zrepl/rpc/client.go

137 lines
2.6 KiB
Go
Raw Normal View History

reimplement io.ReadWriteCloser based RPC mechanism The existing ByteStreamRPC requires writing RPC stub + server code for each RPC endpoint. Does not scale well. Goal: adding a new RPC call should - not require writing an RPC stub / handler - not require modifications to the RPC lib The wire format is inspired by HTTP2, the API by net/rpc. Frames are used for framing messages, i.e. a message is made of multiple frames which are glued together using a frame-bridging reader / writer. This roughly corresponds to HTTP2 streams, although we're happy with just one stream at any time and the resulting non-need for flow control, etc. Frames are typed using a header. The two most important types are 'Header' and 'Data'. The RPC protocol is built on top of this: - Client sends a header => multiple frames of type 'header' - Client sends request body => mulitiple frames of type 'data' - Server reads a header => multiple frames of type 'header' - Server reads request body => mulitiple frames of type 'data' - Server sends response header => ... - Server sends response body => ... An RPC header is serialized JSON and always the same structure. The body is of the type specified in the header. The RPC server and client use some semi-fancy reflection tequniques to automatically infer the data type of the request/response body based on the method signature of the server handler; or the client parameters, respectively. This boils down to a special-case for io.Reader, which are just dumped into a series of data frames as efficiently as possible. All other types are (de)serialized using encoding/json. The RPC layer and Frame Layer log some arbitrary messages that proved useful during debugging. By default, they log to a non-logger, which should not have a big impact on performance. pprof analysis shows the implementation spends its CPU time 60% waiting for syscalls 30% in memmove 10% ... On a Intel(R) Core(TM) i7-6600U CPU @ 2.60GHz CPU, Linux 4.12, the implementation achieved ~3.6GiB/s. Future optimization may include spice(2) / vmspice(2) on Linux, although this doesn't fit so well with the heavy use of io.Reader / io.Writer throughout the codebase. The existing hackaround for local calls was re-implemented to fit the new interface of PRCServer and RPCClient. The 'R'PC method invocation is a bit slower because reflection is involved inbetween, but otherwise performance should be no different. The RPC code currently does not support multipart requests and thus does not support the equivalent of a POST. Thus, the switch to the new rpc code had the following fallout: - Move request objects + constants from rpc package to main app code - Sacrifice the hacky 'push = pull me' way of doing push -> need to further extend RPC to support multipart requests or something to implement this properly with additional interfaces -> should be done after replication is abstracted better than separate algorithms for doPull() and doPush()
2017-08-19 22:37:14 +02:00
package rpc
import (
"bytes"
"encoding/json"
"io"
"reflect"
"github.com/pkg/errors"
)
type Client struct {
ml *MessageLayer
logger Logger
}
func NewClient(rwc io.ReadWriteCloser) *Client {
return &Client{NewMessageLayer(rwc), noLogger{}}
}
func (c *Client) SetLogger(logger Logger, logMessageLayer bool) {
c.logger = logger
if logMessageLayer {
c.ml.logger = logger
} else {
c.ml.logger = noLogger{}
}
}
func (c *Client) Close() (err error) {
c.logger.Printf("sending Close request")
header := Header{
DataType: DataTypeControl,
Endpoint: ControlEndpointClose,
Accept: DataTypeControl,
}
err = c.ml.WriteHeader(&header)
if err != nil {
return
}
c.logger.Printf("reading Close ACK")
ack, err := c.ml.ReadHeader()
if err != nil {
return err
}
c.logger.Printf("received Close ACK: %#v", ack)
if ack.Error != StatusOK {
err = errors.Errorf("error hanging up: remote error (%s) %s", ack.Error, ack.ErrorMessage)
return
reimplement io.ReadWriteCloser based RPC mechanism The existing ByteStreamRPC requires writing RPC stub + server code for each RPC endpoint. Does not scale well. Goal: adding a new RPC call should - not require writing an RPC stub / handler - not require modifications to the RPC lib The wire format is inspired by HTTP2, the API by net/rpc. Frames are used for framing messages, i.e. a message is made of multiple frames which are glued together using a frame-bridging reader / writer. This roughly corresponds to HTTP2 streams, although we're happy with just one stream at any time and the resulting non-need for flow control, etc. Frames are typed using a header. The two most important types are 'Header' and 'Data'. The RPC protocol is built on top of this: - Client sends a header => multiple frames of type 'header' - Client sends request body => mulitiple frames of type 'data' - Server reads a header => multiple frames of type 'header' - Server reads request body => mulitiple frames of type 'data' - Server sends response header => ... - Server sends response body => ... An RPC header is serialized JSON and always the same structure. The body is of the type specified in the header. The RPC server and client use some semi-fancy reflection tequniques to automatically infer the data type of the request/response body based on the method signature of the server handler; or the client parameters, respectively. This boils down to a special-case for io.Reader, which are just dumped into a series of data frames as efficiently as possible. All other types are (de)serialized using encoding/json. The RPC layer and Frame Layer log some arbitrary messages that proved useful during debugging. By default, they log to a non-logger, which should not have a big impact on performance. pprof