Cassandra for Oracle DBA’s Part 3 – Architecture

cassandraenv.sh 

MAX_HEAP_SIZE

SGA is shared global area , where in the Java Heap Memory is used for Cassandra Core Engine

Memtables in Cassandra used for each single table to store the recently used or modifying data, unlike a large buffer cache

suppose you have DML on 4 tables , it will create 4 Memtables in the memory 

When an operation has to perform on a block, one must know the Block location in terms of Block Number, Disk number, node name etc.

To achieve this, Oracle relies on Data Dictionary Cache and get block number and file number and then read from the disk.

In Cassandra, Data Stores on disk based on Partition Key Value and this Key will be stored in Partition Key Cache to effectively get the data from disk 

In Oracle to find the data block in buffer cache, Oracle uses mod function to find the right bucket for the data. This will help process not to traverse entire block chain.

Similarly, Cassandra checks the Bloom filter to discover which SSTables are likely to have the request partition data. The Bloom filter is stored in off-heap memory. Each SSTable has a Bloom filter associated with it.

Create table ..... (bloom_filter_fp_chance = 0.01)

min_index_interval 128 and max_index_interval 2048, or index_interval 128

The partition summary is an off-heap in-memory structure that stores a sampling of the partition index. A partition index contains all partition keys, whereas a partition summary samples every X keys, and maps the location of every Xth key's location in the index file. For example, if the partition summary is set to sample every 20 keys, it will store the location of the first key as the beginning of the SSTable file, the 20th key and its location in the file, and so on.

Compression offset maps holds the offset information for compressed blocks. By default all tables in Cassandra are compressed and more the compression ratio larger the compression offset table.

commitlog_total_space_in_mb

commitlog_segment_size_in_mb

sstable_size_in_mb

sstable_min_threshold

Where the SSTables

• Each table will have a physical SSTable on disk
• SSTables are immutable, well immutable means it cannot be written again, kind of Write and then Readonly (like temp)
• So how about a write that need to modify the Table aka SSTable?
• A transaction will create a new SSTable with new data on same disk with increment the file number
• Both of this SSTables (aka files) will be merged together called Compaction when required or when reached the number of SSTables threshold
• In multinode environment, this SSTables are spawned across nodes with data distributed based on Token Range of that node.

SSTables consists of different types:-

• BloomFilter
• Index
• Compression File
• Statistics File
• Datafile
• Summary File
• TOC.txt

CREATE KEYSPACE TEST
  WITH REPLICATION = { 'class' : 'SimpleStrategy', 'replication_factor' : n };

Keyspace is a schema a collection of Tables a top level namespace

Replication Factor is used to determine the Mirroring of data in a multinode environment

CREATE TABLE TEST.Users(
  ID text,
  firstname text,
  lastname text,
  PRIMARY KEY (ID)
)

A table is often called as ColumnFamily in NO-SQL , created with primary key , 

CREATE TABLE TEST.Users(
  ID text,
  firstname text,
  lastname text,
  PRIMARY KEY (ID)
)

primary key means the partition key to store the data in column pair not the primary key in Oracle.

As like Oracle's Composite Primary Key aka Partitioning you can set multiple columns to the Primary Key

CREATE INDEX user_state
   ON TEST.Users (lastname);

CREATE TABLE TEST.page_view_counts
  (counter_value counter,
  url_name varchar,
  page_name varchar,
  PRIMARY KEY (url_name, page_name)
);

Here the datatype counter depicts , its a sequence based value, so you just can increment in your dml and code 

ex: UPDATE counterks.page_view_counts SET counter_value = counter_value + 1 WHERE url_name='www.datastax.com' AND page_name='home';

CREATE TRIGGER TEST.myTrigger
  ON myTable
  USING 'org.apache.cassandra.triggers.INDEXCHECK'

 CREATE TYPE TEST.address (
  street text,
  city text,
  zip_code int,
  phones set<text>
)