You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@iotdb.apache.org by GitBox <gi...@apache.org> on 2019/03/20 09:23:47 UTC
[GitHub] [incubator-iotdb] jt2594838 commented on a change in pull request
#101: [IOTDB-51]Reimplement post-back module,
rename it to synchronization module
jt2594838 commented on a change in pull request #101: [IOTDB-51]Reimplement post-back module, rename it to synchronization module
URL: https://github.com/apache/incubator-iotdb/pull/101#discussion_r267217901
##########
File path: iotdb/src/main/java/org/apache/iotdb/db/sync/receiver/ServerServiceImpl.java
##########
@@ -0,0 +1,740 @@
+/**
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements. See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership. The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing,
+ * software distributed under the License is distributed on an
+ * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+ * KIND, either express or implied. See the License for the
+ * specific language governing permissions and limitations
+ * under the License.
+ */
+package org.apache.iotdb.db.sync.receiver;
+
+import java.io.BufferedReader;
+import java.io.File;
+import java.io.FileInputStream;
+import java.io.FileNotFoundException;
+import java.io.FileOutputStream;
+import java.io.IOException;
+import java.math.BigInteger;
+import java.nio.ByteBuffer;
+import java.nio.channels.FileChannel;
+import java.security.MessageDigest;
+import java.util.ArrayList;
+import java.util.HashMap;
+import java.util.HashSet;
+import java.util.Iterator;
+import java.util.List;
+import java.util.Map;
+import java.util.Map.Entry;
+import java.util.Objects;
+import java.util.Set;
+import org.apache.iotdb.db.concurrent.ThreadName;
+import org.apache.iotdb.db.conf.IoTDBConfig;
+import org.apache.iotdb.db.conf.IoTDBDescriptor;
+import org.apache.iotdb.db.conf.directories.Directories;
+import org.apache.iotdb.db.engine.filenode.FileNodeManager;
+import org.apache.iotdb.db.engine.filenode.OverflowChangeType;
+import org.apache.iotdb.db.engine.filenode.TsFileResource;
+import org.apache.iotdb.db.exception.FileNodeManagerException;
+import org.apache.iotdb.db.exception.MetadataArgsErrorException;
+import org.apache.iotdb.db.exception.PathErrorException;
+import org.apache.iotdb.db.exception.ProcessorException;
+import org.apache.iotdb.db.metadata.MManager;
+import org.apache.iotdb.db.metadata.MetadataConstant;
+import org.apache.iotdb.db.metadata.MetadataOperationType;
+import org.apache.iotdb.db.qp.executor.OverflowQPExecutor;
+import org.apache.iotdb.db.qp.physical.crud.InsertPlan;
+import org.apache.iotdb.db.sync.conf.Constans;
+import org.apache.iotdb.db.utils.SyncUtils;
+import org.apache.iotdb.service.sync.thrift.SyncDataStatus;
+import org.apache.iotdb.service.sync.thrift.SyncService;
+import org.apache.iotdb.tsfile.file.metadata.ChunkGroupMetaData;
+import org.apache.iotdb.tsfile.file.metadata.ChunkMetaData;
+import org.apache.iotdb.tsfile.file.metadata.TsDeviceMetadata;
+import org.apache.iotdb.tsfile.file.metadata.TsDeviceMetadataIndex;
+import org.apache.iotdb.tsfile.file.metadata.enums.CompressionType;
+import org.apache.iotdb.tsfile.file.metadata.enums.TSDataType;
+import org.apache.iotdb.tsfile.file.metadata.enums.TSEncoding;
+import org.apache.iotdb.tsfile.read.ReadOnlyTsFile;
+import org.apache.iotdb.tsfile.read.TsFileSequenceReader;
+import org.apache.iotdb.tsfile.read.common.Field;
+import org.apache.iotdb.tsfile.read.common.Path;
+import org.apache.iotdb.tsfile.read.common.RowRecord;
+import org.apache.iotdb.tsfile.read.expression.QueryExpression;
+import org.apache.iotdb.tsfile.read.query.dataset.QueryDataSet;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+
+public class ServerServiceImpl implements SyncService.Iface {
+
+ private static final Logger logger = LoggerFactory.getLogger(ServerServiceImpl.class);
+
