cassandra-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From TuxRacer69 <tuxrace...@gmail.com>
Subject Re: Why cassandra single node so slow?
Date Sat, 14 Nov 2009 16:56:28 GMT
Looking at
http://assets.en.oreilly.com/1/event/27/Cassandra_%20Open%20Source%20Bigtable%20+%20Dynamo%20Presentation.pdf
page 26
Cassandra should make around 100 read per sec
and around 10,000 writes per sec (on the data equivalent of  a 50Gb 
mysql db) so your results seem OK

How do your results scale with the database size? I'd say that with 
4,000,000 records your mysql db is still much smaller than 50Gb.
In you mail you did not mention the write results.
Do you see a very fast write with Cassandra (assuming that Laziboy does 
a ZERO consistency level write)?
Do you know the consistency level used by laziboy for reads?
Did you try multi_gets?


ruslan usifov wrote:
> Nо. But i think that all data will be in memory and gstat (FreeBSD 
> utility) show that no any disk activity.
>


Mime
View raw message