hudi 0.9.0适配hbase 2.2.6

总览

在hudi中，hbase可以作为索引数据的存储，hudi默认使用的hbase版本为1.2.3。

在hbase从1.x升级到2.x之后，其api发生了较大的变化，直接修改hudi中hbase的版本是不合适的，即会发生编译错误。

本文对部分源码进行修改以使hbase 2.2.6适配hudi 0.9.0

编译报错

如果我们直接修改hbase的版本为2.2.6的话，会出现如下编译错误：

[ERROR] Failed to execute goal org.apache.maven.plugins:maven-compiler-plugin:3.8.0:compile (default-compile) on project hudi-common: Compilation failure: Compilation failure: 
[ERROR] /root/hudi-0.9.0/hudi-common/src/main/java/org/apache/hudi/common/bootstrap/index/HFileBootstrapIndex.java:[181,34] no suitable method found for createReader(org.apache.hadoop.fs.FileSystem,org.apache.hudi.common.bootstrap.index.HFileBootstrapIndex.HFilePathForReader,org.apache.hadoop.hbase.io.hfile.CacheConfig,org.apache.hadoop.conf.Configuration)
[ERROR]     method org.apache.hadoop.hbase.io.hfile.HFile.createReader(org.apache.hadoop.fs.FileSystem,org.apache.hadoop.fs.Path,org.apache.hadoop.hbase.io.FSDataInputStreamWrapper,long,org.apache.hadoop.hbase.io.hfile.CacheConfig,boolean,org.apache.hadoop.conf.Configuration) is not applicable
[ERROR]       (actual and formal argument lists differ in length)
[ERROR]     method org.apache.hadoop.hbase.io.hfile.HFile.createReader(org.apache.hadoop.fs.FileSystem,org.apache.hadoop.fs.Path,org.apache.hadoop.conf.Configuration) is not applicable
[ERROR]       (actual and formal argument lists differ in length)
[ERROR]     method org.apache.hadoop.hbase.io.hfile.HFile.createReader(org.apache.hadoop.fs.FileSystem,org.apache.hadoop.fs.Path,org.apache.hadoop.hbase.io.hfile.CacheConfig,boolean,org.apache.hadoop.conf.Configuration) is not applicable
[ERROR]       (actual and formal argument lists differ in length)
[ERROR] /root/hudi-0.9.0/hudi-common/src/main/java/org/apache/hudi/common/bootstrap/index/HFileBootstrapIndex.java:[309,93] cannot find symbol
[ERROR]   symbol:   method getKeyValue()
[ERROR]   location: variable scanner of type org.apache.hadoop.hbase.io.hfile.HFileScanner
[ERROR] /root/hudi-0.9.0/hudi-common/src/main/java/org/apache/hudi/common/bootstrap/index/HFileBootstrapIndex.java:[534,51] incompatible types: org.apache.hudi.common.bootstrap.index.HFileBootstrapIndex.HoodieKVComparator cannot be converted to org.apache.hadoop.hbase.CellComparator
[ERROR] /root/hudi-0.9.0/hudi-common/src/main/java/org/apache/hudi/common/bootstrap/index/HFileBootstrapIndex.java:[537,51] incompatible types: org.apache.hudi.common.bootstrap.index.HFileBootstrapIndex.HoodieKVComparator cannot be converted to org.apache.hadoop.hbase.CellComparator
[ERROR] /root/hudi-0.9.0/hudi-common/src/main/java/org/apache/hudi/io/storage/HoodieHFileReader.java:[72,24] no suitable method found for createReader(org.apache.hadoop.fs.FileSystem,org.apache.hadoop.fs.Path,org.apache.hadoop.hbase.io.hfile.CacheConfig,org.apache.hadoop.conf.Configuration)
[ERROR]     method org.apache.hadoop.hbase.io.hfile.HFile.createReader(org.apache.hadoop.fs.FileSystem,org.apache.hadoop.fs.Path,org.apache.hadoop.hbase.io.FSDataInputStreamWrapper,long,org.apache.hadoop.hbase.io.hfile.CacheConfig,boolean,org.apache.hadoop.conf.Configuration) is not applicable
[ERROR]       (actual and formal argument lists differ in length)
[ERROR]     method org.apache.hadoop.hbase.io.hfile.HFile.createReader(org.apache.hadoop.fs.FileSystem,org.apache.hadoop.fs.Path,org.apache.hadoop.conf.Configuration) is not applicable