+ private static final FileNodeManager fileNodeManager = FileNodeManager.getInstance();
+ /**
+ * Metadata manager
+ **/
+ private static final MManager metadataManger = MManager.getInstance();
+
+ private static final String SYNC_SERVER = Constans.SYNC_SERVER;
+
+ private ThreadLocal<String> uuid = new ThreadLocal<>();
+ /**
+ * String means storage group,List means the set of new files(path) in local IoTDB and String
+ * means path of new Files
+ **/
+ private ThreadLocal<Map<String, List<String>>> fileNodeMap = new ThreadLocal<>();
+ /**
+ * Map String1 means timeseries String2 means path of new Files, long means startTime
+ **/
+ private ThreadLocal<Map<String, Map<String, Long>>> fileNodeStartTime = new ThreadLocal<>();
+ /**
+ * Map String1 means timeseries String2 means path of new Files, long means endTime
+ **/
+ private ThreadLocal<Map<String, Map<String, Long>>> fileNodeEndTime = new ThreadLocal<>();
+
+ /**
+ * Total num of files that needs to be loaded
+ */
+ private ThreadLocal<Integer> fileNum = new ThreadLocal<>();
+
+ /**
+ * IoTDB config
+ **/
+ private IoTDBConfig config = IoTDBDescriptor.getInstance().getConfig();
+
+ /**
+ * IoTDB data directory
+ **/
+ private String dataPath = config.getDataDir();
+
+ /**
+ * IoTDB multiple bufferWrite directory
+ **/
+ private String[] bufferWritePaths = config.getBufferWriteDirs();
+
+ /**
+ * The path to store metadata file of sender
+ */
+ private ThreadLocal<String> schemaFromSenderPath = new ThreadLocal<>();
+
+ /**
+ * Sync folder path of server
+ **/
+ private String syncFolderPath;
+
+ /**
+ * Sync data path of server
+ */
+ private String syncDataPath;
+
+ /**
+ * Init threadLocal variable
+ */
+ @Override
+ public void init(String storageGroup) {
+ if (logger.isInfoEnabled()) {
+ logger.info("Sync process starts to receive data of storage group {}", storageGroup);
+ }
+ fileNum.set(0);
+ fileNodeMap.set(new HashMap<>());
+ fileNodeStartTime.set(new HashMap<>());
+ fileNodeEndTime.set(new HashMap<>());
+ }
+
+ /**
+ * Verify IP address of sender
+ */
+ @Override
+ public boolean checkIdentity(String uuid, String ipAddress) {
+ Thread.currentThread().setName(ThreadName.SYNC_SERVER.getName());
+ this.uuid.set(uuid);
+ initPath();
+ return SyncUtils.verifyIPSegment(config.getIpWhiteList(), ipAddress);
+ }
+
+ /**
+ * Init file path and clear data if last sync process failed.
+ */
+ private void initPath() {
+ if (dataPath.length() > 0 && dataPath.charAt(dataPath.length() - 1) != File.separatorChar) {
+ dataPath = dataPath + File.separatorChar;
+ }
+ syncFolderPath = dataPath + SYNC_SERVER + File.separatorChar + this.uuid.get();
+ syncDataPath = syncFolderPath + File.separatorChar + Constans.DATA_SNAPSHOT_NAME;
+ schemaFromSenderPath
+ .set(syncFolderPath + File.separator + MetadataConstant.METADATA_LOG);
+ File syncFileDirectory = new File(syncFolderPath, this.uuid.get());
+ if (syncFileDirectory.exists()
+ && Objects.requireNonNull(syncFileDirectory.list()).length != 0) {
+ SyncUtils.deleteFile(syncFileDirectory);
+ }
+ for (String bufferWritePath : bufferWritePaths) {
+ if (bufferWritePath.length() > 0
+ && bufferWritePath.charAt(bufferWritePath.length() - 1) != File.separatorChar) {
+ bufferWritePath = bufferWritePath + File.separatorChar;
+ }
+ String backupPath = bufferWritePath + SYNC_SERVER + File.separator;
+ File backupDirectory = new File(backupPath, this.uuid.get());
+ if (backupDirectory.exists() && Objects.requireNonNull(backupDirectory.list()).length != 0) {
+ /** if does not exist, it means that the last time sync failed, clear data in the uuid directory and receive the data again **/
+ SyncUtils.deleteFile(backupDirectory);
+ }
+ }
+ }
+
+ /**
+ * Acquire schema from sender
+ *
+ * @param status: SUCCESS_STATUS or PROCESSING_STATUS. status = SUCCESS_STATUS : finish receiving
+ * schema file, start to sync schema. status = SUCCESS_STATUS : the schema file has not received
+ * completely.
