FWIW, as I mentioned in the 1497 comments, the patch makes it abstract so that you can have
any rpc/marshalling format you want with a simple extension point. So if we want to move
to something besides avro, or even like I mentioned do something with Dumbo for streaming,
it's easy to extend.
On May 19, 2011, at 9:23 AM, Jonathan Ellis wrote:
> Avro is there because (1) a long time ago (it now seems) we thought we
> were going to move the main RPC layer to Avro and (2) it simplifies
> using Avro for Streaming, which also seems misguided now
> (https://issues.apache.org/jira/browse/CASSANDRA-1497).
>
> Using "native" Thrift mutations makes the most sense to me now, which
> would keep it similar in structure but avoid the avroToThrift copy.
>
> On Thu, May 19, 2011 at 2:30 AM, Mck <mck@apache.org> wrote:
>> I'd like to investigate using Counters in hadoop using
>> ColumnFamilyOutputFormat.
>>
>> But i see that this class uses outdated ..hadoop.arvo classes
>>
>> Does it make sense to use counters for hadoop output?
>>
>> If i try rewriting ColumnFamilyOutputFormat and friends should it be to
>> the normal ..avro classes, or to something else?
>>
>> ~mck
>>
>>
>
>
>
> --
> Jonathan Ellis
> Project Chair, Apache Cassandra
> co-founder of DataStax, the source for professional Cassandra support
> http://www.datastax.com
|