cassandra-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Brandon Williams <dri...@gmail.com>
Subject Re: Time to insert bulk data is very high comparing to database
Date Fri, 06 Nov 2009 18:57:15 GMT
On Fri, Nov 6, 2009 at 10:54 AM, Richard grossman <richiesgr@gmail.com>wrote:

> Hi,
>
> We've started to use cassandra as a viable solution for distributed
> database.
> I've only define 1 column family simple as that :
>      <ColumnFamily CompareWith="UTF8Type"
>                       Name="channelShow"
>                       FlushPeriodInMinutes="15"/>
>
> The average time i get is ~680 rows/second. The actual performance for our
> oracle db is ~1400 rows/second.
> I know the test is not very accurate because the cassandra and oracle db
> doesn't run on the same hardware but there is really a big difference.
> Is there something to check why it's so slow ??
>

This is exceptionally slow, can you tell me more about your hardware setup?
 How many columns are you inserting per row and how large are the values?
 Are you running the benchmark on the same machine?

On a 4 core machine I get several thousands of inserts per second using
test/system/stress.py (from another machine) on both 0.4 and trunk.

-Brandon

Mime
View raw message