Michael Dagaev wrote:
Michael,Could you explain, please, why the Hbase PRC has such alimitation?
Its by design (HBase RPC is effectively Hadoop RPC). Server-side it uses non-blocking i/o but on the client-side, its currently as described below.
As I remember, Hbase uses Thrift as an PRC mechanism. Is it a Thrift limitation? Can it be changed?
No. HBase does not use thrift natively. Thrift is an option. Hadoop's RPC was originally written as a more performant alternative to RMI.
St.Ack