+ */
+ @Override
+ public void syncSchema(ByteBuffer schema, SyncDataStatus status) {
+ if (status == SyncDataStatus.SUCCESS_STATUS) {
+ /** sync metadata, include storage group and timeseries **/
+ loadMetadata();
+ } else {
+ File file = new File(schemaFromSenderPath.get());
+ if (!file.getParentFile().exists()) {
+ try {
+ file.getParentFile().mkdirs();
+ if (!file.createNewFile()) {
+ logger.error("Cannot create file {}", file.getPath());
+ }
+ } catch (IOException e) {
+ logger.error("Cannot make schema file {}.", file.getPath(), e);
+ }
+ }
+ try (FileOutputStream fos = new FileOutputStream(file, true);
+ FileChannel channel = fos.getChannel()) {
+ channel.write(schema);
+ } catch (Exception e) {
+ logger.error("Cannot write data to file {}.", file.getPath(), e);
+ }
+ }
+
+ }
+
+ /**
+ * Load metadata from sender
+ */
+ private void loadMetadata() {
+ if (new File(schemaFromSenderPath.get()).exists()) {
+ try (BufferedReader br = new BufferedReader(
+ new java.io.FileReader(schemaFromSenderPath.get()))) {
+ String metadataOperation;
+ while ((metadataOperation = br.readLine()) != null) {
+ operation(metadataOperation);
+ }
+ } catch (FileNotFoundException e) {
+ logger.error("Cannot read the file {}.",
+ schemaFromSenderPath.get(), e);
+ } catch (IOException e) {
+ //TODO: how to deal with multiple insert schema
+ } catch (Exception e) {
+ logger.error("Parse metadata operation failed.", e);
+ }
+ }
+ }
+
+ /**
+ * Operate metadata operation in MManager
+ *
+ * @param cmd metadata operation
+ */
+ private void operation(String cmd)
+ throws PathErrorException, IOException, MetadataArgsErrorException {
+ String[] args = cmd.trim().split(",");
+ switch (args[0]) {
+ case MetadataOperationType.ADD_PATH_TO_MTREE:
+ Map<String, String> props = null;
+ String[] kv;
+ props = new HashMap<>(args.length - 5 + 1, 1);
+ for (int k = 5; k < args.length; k++) {
+ kv = args[k].split("=");
+ props.put(kv[0], kv[1]);
+ }
+ metadataManger.addPathToMTree(args[1], TSDataType.deserialize(Short.valueOf(args[2])),
+ TSEncoding.deserialize(Short.valueOf(args[3])),
+ CompressionType.deserialize(Short.valueOf(args[4])),
+ props);
+ break;
+ case MetadataOperationType.DELETE_PATH_FROM_MTREE:
+ metadataManger.deletePathFromMTree(args[1]);
+ break;
+ case MetadataOperationType.SET_STORAGE_LEVEL_TO_MTREE:
+ metadataManger.setStorageLevelToMTree(args[1]);
+ break;
+ case MetadataOperationType.ADD_A_PTREE:
+ metadataManger.addAPTree(args[1]);
+ break;
+ case MetadataOperationType.ADD_A_PATH_TO_PTREE:
+ metadataManger.addPathToPTree(args[1]);
+ break;
+ case MetadataOperationType.DELETE_PATH_FROM_PTREE:
+ metadataManger.deletePathFromPTree(args[1]);
+ break;
+ case MetadataOperationType.LINK_MNODE_TO_PTREE:
+ metadataManger.linkMNodeToPTree(args[1], args[2]);
+ break;
+ case MetadataOperationType.UNLINK_MNODE_FROM_PTREE:
+ metadataManger.unlinkMNodeFromPTree(args[1], args[2]);
+ break;
+ default:
+ logger.error("Unrecognizable command {}", cmd);
+ }
+ }
+
+ /**
+ * Start receiving tsfile from sender
+ *
+ * @param status status = SUCCESS_STATUS : finish receiving one tsfile status = PROCESSING_STATUS
+ * : tsfile has not received completely.
