Re: testing retry policy

[Sylvain Lebresne <sy...@datastax.com>]

On Wed, Aug 31, 2016 at 6:56 PM, Jimmy Lin <y2...@gmail.com> wrote:

> hi all,
> I have some customized retry policies that want to test.
> In my single node local cluster, is there anyway to simulate the
> read/write timeout and or unavailable exception?
> I tried to kill the Cassandra process but it won't result in unavailable
> exception but no host available exception and so not going through the
> retry policy logic....
>

This is really more a driver question and you should ideally use the
mailing list of whatever driver you have written your policy for for such
question. But read/write timeout or unavailable exceptions just cannot
happen on a single node cluster since those are errors send by the server,
and if you kill your only server, there is no-one to send you the error. So
you'll need at least 2 nodes to test those.