LSM algorithm
hfile
Index, Level Two index
the problem of HBase
1.hbase How to pre-partition.
2.hbase how to provide an interface to the Web foreground to access.
There is no thread-safety problem with the 3.htable API, whether it is a single case or multiple cases in the program.
4. Our hbase probably in the company's business (mainly online mall) is probably a few tables, a few table clusters, probably have what kind of data.
5.hbase concurrency problem.
Part II: HBase Core Knowledge points (for hbase implementation of the core LSM, will be introduced in the next article) 11. The LSM meaning is. A a log structure merging tree (log-structured) B two fork tree C balanced binary tree D long balanced binary tree
12. The next face LSM structure is described correctly. A, c a sequential storage B direct write hard drive C need to flush data to disk D is a search balance tree
LSM is more able to guarantee the performance of any kind of operation. b A Read B write C random read D merge
LSM's read and write operations are independent. A A is. B No. C LSM does not distinguish between read and write D LSM, read and write is the same operation
The LSM structure's data is first stored in (). b A hard drive on B in-memory C disk array D Flash
The data field in the hfile date format is used (). A a stores the actual keyvalue data B The starting point for storing data C Specify the length of the field D stores the starting point of the data block
The Metaindex field in the hfile data format is used (). D A meta block length B meta block end point C meta block data content D Meta block starting point
The Magic field in the hfile data format is used (). A a stores random numbers, prevents data corruption B stores the starting point of the data C stores the start of the data block D specify the length of the field
The KeyValue data format in the hfile data format, the following options describe the correct (). A, D A is byte[] array B has no fixed structure C the size of the data is fixed the length of D has a definite structure
The value portion of the KeyValue data format in the hfile data format is (). C A string B string with complex structure C binary data D compressed data
Part III: Introduction to HBase Advanced application
HBase How to build a Level two index. A, b a MapReduce b coprocessor (hbase after 0.92 introduced the coprocessor (coprocessors) to achieve some exciting new features: the ability to easily establish two indexes, complex filters (predicate push) and access control, etc. C Bloom Filter D Filter
35. For the description of the HBase Level two index, which is correct. A, b a core is inverted table B two level index concept is corresponding to rowkey this "first level" index C two level index using the balanced binary tree D Two-level index using the LSM structure
36. The following description of Bloom filter is correct. A, c A is a very long binary vector and a series of random mapping function B No error rate C has a certain error rate D can remove elements in bloom filter
-------------------------------------Gorgeous split Line-------------------------------------------