+ */
+ @Override
+ public String syncData(String md5OfSender, List<String> filePathSplit,
+ ByteBuffer dataToReceive, SyncDataStatus status) {
+ String md5OfReceiver = "";
+ StringBuilder filePathBuilder = new StringBuilder();
+ FileChannel channel;
+ /** Recombination File Path **/
+ for (int i = 0; i < filePathSplit.size(); i++) {
+ if (i == filePathSplit.size() - 1) {
+ filePathBuilder.append(filePathSplit.get(i));
+ } else {
+ filePathBuilder.append(filePathSplit.get(i)).append(File.separator);
+ }
+ }
+ String filePath = filePathBuilder.toString();
+ if (syncDataPath.length() > 0
+ && syncDataPath.charAt(syncDataPath.length() - 1) != File.separatorChar) {
+ syncDataPath = syncDataPath + File.separatorChar;
+ }
+ filePath = syncDataPath + filePath;
+ if (status == SyncDataStatus.PROCESSING_STATUS) { // there are still data stream to add
+ File file = new File(filePath);
+ if (!file.getParentFile().exists()) {
+ try {
+ file.getParentFile().mkdirs();
+ if (!file.createNewFile()) {
+ logger.error("cannot create file {}", file.getPath());
+ }
+ } catch (IOException e) {
+ logger.error("cannot make file {}", file.getPath(), e);
+ }
+ }
+ try (FileOutputStream fos = new FileOutputStream(file, true)) {// append new data
+ channel = fos.getChannel();
+ channel.write(dataToReceive);
+ } catch (IOException e) {
+ logger.error("cannot write data to file {}", file.getPath(), e);
+
+ }
+ } else { // all data in the same file has received successfully
+ try (FileInputStream fis = new FileInputStream(filePath)) {
+ MessageDigest md = MessageDigest.getInstance("MD5");
+ byte[] buffer = new byte[Constans.DATA_CHUNK_SIZE];
+ int n;
+ while ((n = fis.read(buffer)) != -1) {
+ md.update(buffer, 0, n);
+ }
+ md5OfReceiver = (new BigInteger(1, md.digest())).toString(16);
+ if (md5OfSender.equals(md5OfReceiver)) {
+ fileNum.set(fileNum.get() + 1);
+ if (logger.isInfoEnabled()) {
+ logger.info(String.format("Receiver has received %d files from sender", fileNum.get()));
+ }
+ } else {
+ if (!new File(filePath).delete()) {
+ logger.error("Receiver can not delete file {}", new File(filePath).getPath());
+ }
+ }
+ } catch (Exception e) {
+ logger.error("Receiver cannot generate md5", e);
+ }
+ }
+ return md5OfReceiver;
+ }
+
+
+ @Override
+ public boolean load() {
+ getFileNodeInfo();
+ loadData();
+ SyncUtils.deleteFile(new File(syncDataPath));
+ for (String bufferWritePath : bufferWritePaths) {
+ if (bufferWritePath.length() > 0
+ && bufferWritePath.charAt(bufferWritePath.length() - 1) != File.separatorChar) {
+ bufferWritePath = bufferWritePath + File.separatorChar;
+ }
+ String backupPath = bufferWritePath + SYNC_SERVER + File.separator;
+ File backupDirectory = new File(backupPath, this.uuid.get());
+ if (backupDirectory.exists() && Objects.requireNonNull(backupDirectory.list()).length != 0) {
+ SyncUtils.deleteFile(backupDirectory);
+ }
+ }
+ return true;
+ }
+
+ /**
+ * Get all tsfiles' info which are sent from sender, it is prepare for merging these data
+ */
+ public void getFileNodeInfo() {
+ File dataFileRoot = new File(syncDataPath);
+ File[] files = dataFileRoot.listFiles();
+ int processedNum = 0;
+ for (File storageGroupPB : files) {
+ List<String> filesPath = new ArrayList<>();
+ File[] filesSG = storageGroupPB.listFiles();
+ for (File fileTF : filesSG) { // fileTF means TsFiles
+ Map<String, Long> startTimeMap = new HashMap<>();
+ Map<String, Long> endTimeMap = new HashMap<>();
+ TsFileSequenceReader reader = null;
+ try {
+ reader = new TsFileSequenceReader(fileTF.getPath());