[ERROR]       (actual and formal argument lists differ in length)
[ERROR]     method org.apache.hadoop.hbase.io.hfile.HFile.createReader(org.apache.hadoop.fs.FileSystem,org.apache.hadoop.fs.Path,org.apache.hadoop.hbase.io.hfile.CacheConfig,boolean,org.apache.hadoop.conf.Configuration) is not applicable
[ERROR]       (actual and formal argument lists differ in length)
[ERROR] /root/hudi-0.9.0/hudi-common/src/main/java/org/apache/hudi/io/storage/HoodieHFileReader.java:[80,24] no suitable method found for createReader(org.apache.hadoop.fs.FileSystem,org.apache.hadoop.fs.Path,org.apache.hadoop.hbase.io.FSDataInputStreamWrapper,int,org.apache.hadoop.hbase.io.hfile.CacheConfig,org.apache.hadoop.conf.Configuration)
[ERROR]     method org.apache.hadoop.hbase.io.hfile.HFile.createReader(org.apache.hadoop.fs.FileSystem,org.apache.hadoop.fs.Path,org.apache.hadoop.hbase.io.FSDataInputStreamWrapper,long,org.apache.hadoop.hbase.io.hfile.CacheConfig,boolean,org.apache.hadoop.conf.Configuration) is not applicable
[ERROR]       (actual and formal argument lists differ in length)
[ERROR]     method org.apache.hadoop.hbase.io.hfile.HFile.createReader(org.apache.hadoop.fs.FileSystem,org.apache.hadoop.fs.Path,org.apache.hadoop.conf.Configuration) is not applicable
[ERROR]       (actual and formal argument lists differ in length)
[ERROR]     method org.apache.hadoop.hbase.io.hfile.HFile.createReader(org.apache.hadoop.fs.FileSystem,org.apache.hadoop.fs.Path,org.apache.hadoop.hbase.io.hfile.CacheConfig,boolean,org.apache.hadoop.conf.Configuration) is not applicable
[ERROR]       (actual and formal argument lists differ in length)
[ERROR] /root/hudi-0.9.0/hudi-common/src/main/java/org/apache/hudi/io/storage/HoodieHFileReader.java:[114,56] incompatible types: org.apache.hadoop.hbase.io.hfile.HFileBlock cannot be converted to java.nio.ByteBuffer
[ERROR] /root/hudi-0.9.0/hudi-common/src/main/java/org/apache/hudi/io/storage/HoodieHFileReader.java:[149,27] cannot find symbol
[ERROR]   symbol:   method getKeyValue()
[ERROR]   location: variable scanner of type org.apache.hadoop.hbase.io.hfile.HFileScanner
[ERROR] /root/hudi-0.9.0/hudi-common/src/main/java/org/apache/hudi/io/storage/HoodieHFileReader.java:[180,54] cannot find symbol
[ERROR]   symbol:   method getKeyValue()
[ERROR]   location: variable scanner of type org.apache.hadoop.hbase.io.hfile.HFileScanner
[ERROR] /root/hudi-0.9.0/hudi-common/src/main/java/org/apache/hudi/io/storage/HoodieHFileReader.java:[200,50] cannot find symbol
[ERROR]   symbol:   method getKeyValue()
[ERROR]   location: variable scanner of type org.apache.hadoop.hbase.io.hfile.HFileScanner
[ERROR] /root/hudi-0.9.0/hudi-common/src/main/java/org/apache/hudi/io/storage/HoodieHFileReader.java:[224,28] cannot find symbol
[ERROR]   symbol:   method getKeyValue()
[ERROR]   location: variable keyScanner of type org.apache.hadoop.hbase.io.hfile.HFileScanner
[ERROR] -> [Help 1]
[ERROR] 
[ERROR] To see the full stack trace of the errors, re-run Maven with the -e switch.
[ERROR] Re-run Maven using the -X switch to enable full debug logging.
[ERROR] 
[ERROR] For more information about the errors and possible solutions, please read the following articles:
[ERROR] [Help 1] http://cwiki.apache.org/confluence/display/MAVEN/MojoFailureException
[ERROR] 
[ERROR] After correcting the problems, you can resume the build with the command
[ERROR]   mvn <goals> -rf :hudi-common

