[GitHub] [hive] aasha commented on a change in pull request #1358: HIVE-23955 : Classification of Error Codes in Replication
aasha commented on a change in pull request #1358: URL: https://github.com/apache/hive/pull/1358#discussion_r466579718 ## File path: ql/src/java/org/apache/hadoop/hive/ql/exec/repl/AtlasDumpTask.java ## @@ -196,12 +203,14 @@ private String checkHiveEntityGuid(AtlasRequestBuilder atlasRequestBuilder, Stri AtlasObjectId objectId = atlasRequestBuilder.getItemToExport(clusterName, srcDb); Set> entries = objectId.getUniqueAttributes().entrySet(); if (entries == null || entries.isEmpty()) { - throw new SemanticException("Could find entries in objectId for:" + clusterName); + throw new SemanticException(ErrorMsg.REPL_INVALID_CONFIG_FOR_SERVICE.format("Could find " + +"entries in objectId for:" + clusterName, "atlas")); Review comment: ok ok. yes will do that 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: us...@infra.apache.org - To unsubscribe, e-mail: gitbox-unsubscr...@hive.apache.org For additional commands, e-mail: gitbox-h...@hive.apache.org
[GitHub] [hive] aasha commented on a change in pull request #1358: HIVE-23955 : Classification of Error Codes in Replication
aasha commented on a change in pull request #1358: URL: https://github.com/apache/hive/pull/1358#discussion_r466574105 ## File path: ql/src/java/org/apache/hadoop/hive/ql/exec/repl/AtlasDumpTask.java ## @@ -132,31 +130,40 @@ private AtlasReplInfo createAtlasReplInfo() throws SemanticException, MalformedU private long lastStoredTimeStamp() throws SemanticException { Path prevMetadataPath = new Path(work.getPrevAtlasDumpDir(), EximUtil.METADATA_NAME); -BufferedReader br = null; +Retryable retryable = Retryable.builder() + .withHiveConf(conf) + .withRetryOnException(IOException.class) + .withFailOnException(FileNotFoundException.class).build(); try { - FileSystem fs = prevMetadataPath.getFileSystem(conf); - br = new BufferedReader(new InputStreamReader(fs.open(prevMetadataPath), Charset.defaultCharset())); - String line = br.readLine(); - if (line == null) { -throw new SemanticException("Could not read lastStoredTimeStamp from atlas metadata file"); - } - String[] lineContents = line.split("\t", 5); - return Long.parseLong(lineContents[1]); -} catch (Exception ex) { - throw new SemanticException(ex); -} finally { - if (br != null) { + return retryable.executeCallable(() -> { +BufferedReader br = null; try { - br.close(); -} catch (IOException e) { - throw new SemanticException(e); + FileSystem fs = prevMetadataPath.getFileSystem(conf); + br = new BufferedReader(new InputStreamReader(fs.open(prevMetadataPath), Charset.defaultCharset())); + String line = br.readLine(); + if (line == null) { +throw new SemanticException(ErrorMsg.REPL_INVALID_CONFIG_FOR_SERVICE Review comment: Named it as REPL_INVALID_INTERNAL_CONFIG_FOR_SERVICE 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: us...@infra.apache.org - To unsubscribe, e-mail: gitbox-unsubscr...@hive.apache.org For additional commands, e-mail: gitbox-h...@hive.apache.org
[GitHub] [hive] aasha commented on a change in pull request #1358: HIVE-23955 : Classification of Error Codes in Replication
