[
https://issues.apache.org/jira/browse/CASSANDRA-5268?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13611877#comment-13611877
]
Jonathan Ellis commented on CASSANDRA-5268:
-------------------------------------------
The compaction throughput is not a magic turbo button; the problem is you're
CPU bound.
You can enable multithreaded_compaction to partially mitigate this; how much it
helps depends on whether your rows are small enough to deserialize to memory.
> nodetool setcompactionthroughput has no effect
> ----------------------------------------------
>
> Key: CASSANDRA-5268
> URL: https://issues.apache.org/jira/browse/CASSANDRA-5268
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.1.10
> Environment: Linux gravity-jasmapp-1 3.2.0-0.bpo.4-amd64 #1 SMP
> Debian 3.2.32-1~bpo60+1 x86_64 GNU/Linux,
> java version "1.7.0_07"
> Reporter: Pas
>
> issuing a nodetool setcompactionthroughput 64 has no effect (not even a line
> in the log), and compactions go merrily about ~20MB/s on an SSD with fairly
> low utilization (iostat -xm 1 shows ~30-60% util, probably due to reads and
> other processes on the host)
> Token : Token(bytes[3030])
> Gossip active : true
> Thrift active : true
> Load : 106 GB
> Generation No : 1361216406
> Uptime (seconds) : 13825
> Heap Memory (MB) : 2829.19 / 9158.88
> Data Center : datacenter1
> Rack : rack1
> Exceptions : 0
> Key Cache : size 104857584 (bytes), capacity 104857584 (bytes),
> 20638124 hits, 20948679 requests, 0.985 recent hit rate, 14400 save period in
> seconds
> Row Cache : size 0 (bytes), capacity 0 (bytes), 0 hits, 0 requests,
> NaN recent hit rate, 0 save period in seconds
--
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
