserialize the commit log entry */
ByteArrayInputStream bufIn = new ByteArrayInputStream(bytes, 0, serializedSize);
RowMutation rm = null;
try
{
rm = RowMutation.serializer().deserialize(new DataInputStream(bufIn));
}
catch (UnserializableColumnFamilyException ex)
{
AtomicInteger i = invalidMutations.get(ex.cfId);
if (i == null)
{
i = new AtomicInteger(1);
invalidMutations.put(ex.cfId, i);
}
else
i.incrementAndGet();
continue;
}
if (logger.isDebugEnabled())
logger.debug(String.format("replaying mutation for %s.%s: %s",
rm.getTable(),
ByteBufferUtil.bytesToHex(rm.key()),
"{" + StringUtils.join(rm.getColumnFamilies(), ", ") + "}"));
final Table table = Table.open(rm.getTable());
tablesRecovered.add(table);
final Collection<ColumnFamily> columnFamilies = new ArrayList<ColumnFamily>(rm.getColumnFamilies());
final long entryLocation = reader.getFilePointer();
final CommitLogHeader finalHeader = clHeader;
final RowMutation frm = rm;
Runnable runnable = new WrappedRunnable()
{
public void runMayThrow() throws IOException