aasha commented on a change in pull request #1358: URL: https://github.com/apache/hive/pull/1358#discussion_r466563797 ## File path: ql/src/java/org/apache/hadoop/hive/ql/exec/repl/AtlasDumpTask.java ## @@ -196,12 +203,14 @@ private String checkHiveEntityGuid(AtlasRequestBuilder atlasRequestBuilder, Stri AtlasObjectId objectId = atlasRequestBuilder.getItemToExport(clusterName, srcDb); Set> entries = objectId.getUniqueAttributes().entrySet(); if (entries == null || entries.isEmpty()) { - throw new SemanticException("Could find entries in objectId for:" + clusterName); + throw new SemanticException(ErrorMsg.REPL_INVALID_CONFIG_FOR_SERVICE.format("Could find " + +"entries in objectId for:" + clusterName, "atlas")); Review comment: Format will replace the config name and service name and helps us to reuse the same error code. 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: us...@infra.apache.org - To unsubscribe, e-mail: gitbox-unsubscr...@hive.apache.org For additional commands, e-mail: gitbox-h...@hive.apache.org
[GitHub] [hive] aasha commented on a change in pull request #1358: HIVE-23955 : Classification of Error Codes in Replication
aasha commented on a change in pull request #1358: URL: https://github.com/apache/hive/pull/1358#discussion_r466546396 ## File path: ql/src/java/org/apache/hadoop/hive/ql/exec/repl/DirCopyTask.java ## @@ -86,107 +87,59 @@ private boolean checkIfPathExist(Path sourcePath, UserGroupInformation proxyUser return proxyUser.doAs((PrivilegedExceptionAction) () -> sourcePath.getFileSystem(conf).exists(sourcePath)); } - private int handleException(Exception e, Path sourcePath, Path targetPath, - int currentRetry, UserGroupInformation proxyUser) { -try { - LOG.info("Checking if source path " + sourcePath + " is missing for exception ", e); - if (!checkIfPathExist(sourcePath, proxyUser)) { -LOG.info("Source path is missing. Ignoring exception."); -return 0; - } -} catch (Exception ex) { - LOG.warn("Source path missing check failed. ", ex); -} -// retry logic only for i/o exception -if (!(e instanceof IOException)) { - LOG.error("Unable to copy {} to {}", sourcePath, targetPath, e); - setException(e); - return ErrorMsg.getErrorMsg(e.getMessage()).getErrorCode(); -} - -if (currentRetry <= MAX_COPY_RETRY) { - LOG.warn("Unable to copy {} to {}", sourcePath, targetPath, e); -} else { - LOG.error("Unable to copy {} to {} even after retrying for {} time", sourcePath, targetPath, currentRetry, e); - setException(e); - return ErrorMsg.REPL_FILE_SYSTEM_OPERATION_RETRY.getErrorCode(); -} -int sleepTime = FileUtils.getSleepTime(currentRetry); -LOG.info("Sleep for " + sleepTime + " milliseconds before retry no " + (currentRetry)); -try { - Thread.sleep(sleepTime); -} catch (InterruptedException timerEx) { - LOG.info("Sleep interrupted", timerEx.getMessage()); -} -try { - if (proxyUser == null) { -proxyUser = Utils.getUGI(); - } - FileSystem.closeAllForUGI(proxyUser); -} catch (Exception ex) { - LOG.warn("Unable to closeAllForUGI for user " + proxyUser, ex); -} -return ErrorMsg.getErrorMsg(e.getMessage()).getErrorCode(); - } - @Override public int execute() { String distCpDoAsUser = conf.getVar(HiveConf.ConfVars.HIVE_DISTCP_DOAS_USER); +Retryable retryable = Retryable.builder() + .withHiveConf(conf) + .withRetryOnException(IOException.class).build(); +try { + return retryable.executeCallable(() -> { +UserGroupInformation proxyUser = null; +Path sourcePath = work.getFullyQualifiedSourcePath(); +Path targetPath = work.getFullyQualifiedTargetPath(); +try { + if (conf.getBoolVar(HiveConf.ConfVars.REPL_ADD_RAW_RESERVED_NAMESPACE)) { +sourcePath = reservedRawPath(work.getFullyQualifiedSourcePath().toUri()); +targetPath = reservedRawPath(work.getFullyQualifiedTargetPath().toUri()); + } + UserGroupInformation ugi = Utils.getUGI(); + String currentUser = ugi.getShortUserName(); + if (distCpDoAsUser != null && !currentUser.equals(distCpDoAsUser)) { +proxyUser = UserGroupInformation.createProxyUser( + distCpDoAsUser, UserGroupInformation.getLoginUser()); + } -Path sourcePath = work.getFullyQualifiedSourcePath(); -Path targetPath = work.getFullyQualifiedTargetPath(); -if (conf.getBoolVar(HiveConf.ConfVars.REPL_ADD_RAW_RESERVED_NAMESPACE)) { - sourcePath = reservedRawPath(work.getFullyQualifiedSourcePath().toUri()); - targetPath = reservedRawPath(work.getFullyQualifiedTargetPath().toUri()); -} -int currentRetry = 0; -int error = 0; -UserGroupInformation proxyUser = null; -while (currentRetry <= MAX_COPY_RETRY) { - try { -UserGroupInformation ugi = Utils.getUGI(); -String currentUser = ugi.getShortUserName(); -if (distCpDoAsUser != null && !currentUser.equals(distCpDoAsUser)) { - proxyUser = UserGroupInformation.createProxyUser( - distCpDoAsUser, UserGroupInformation.getLoginUser()); -} - -setTargetPathOwner(targetPath, sourcePath, proxyUser); - -// do we create a new conf and only here provide this additional option so that we get away from -// differences of data in two location for the same directories ? -// basically add distcp.options.delete to hiveconf new object ? -FileUtils.distCp( -sourcePath.getFileSystem(conf), // source file system -Collections.singletonList(sourcePath), // list of source paths -targetPath, -false, -proxyUser, -conf, -ShimLoader.getHadoopShims()); -return 0; - } catch (Exception e) { -currentRetry++; -error = handleException(e, sourcePath, targetPath, currentRetry, proxyUser); -if (error == 0) { -
[GitHub] [hive] aasha commented on a change in pull request #1358: HIVE-23955 : Classification of Error Codes in Replication
aasha commented on a change in pull request #1358: URL: https://github.com/apache/hive/pull/1358#discussion_r466544742 ## File path: ql/src/java/org/apache/hadoop/hive/ql/exec/repl/AtlasDumpTask.java ## @@ -132,31 +130,40 @@ private AtlasReplInfo createAtlasReplInfo() throws SemanticException, MalformedU private long lastStoredTimeStamp() throws SemanticException { Path prevMetadataPath = new Path(work.getPrevAtlasDumpDir(), EximUtil.METADATA_NAME); -BufferedReader br = null; +Retryable retryable = Retryable.builder() + .withHiveConf(conf) + .withRetryOnException(IOException.class) + .withFailOnException(FileNotFoundException.class).build(); try { - FileSystem fs = prevMetadataPath.getFileSystem(conf); - br = new BufferedReader(new InputStreamReader(fs.open(prevMetadataPath), Charset.defaultCharset())); - String line = br.readLine(); - if (line == null) { -throw new SemanticException("Could not read lastStoredTimeStamp from atlas metadata file"); - } - String[] lineContents = line.split("\t", 5); - return Long.parseLong(lineContents[1]); -} catch (Exception ex) { - throw new SemanticException(ex); -} finally { - if (br != null) { + return retryable.executeCallable(() -> { +BufferedReader br = null; try { - br.close(); -} catch (IOException e) { - throw new SemanticException(e); + FileSystem fs = prevMetadataPath.getFileSystem(conf); + br = new BufferedReader(new InputStreamReader(fs.open(prevMetadataPath), Charset.defaultCharset())); + String line = br.readLine(); + if (line == null) { +throw new SemanticException(ErrorMsg.REPL_INVALID_CONFIG_FOR_SERVICE + .format("Could not read lastStoredTimeStamp from atlas metadata file", "atlas")); + } + String[] lineContents = line.split("\t", 5); + return Long.parseLong(lineContents[1]); +} finally { + if (br != null) { +try { + br.close(); +} catch (IOException e) { + //Do nothing +} + } } - } + }); +} catch (Exception e) { + throw new SemanticException(ErrorMsg.REPL_RETRY_EXHAUSTED.format(e.getMessage()), e); Review comment: Yes but this is of type exception 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: us...@infra.apache.org - To unsubscribe, e-mail: gitbox-unsubscr...@hive.apache.org For additional commands, e-mail: gitbox-h...@hive.apache.org