We just do double (triple) writes on the application side. We're shadowing
partial of the traffic to a smaller staging cluster for new release test,
performance/configuration tuning.
On Sat, Jun 2, 2018 at 7:47 PM Nate McCall wrote:
> > Thanks for sharing, Jay.
> >
> > Could you say a bit more
On Sat, Jun 2, 2018 at 4:46 PM, Jay Zhuang wrote:
> We're shadowing some production traffics to a Rocksandra cluster (
> https://github.com/Instagram/cassandra/tree/rocks_3.0), the P99 latency is
> significantly improved (about 6x for read, 12x for write). Here are the
> test details:
>
>
> Thanks for sharing, Jay.
>
> Could you say a bit more about how you’ve approached shadowing traffic
> against an alternate cluster? (Asking not so much with regard to the
> Rocksandra test, but toward shadowing methodology in general).
+1 to this. If someone out there had a method to
Thanks for sharing, Jay.
Could you say a bit more about how you’ve approached shadowing traffic against
an alternate cluster? (Asking not so much with regard to the Rocksandra test,
but toward shadowing methodology in general).
– Scott
> On Jun 1, 2018, at 9:46 PM, Jay Zhuang wrote:
>
>
We're shadowing some production traffics to a Rocksandra cluster (
https://github.com/Instagram/cassandra/tree/rocks_3.0), the P99 latency is
significantly improved (about 6x for read, 12x for write). Here are the
test details:
