cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Binh Van Nguyen (JIRA)" <>
Subject [jira] [Updated] (CASSANDRA-5182) Deletable rows are sometimes not removed during compaction
Date Wed, 23 Jan 2013 01:44:12 GMT


Binh Van Nguyen updated CASSANDRA-5182:

    Attachment: test_ttl.tar.gz

Here is simple code to reproduce the issue. The simple code use 10 threads to continuously
write to C*. All the column has 300 in its TTL. While running the code, check the number of
sstable of column family, you will see that it keeps growing and never stop
Here are steps to run:
- Start C* on your localhost
- Create a keyspace named test
- Use the following command to create column family cf
CREATE COLUMN FAMILY cf WITH comparator = UTF8Type AND key_validation_class = UTF8Type AND
default_validation_class = UTF8Type AND gc_grace = 0 AND caching = none AND bloom_filter_fp_chance
= 1.0 AND compaction_strategy='LeveledCompactionStrategy' AND compaction_strategy_options
= { sstable_size_in_mb : 1 } AND compression_options = { chunk_length_kb : 64, sstable_compression
: 'SnappyCompressor' };

- Extract the code switch to extracted directory
- Run mvn package
- Run java -jar target/test_ttl-1.0-SNAPSHOT-jar-with-dependencies.jar 10

> Deletable rows are sometimes not removed during compaction
> ----------------------------------------------------------
>                 Key: CASSANDRA-5182
>                 URL:
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Core
>    Affects Versions: 1.1.5
>            Reporter: Binh Van Nguyen
>         Attachments: test_ttl.tar.gz
> Our use case is write heavy and read seldom.  To optimize the space used, we've set the
bloom_filter_fp_ratio=1.0  That along with the fact that each row is only written to one time
and that there are more than 20 SSTables keeps the rows from ever being compacted. Here is
the code:
> We hit this conner case and because of this C* keeps consuming more and more space on
disk while it should not.

This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see:

View raw message