Pyarrow read timeout always 15 seconds bug?

Hi Dremio Team,

i am using the python client to connect to dremio via pyarrow’s flight protocol.

i follow the pyarrow example and do the following :

options = flight.FlightCallOptions(timeout=600000, headers=headers)
flight_info = cls.__client.get_flight_info(flight.FlightDescriptor.for_command(sqlquery), options)
reader = cls.__client.do_get(flight_info.endpoints[0].ticket, options)
reader.read_pandas()

what i realise is that every time i read execute the sqlquery, if the query exceeds 15 seconds, i got an timeout error even though i set the timeout in the FlightCallOption to be 6000000, the exact exception is below:

Exception: FlightUnavailableError(‘gRPC returned unavailable error, with message: upstream request timeout. Client context: IOError: Server never sent a data message. Detail: Internal. gRPC client debug context: {“created”:"@1645525023.713203665",“description”:“Error received from peer ipv4:172.29.123.223:8000”,“file”:"/opt/vcpkg/buildtrees/grpc/src/17cc203898-db2679e7f2.clean/src/core/lib/surface/call.cc",“file_line”:1067,“grpc_message”:“upstream request timeout”,“grpc_status”:14}’)

do you know why?

@xieyonggang There seems to be an open feature request to support this, let me confirm and get back to you

Thanks, let me know if you found it and any work around now ?

@xieyonggang We are discussing prioritization of this and will keep you posted