Hey Boris,

Could you post specifics? Last I checked, the commit log in Cassandra was
designed to be run on a dedicated disk and thus should be writing
sequentially. I wouldn't expect a significant speed boost unless you are
running the commit log on a shared disk.

When the number of SSTables corresponding to a region grows large, however,
I'd expect a significant improvement in Cassandra's read performance. Quite
curious to see these results, really.

Thanks,
Jeff

On Wed, Jun 9, 2010 at 2:59 AM, Boris Shulman <shulm...@gmail.com> wrote:

> Using SSD for a commitlog device can bost performance while using
> cassandra in a batch mode for fsync operations. From my experience ta
> write operation can be 10 times faster when usding SSD for commitlog.
>
> On Mon, Jun 7, 2010 at 6:37 PM, Héctor Izquierdo <izquie...@strands.com>
> wrote:
> > Hi everyone.
> >
> > I wanted to know if anybody has had any experience with cassandra on
> flash
> > storage. At work we have a cluster of 6 machines running Tokyotyrant on
> > flash-io drives (320GB) each, but performance is not what we expected,
> and
> > we'are having some issues with replication and availability. It's also
> hard
> > to manage, and adding/removing nodes is pure hell.
> >
> > We can't afford test hardware with flash storage right now, so could
> > somebody share his experience?
> >
> > Thank you very much
> >
> > Héctor Izquierdo
> >
>

Reply via email to