zrepl/rpc/shared.go

112 lines
2.5 KiB
Go
Raw Permalink 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 (
"fmt"
"github.com/pkg/errors"
"io"
"reflect"
)
type RPCServer interface {
Serve() (err error)
RegisterEndpoint(name string, handler interface{}) (err error)
}
type RPCClient interface {
Call(endpoint string, in, out interface{}) (err error)
Close() (err error)
}
type Logger interface {
Printf(format string, args ...interface{})
}
type noLogger struct{}
func (l noLogger) Printf(format string, args ...interface{}) {}
func typeIsIOReader(t reflect.Type) bool {
return t == reflect.TypeOf((*io.Reader)(nil)).Elem()
}
func typeIsIOReaderPtr(t reflect.Type) bool {
return t == reflect.TypeOf((*io.Reader)(nil))
}
// An error returned by the Client if the response indicated a status code other than StatusOK
type RPCError struct {
ResponseHeader *Header
}
func (e *RPCError) Error() string {
return fmt.Sprintf("%s: %s", e.ResponseHeader.Error, e.ResponseHeader.ErrorMessage)
}
type RPCProtoError struct {
Message string
UnderlyingError error
}
func (e *RPCProtoError) Error() string {
return e.Message
}
func checkRPCParamTypes(in, out reflect.Type) (err error) {
if !(in.Kind() == reflect.Ptr || typeIsIOReader(in)) {
err = errors.Errorf("input parameter must be a pointer or an io.Reader, is of kind %s, type %s", in.Kind(), in)
return
}
if !(out.Kind() == reflect.Ptr) {
err = errors.Errorf("second input parameter (the non-error output parameter) must be a pointer or an *io.Reader")
return
}
return nil
}
func checkRPCReturnType(rt reflect.Type) (err error) {
errInterfaceType := reflect.TypeOf((*error)(nil)).Elem()
if !rt.Implements(errInterfaceType) {
err = errors.Errorf("handler must return an error")
return
}
return nil
}
func makeEndpointDescr(handler interface{}) (descr endpointDescr, err error) {
ht := reflect.TypeOf(handler)
if ht.Kind() != reflect.Func {
err = errors.Errorf("handler must be of kind reflect.Func")
return
}
if ht.NumIn() != 2 || ht.NumOut() != 1 {
err = errors.Errorf("handler must have exactly two input parameters and one output parameter")
return
}
if err = checkRPCParamTypes(ht.In(0), ht.In(1)); err != nil {
return
}
if err = checkRPCReturnType(ht.Out(0)); err != nil {
return
}
descr.handler = reflect.ValueOf(handler)
descr.inType.local = ht.In(0)
descr.outType.local = ht.In(1)
if typeIsIOReader(ht.In(0)) {
descr.inType.proto = DataTypeOctets
} else {
descr.inType.proto = DataTypeMarshaledJSON
}
if typeIsIOReaderPtr(ht.In(1)) {
descr.outType.proto = DataTypeOctets
} else {
descr.outType.proto = DataTypeMarshaledJSON
}
return
}