针对上述问题，我们发现主要的兼容性问题有4个：

1. HFileBootstrapIndex#createReader方法参数问题
2. HFileScanner#getKeyValue方法在hbase 2.2.6中已经不存在了
3. HFile#withComparator方法的入参为CellComparator类型，而源码中的则是HoodieKVComparator
4. HFileReaderImpl#getMetaBlock方法返回参数的变化由ByteBuffer变为HFileBlock

hudi源码修改

针对上述问题，我们进行如下修改：

1. 查看Hbase源码，我们发现HFileBootstrapIndex#createReader增加了一个类型为boolean的参数primaryReplicaReader，该参数说明如下：

   /**
   * Creates reader with cache configuration disabled
   * @param fs filesystem
   * @param path Path to file to read
   * @return an active Reader instance
   * @throws IOException Will throw a CorruptHFileException
   * (DoNotRetryIOException subtype) if hfile is corrupt/invalid.
   */
   public static Reader createReader(FileSystem fs, Path path, Configuration conf)
     throws IOException {
   // The primaryReplicaReader is mainly used for constructing block cache key, so if we do not use
   // block cache then it is OK to set it as any value. We use true here.
   return createReader(fs, path, CacheConfig.DISABLED, true, conf);
   }

   所以我们可以对HFileBootstrapIndex类的182-183行进行如下修改从而解决这个问题(增加了一个true参数)：

     HFile.Reader reader = HFile.createReader(fileSystem, new HFilePathForReader(hFilePath),
         new CacheConfig(conf), true, conf);

   （在编译中可能会遇到其他相同的createReader参数问题，用上述方法进行修改即可）

2. HFileScanner#getKeyValue方法在hbase升级之后已经替换为HFileScanner#getCell，具体可参考HFile的提交历史.

所以我们可以对HFileBootstrapIndex类的309行修改为：

          keys.add(converter.apply(getUserKeyFromCellKey(CellUtil.getCellKeyAsString(scanner.getCell()))));

（在编译中可能会遇到其他相同的getKeyValue方法问题，用上述方法进行修改即可）

3. 在hbase升级之后，我们可以看到HFile#withComparator需要的参数为CellComparator：

所以我们可以通过修改HFileBootstrapIndex的585-586行，使HoodieKVComparator继承CellComparatorImpl

  public static class HoodieKVComparator extends CellComparatorImpl {
  }

（在编译中可能会遇到其他相同问题，用上述方法进行修改即可）

4. HFileReaderImpl#getMetaBlock方法返回参数具体问题如下：

我们再来看一下这个方法的变更历史：

由此我们知道可以通过修改HoodieHFileReader的115行，改为如下：

      ByteBuff serializedFilter = reader.getMetaBlock(KEY_BLOOM_FILTER_META_BLOCK, false).getBufferWithoutHeader();

（在编译中可能会遇到相同问题，用上述方法进行修改即可）

其他问题

由于hudi 0.9.0的jetty版本和hbase 2.2.6的jetty版本存在冲突，所以我们需要排除掉hbase的jetty，使用hudi要求的jetty版本。
这个问题可能在编译阶段不会报错，但是在运行阶段是会报错的。

那么解决了上述问题之后，就可以让hudi使用hbase 2.2.6啦！