incubator-cassandra-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sylvain Lebresne <sylv...@datastax.com>
Subject Re: Is there guidance about compaction thresholds and setting minthreshold to 2?
Date Thu, 04 Apr 2013 17:46:54 GMT
More importantly than CPU, you'll use more I/O. Say you have 4 (like-size)
sstables, compact them all into one file (which is really what
SizeTieredCompaction will try to do) will require twice as much I/O that
with min_compaction=2 versus 4.


On Thu, Apr 4, 2013 at 7:26 PM, Edward Capriolo <edlinuxguru@gmail.com>wrote:

> One would think, but remember only "like sized" sstables compact. You want
> more files roughlt the same size rather then few big ones in most cases,
> but there are no hard fast rules.
>
>
> On Thu, Apr 4, 2013 at 11:36 AM, Peter Haggerty <
> peter.haggerty@librato.com> wrote:
>
>> The default minthreshold for compactions is 4:
>>
>> http://www.datastax.com/docs/1.1/references/nodetool#nodetool-setcompactionthreshold
>>
>> Is there a reason that this value is not "2", the lowest possible value?
>>  If we change this to 2 what should we expect to see? Should we see less
>> growth in storage load and fewer files to seek through when reading but at
>> the cost of higher CPU usage?
>>
>>
>> Thanks,
>>
>> Peter
>>
>>
>

Mime
View raw message