cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Oleg Kibirev (JIRA)" <j...@apache.org>
Subject [jira] [Created] (CASSANDRA-5278) Throttle streaming during repair to control disk space use
Date Thu, 21 Feb 2013 01:50:11 GMT
Oleg Kibirev created CASSANDRA-5278:
---------------------------------------

             Summary: Throttle streaming during repair to control disk space use
                 Key: CASSANDRA-5278
                 URL: https://issues.apache.org/jira/browse/CASSANDRA-5278
             Project: Cassandra
          Issue Type: Improvement
          Components: Core
    Affects Versions: 1.1.10
            Reporter: Oleg Kibirev


Currently, doing a repair can consume a multiple of disk space originally needed on the node.
For example, consider a cluster with 4 replicas. In the worst case, a node will have to stream
half of the data from each of the 3 other replicas, consuming 250% of original disk space.

We need some mechanism to allow compaction to catch up and restore reasonable disk usage.
For example, the code that receives a stream can wait until a configurable percentage of available
disk space is free, or until the number of compacted sstable is less than a certain threshold.

--
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: http://www.atlassian.com/software/jira

Mime
View raw message