You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@cassandra.apache.org by Or Sher <or...@gmail.com> on 2014/10/27 13:45:43 UTC

Hector latency related configuration

Hi all,

We're using Hector in one of our older use cases with C* 1.0.9.
We suspect it increases our total round trip write latency to Cassandra.
C* metrics shows low latency so we assume the problem is somewhere else.
What are the configuration parameters you would recommend to
investigate/change in order to decrease latency.

-- 
Or Sher

Re: Hector latency related configuration

Posted by Jason Wee <pe...@gmail.com>.
Hi,

What version of Hector are you using? Probably start with different
consistency level? Does your node in cluster having memory pressure (you
can check in cassandra system log)?  what is the average node load per node
currently? Also read concurrent_writes in cassandra.yaml if you can
increase higher. You can also use nodetool cfstats to read for the write
latency.

Thanks.

Jason

On Mon, Oct 27, 2014 at 8:45 PM, Or Sher <or...@gmail.com> wrote:

> Hi all,
>
> We're using Hector in one of our older use cases with C* 1.0.9.
> We suspect it increases our total round trip write latency to Cassandra.
> C* metrics shows low latency so we assume the problem is somewhere else.
> What are the configuration parameters you would recommend to
> investigate/change in order to decrease latency.
>
> --
> Or Sher
>