analysis shows the implementation spends its CPU time 60% waiting for syscalls 30% in memmove 10% ... On a Intel(R) Core(TM) i7-6600U CPU @ 2.60GHz CPU, Linux 4.12, the implementation achieved ~3.6GiB/s. Future optimization may include spice(2) / vmspice(2) on Linux, although this doesn't fit so well with the heavy use of io.Reader / io.Writer throughout the codebase. The existing hackaround for local calls was re-implemented to fit the new interface of PRCServer and RPCClient. The 'R'PC method invocation is a bit slower because reflection is involved inbetween, but otherwise performance should be no different. The RPC code currently does not support multipart requests and thus does not support the equivalent of a POST. Thus, the switch to the new rpc code had the following fallout: - Move request objects + constants from rpc package to main app code - Sacrifice the hacky 'push = pull me' way of doing push -> need to further extend RPC to support multipart requests or something to implement this properly with additional interfaces -> should be done after replication is abstracted better than separate algorithms for doPull() and doPush()
2017-08-19 22:37:14 +02:00
}
c.logger.Printf("closing MessageLayer")
if err = c.ml.Close(); err != nil {
c.logger.Printf("error closing RWC: %+v", err)
return
}
reimplement io.ReadWriteCloser based RPC mechanism The existing ByteStreamRPC requires writing RPC stub + server code for each RPC endpoint. Does not scale well. Goal: adding a new RPC call should - not require writing an RPC stub / handler - not require modifications to the RPC lib The wire format is inspired by HTTP2, the API by net/rpc. Frames are used for framing messages, i.e. a message is made of multiple frames which are glued together using a frame-bridging reader / writer. This roughly corresponds to HTTP2 streams, although we're happy with just one stream at any time and the resulting non-need for flow control, etc. Frames are typed using a header. The two most important types are 'Header' and 'Data'. The RPC protocol is built on top of this: - Client sends a header => multiple frames of type 'header' - Client sends request body => mulitiple frames of type 'data' - Server reads a header => multiple frames of type 'header' - Server reads request body => mulitiple frames of type 'data' - Server sends response header => ... - Server sends response body => ... An RPC header is serialized JSON and always the same structure. The body is of the type specified in the header. The RPC server and client use some semi-fancy reflection tequniques to automatically infer the data type of the request/response body based on the method signature of the server handler; or the client parameters, respectively. This boils down to a special-case for io.Reader, which are just dumped into a series of data frames as efficiently as possible. All other types are (de)serialized using encoding/json. The RPC layer and Frame Layer log some arbitrary messages that proved useful during debugging. By default, they log to a non-logger, which should not have a big impact on performance. pprof analysis shows the implementation spends its CPU time 60% waiting for syscalls 30% in memmove 10% ... On a Intel(R) Core(TM) i7-6600U CPU @ 2.60GHz CPU, Linux 4.12, the implementation achieved ~3.6GiB/s. Future optimization may include spice(2) / vmspice(2) on Linux, although this doesn't fit so well with the heavy use of io.Reader / io.Writer throughout the codebase. The existing hackaround for local calls was re-implemented to fit the new interface of PRCServer and RPCClient. The 'R'PC method invocation is a bit slower because reflection is involved inbetween, but otherwise performance should be no different. The RPC code currently does not support multipart requests and thus does not support the equivalent of a POST. Thus, the switch to the new rpc code had the following fallout: - Move request objects + constants from rpc package to main app code - Sacrifice the hacky 'push = pull me' way of doing push -> need to further extend RPC to support multipart requests or something to implement this properly with additional interfaces -> should be done after replication is abstracted better than separate algorithms for doPull() and doPush()
2017-08-19 22:37:14 +02:00
return err
}
func (c *Client) recvResponse() (h *Header, err error) {
h, err = c.ml.ReadHeader()
if err != nil {
return nil, errors.Wrap(err, "cannot read header")
}
// TODO validate
return
}
func (c *Client) writeRequest(h *Header) (err error) {
// TODO validate
err = c.ml.WriteHeader(h)
if err != nil {
return errors.Wrap(err, "cannot write header")
}
return
}
func (c *Client) Call(endpoint string, in, out interface{}) (err error) {
var accept DataType
{
outType := reflect.TypeOf(out)
if typeIsIOReaderPtr(outType) {
accept = DataTypeOctets
} else {
accept = DataTypeMarshaledJSON
}
}
h := Header{
Endpoint: endpoint,
DataType: DataTypeMarshaledJSON,
Accept: accept,
}
if err = c.writeRequest(&h); err != nil {
return err
}
var buf bytes.Buffer
if err = json.NewEncoder(&buf).Encode(in); err != nil {
panic("cannot encode 'in' parameter")
}
if err = c.ml.WriteData(&buf); err != nil {
return err
}
rh, err := c.recvResponse()
if err != nil {
return err
}
if rh.Error != StatusOK {
return &RPCError{rh}
}
rd := c.ml.ReadData()
switch accept {
case DataTypeOctets:
c.logger.Printf("setting out to ML data reader")
outPtr := out.(*io.Reader) // we checked that above
*outPtr = rd
case DataTypeMarshaledJSON:
c.logger.Printf("decoding marshaled json")
if err = json.NewDecoder(c.ml.ReadData()).Decode(out); err != nil {
return errors.Wrap(err, "cannot decode marshaled reply")
}
default:
panic("implementation error") // accept is controlled by us
}
return
}