You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@cassandra.apache.org by "Hiller, Dean" <De...@nrel.gov> on 2013/04/17 16:50:23 UTC
looking at making astyanax asynchronous but cassandra-thrift-1.1.1
doesn't look right
Is cassandra-thrift-1.1.1.jar the generated code? I see a send() and recv() but I don't see a send(Callback cb) that is typicaly of true asynchronous platforms. Ie. I don't know when to call recv myself obviously if I am trying to make astyanax truly asynchronous.
The reason I ask is we have a 100k row upload that with synchronous 20 threads takes around 30 seconds and with simulation, we predict this would be done in 3 seconds with an asynch api as our threads would not get held up like they do now. I guess we can try to crank it up to 100 threads to get it running a bit faster for now :( :(.
Thanks,
Dean
Re: looking at making astyanax asynchronous but cassandra-thrift-1.1.1 doesn't look right
Posted by aaron morton <aa...@thelastpickle.com>.
Here's an example I did in python a long time ago http://www.mail-archive.com/user@cassandra.apache.org/msg04775.html
Call send() then select on the file handle, when it's ready to read call recv().
Or just add more threads on your side :)
Cheers
-----------------
Aaron Morton
Freelance Cassandra Consultant
New Zealand
@aaronmorton
http://www.thelastpickle.com
On 18/04/2013, at 2:50 AM, "Hiller, Dean" <De...@nrel.gov> wrote:
> Is cassandra-thrift-1.1.1.jar the generated code? I see a send() and recv() but I don't see a send(Callback cb) that is typicaly of true asynchronous platforms. Ie. I don't know when to call recv myself obviously if I am trying to make astyanax truly asynchronous.
>
> The reason I ask is we have a 100k row upload that with synchronous 20 threads takes around 30 seconds and with simulation, we predict this would be done in 3 seconds with an asynch api as our threads would not get held up like they do now. I guess we can try to crank it up to 100 threads to get it running a bit faster for now :( :(.
>
> Thanks,
> Dean