GitHub user kevinjmh opened a pull request:
https://github.com/apache/carbondata/pull/3005 [CARBONDATA-3185] Fix alluxio file rename **Problem** Exception thrown when create table on alluxio because rename schema file failed **Analyse** Re-run the command after adding some logs, I found that the file system object it uses is `alluxio.hadoop.FileSystem`, and it is not an instanceof `DistributedFileSystem`, such that the renameForce method return false. **Solution** By checking hierarchy of class [`alluxio.hadoop.FileSystem`](https://github.com/Alluxio/alluxio/blob/branch-1.8/core/client/hdfs/src/main/java/alluxio/hadoop/FileSystem.java), it is ok to use method of hadoop FileSystem directly. We can fix it by remove instant type check and rewrite the overwrite file code by hand. ``` FileSystem (org.apache.hadoop.fs) |__ AbstractFileSystem (alluxio.hadoop) |__FileSystem (alluxio.hadoop) ``` *Log* ``` 18/12/20 09:31:50 ERROR thriftserver.SparkExecuteStatementOperation: Error executing query, currentState RUNNING, org.apache.carbondata.spark.exception.ProcessMetaDataException: operation failed for default.alluxio: Create table'alluxio' in database 'default' failed, temporary file renaming failed, src=alluxio://localhost:19998/user/hive/warehouse/carbon.store/default/alluxio/Metadata/schema.write, dest=alluxio://localhost:19998/user/hive/warehouse/carbon.store/default/alluxio/Metadata/schema at org.apache.spark.sql.execution.command.MetadataProcessOpeation$class.throwMetadataException(package.scala:55) at org.apache.spark.sql.execution.command.MetadataCommand.throwMetadataException(package.scala:120) at org.apache.spark.sql.execution.command.table.CarbonCreateTableCommand.processMetadata(CarbonCreateTableCommand.scala:179) at org.apache.spark.sql.execution.command.MetadataCommand$$anonfun$run$1.apply(package.scala:122) at org.apache.spark.sql.execution.command.MetadataCommand$$anonfun$run$1.apply(package.scala:122) at org.apache.spark.sql.execution.command.Auditable$class.runWithAudit(package.scala:104) at org.apache.spark.sql.execution.command.MetadataCommand.runWithAudit(package.scala:120) at org.apache.spark.sql.execution.command.MetadataCommand.run(package.scala:122) at org.apache.spark.sql.execution.command.ExecutedCommandExec.sideEffectResult$lzycompute(commands.scala:58) at org.apache.spark.sql.execution.command.ExecutedCommandExec.sideEffectResult(commands.scala:56) at org.apache.spark.sql.execution.command.ExecutedCommandExec.executeCollect(commands.scala:67) at org.apache.spark.sql.Dataset.<init>(Dataset.scala:183) at org.apache.spark.sql.CarbonSession$$anonfun$sql$1.apply(CarbonSession.scala:91) at org.apache.spark.sql.CarbonSession$$anonfun$sql$1.apply(CarbonSession.scala:90) at org.apache.spark.sql.CarbonSession.withProfiler(CarbonSession.scala:136) at org.apache.spark.sql.CarbonSession.sql(CarbonSession.scala:88) ``` Be sure to do all of the following checklist to help us incorporate your contribution quickly and easily: - [ ] Any interfaces changed? - [ ] Any backward compatibility impacted? - [ ] Document update required? - [ ] Testing done Please provide details on - Whether new unit test cases have been added or why no new tests are required? - How it is tested? Please attach test report. - Is it a performance related change? Please attach the performance test report. - Any additional information to help reviewers in testing this change. - [ ] For large changes, please consider breaking it into sub-tasks under an umbrella JIRA. You can merge this pull request into a Git repository by running: $ git pull https://github.com/kevinjmh/carbondata alluxio_fs Alternatively you can review and apply these changes as the patch at: https://github.com/apache/carbondata/pull/3005.patch To close this pull request, make a commit to your master/trunk branch with (at least) the following in the commit message: This closes #3005 ---- commit 78f5d72a5b169b38d7d6bd0f7fbd9ebf8bfa6d55 Author: Manhua <kevinjmh@...> Date: 2018-12-20T02:34:48Z fix alluxio rename fail ---- --- |
Github user xubo245 commented on a diff in the pull request:
https://github.com/apache/carbondata/pull/3005#discussion_r243150083 --- Diff: core/src/main/java/org/apache/carbondata/core/datastore/filesystem/AlluxioCarbonFile.java --- @@ -95,15 +94,13 @@ public boolean renameForce(String changeToName) { FileSystem fs; try { fs = fileStatus.getPath().getFileSystem(FileFactory.getConfiguration()); - if (fs instanceof DistributedFileSystem) { - ((DistributedFileSystem) fs).rename(fileStatus.getPath(), new Path(changeToName), - org.apache.hadoop.fs.Options.Rename.OVERWRITE); - return true; - } else { - return false; + Path targetPath = new Path(changeToName); + if (fs.exists(targetPath)) { + fs.delete(targetPath, true); --- End diff -- are there any risk for deleting path? --- |
In reply to this post by qiuchenjian-2
Github user xubo245 commented on the issue:
https://github.com/apache/carbondata/pull/3005 There is a PR to fix the same issue: https://github.com/apache/carbondata/pull/2161 --- |
In reply to this post by qiuchenjian-2
Github user CarbonDataQA commented on the issue:
https://github.com/apache/carbondata/pull/3005 Build Success with Spark 2.1.0, Please check CI http://136.243.101.176:8080/job/ApacheCarbonPRBuilder2.1/1861/ --- |
In reply to this post by qiuchenjian-2
Github user kevinjmh commented on a diff in the pull request:
https://github.com/apache/carbondata/pull/3005#discussion_r243150578 --- Diff: core/src/main/java/org/apache/carbondata/core/datastore/filesystem/AlluxioCarbonFile.java --- @@ -95,15 +94,13 @@ public boolean renameForce(String changeToName) { FileSystem fs; try { fs = fileStatus.getPath().getFileSystem(FileFactory.getConfiguration()); - if (fs instanceof DistributedFileSystem) { - ((DistributedFileSystem) fs).rename(fileStatus.getPath(), new Path(changeToName), - org.apache.hadoop.fs.Options.Rename.OVERWRITE); - return true; - } else { - return false; + Path targetPath = new Path(changeToName); + if (fs.exists(targetPath)) { + fs.delete(targetPath, true); --- End diff -- This should be managed by developer. If developer wants a simple rename, he should call `renameTo` instead of `renameForce`. Codes in this PR implements the OVERWRITE(delete before rename), as `S3CarbonFile` , `ViewFSCarbonFile`, `LocalCarbonFile` do --- |
In reply to this post by qiuchenjian-2
Github user CarbonDataQA commented on the issue:
https://github.com/apache/carbondata/pull/3005 Build Failed with Spark 2.2.1, Please check CI http://95.216.28.178:8080/job/ApacheCarbonPRBuilder1/2069/ --- |
In reply to this post by qiuchenjian-2
Github user CarbonDataQA commented on the issue:
https://github.com/apache/carbondata/pull/3005 Build Failed with Spark 2.3.2, Please check CI http://136.243.101.176:8080/job/carbondataprbuilder2.3/10116/ --- |
In reply to this post by qiuchenjian-2
Github user kevinjmh commented on the issue:
https://github.com/apache/carbondata/pull/3005 > There is a PR to fix the same issue: #2161 I see. It fixed LockFile problem I faced now too. Hope it got fix soon. And I am closing this PR now --- |
In reply to this post by qiuchenjian-2
|
Free forum by Nabble | Edit this page |