+ Map<String, TsDeviceMetadataIndex> deviceIdMap = reader.readFileMetadata().getDeviceMap();
+ Iterator<String> it = deviceIdMap.keySet().iterator();
+ while (it.hasNext()) {
+ String key = it.next();
+ TsDeviceMetadataIndex device = deviceIdMap.get(key);
+ startTimeMap.put(key, device.getStartTime());
+ endTimeMap.put(key, device.getEndTime());
+ }
+ } catch (Exception e) {
+ logger.error("Unable to read tsfile {}", fileTF.getPath(), e);
+ } finally {
+ try {
+ if (reader != null) {
+ reader.close();
+ }
+ } catch (IOException e) {
+ logger.error("Cannot close tsfile stream {}", fileTF.getPath(), e);
+ }
+ }
+ fileNodeStartTime.get().put(fileTF.getPath(), startTimeMap);
+ fileNodeEndTime.get().put(fileTF.getPath(), endTimeMap);
+ filesPath.add(fileTF.getPath());
+ processedNum++;
+ if (logger.isInfoEnabled()) {
+ logger.info(String
+ .format("Get tsfile info has complete : %d/%d", processedNum, fileNum.get()));
+ }
+ fileNodeMap.get().put(storageGroupPB.getName(), filesPath);
+ }
+ }
+ }
+
+
+ /**
+ * It is to merge data. If data in the tsfile is new, append the tsfile to the storage group
+ * directly. If data in the tsfile is old, it has two strategy to merge.It depends on the
+ * possibility of updating historical data.
+ */
+ public void loadData() {
+ if (syncDataPath.length() > 0
+ && syncDataPath.charAt(syncDataPath.length() - 1) != File.separatorChar) {
+ syncDataPath = syncDataPath + File.separatorChar;
+ }
+ int processedNum = 0;
+ for (String storageGroup : fileNodeMap.get().keySet()) {
+ List<String> filesPath = fileNodeMap.get().get(storageGroup);
+ /** before load external tsFile, it is necessary to order files in the same storage group **/
+ for (int i = 0; i < filesPath.size(); i++) {
+ for (int j = i + 1; j < filesPath.size(); j++) {
+ boolean swapOrNot = false;
+ Map<String, Long> startTimeI = fileNodeStartTime.get().get(filesPath.get(i));
+ Map<String, Long> endTimeI = fileNodeStartTime.get().get(filesPath.get(i));
+ Map<String, Long> startTimeJ = fileNodeStartTime.get().get(filesPath.get(j));
+ Map<String, Long> endTimeJ = fileNodeStartTime.get().get(filesPath.get(j));
+ for (String deviceId : endTimeI.keySet()) {
+ if (startTimeJ.containsKey(deviceId) && startTimeI.get(deviceId) >
+ endTimeJ.get(deviceId)) {
+ swapOrNot = true;
+ break;
+ }
+ }
+ if (swapOrNot) {
+ String temp = filesPath.get(i);
+ filesPath.set(i, filesPath.get(j));
+ filesPath.set(j, temp);
+ }
+ }
+ }
+
+ for (String path : filesPath) {
+ // get startTimeMap and endTimeMap
+ Map<String, Long> startTimeMap = fileNodeStartTime.get().get(path);
+ Map<String, Long> endTimeMap = fileNodeEndTime.get().get(path);
+
+ // create a new fileNode
+ String header = syncDataPath;
+ String relativePath = path.substring(header.length());
+ TsFileResource fileNode = new TsFileResource(startTimeMap, endTimeMap,
+ OverflowChangeType.NO_CHANGE,
+ Directories.getInstance().getNextFolderIndexForTsFile(), relativePath);
+ // call interface of load external file
+ try {
+ if (!fileNodeManager.appendFileToFileNode(storageGroup, fileNode, path)) {
+ // it is a file with overflow data
+ if (config.isUpdate_historical_data_possibility()) {
+ loadOldData(path);
+ } else {
+ List<String> overlapFiles = fileNodeManager.getOverlapFilesFromFileNode(
+ storageGroup,
+ fileNode, uuid.get());
+ if (overlapFiles.isEmpty()) {
+ loadOldData(path);
Review comment:
Maybe you can insert this new file between two existing files?
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
With regards,
Apache Git Services