searching mechanism in cassandra with memtables

Question

can anyone please give some information on how search process happens in cassandra and, how a particular key is identified from the memtables when a search process happens?

Is it like the column keys are stored in memtables and then if it matches it goes into SSTables to gather column data? If yes, will searching pattern searches for every memtable or any mechanism to search for the right memtable too?

Is Searching done by co-ordinator node or the co-ordinator node hands off the search to the right node where the data can be found through any mechanism in cassandra?

score 2 · Accepted Answer · answered Oct 26 '17 at 07:21

Read mechanism in cassandra is pretty amazing.
Cassandra is supposed to combine results from the active memtable and potentially multiple SSTables.
While reading the records from Cassandra, if a record is present in the Row Cache, the read would be served from the row cache without needing to look into any other location This is fastest read path available in Cassandra.
Cassandra starts with checking the Bloom filter to discover which SSTables are likely to have the request partition data. Bloom filter speeds up the process of partition key lookup by narrowing the pool of keys.
If the Bloom filter does not rule out an SSTable, Cassandra checks the partition key cache However, if you couldn't find it in the key cache i.e if the partition key is not there in the key cache, Cassandra will look into the Partition Summaries which is just a sampling of the Partition Indexes. Partition Summary helps to jump into a specific offset in the Partition Index.

Once we are in Partition Index, we now have the offset of the partition key in SSTable and we can directly fetch (into memtable) the record from that offset of SSTable.

Cassandra always consults the Compression offsets to be able to read the data from compressed blocks.

Now, as partition data in Cassandra exists in multiple SSTables ,Cassandra attaches a timestamp to each version of the record (specifically to each column/field) and uses this timestamp to merge records from different SSTables and memtable to present the current version of the complete record.

So for your answer as a matter of fact coordinator starts by asking the query from the preferred node (which it evaluates from the gossip state).This node than performs the above operations.
For more info read
http://docs.datastax.com/en/archived/cassandra/3.x/cassandra/dml/dmlAboutReads.html

Hope it helps !!

Yeah, is it the best information can we get on row cache and bloom filter or would you like to suggest some useful links on this topics, especially bloom filter and how the best SSTable is chosen based on gossip — Chakri Stark, Oct 26 '17 at 09:21
you can go through this https://stackoverflow.com/questions/39327427/what-is-role-of-bloom-filter-in-cassandra — Abhishek Raj, Oct 26 '17 at 09:25
while inserting large chunks of data, it is storing randomly.. is it in-built or the threads generation is random? — Chakri Stark, Nov 06 '17 at 06:04
The data storage is based on the token distribution across the nodes. — Abhishek Raj, Nov 06 '17 at 11:21
yeah, just now i checked it.. The hash functions will generate random tokens and data storage will be random.. — Chakri Stark, Nov 06 '17 at 11:47
But why Cassandra checks ss table if data is found in memtable? — Mahesh Bhuva, Jun 27 '21 at 13:14

searching mechanism in cassandra with memtables

1 Answers1