memgraph/src/distributed/plan_dispatcher.hpp
Matej Ferencevic 53c405c699 Throw exceptions on RPC failure and Distributed error handling
Summary:
This diff changes the RPC layer to directly return `TResponse` to the user when
issuing a `Call<...>` RPC call. The call throws an exception on failure
(instead of the previous return `nullopt`).

All servers (network, RPC and distributed) are set to have explicit `Shutdown`
methods so that a controlled shutdown can always be performed. The object
destructors now have `CHECK`s to enforce that the `AwaitShutdown` methods were
called.

The distributed memgraph is changed that none of the binaries (master/workers)
crash when there is a communication failure. Instead, the whole cluster starts
a graceful shutdown when a persistent communication error is detected.
Transient errors are allowed during execution. The transaction that errored out
will be aborted on the whole cluster. The cluster state is managed using a new
Heartbeat RPC call.

Reviewers: buda, teon.banek, msantl

Reviewed By: teon.banek

Subscribers: pullbot

Differential Revision: https://phabricator.memgraph.io/D1604
2018-09-27 16:27:40 +02:00

30 lines
841 B
C++

#pragma once
#include "distributed/coordination.hpp"
#include "distributed/plan_rpc_messages.hpp"
#include "query/frontend/semantic/symbol_table.hpp"
#include "query/plan/operator.hpp"
namespace distributed {
/** Handles plan dispatching to all workers. Uses MasterCoordination to
* acomplish that. Master side.
*/
class PlanDispatcher {
public:
explicit PlanDispatcher(Coordination *coordination);
/** Dispatch a plan to all workers and wait for their acknowledgement. */
void DispatchPlan(int64_t plan_id,
std::shared_ptr<query::plan::LogicalOperator> plan,
const query::SymbolTable &symbol_table);
/** Remove a plan from all workers and wait for their acknowledgement. */
void RemovePlan(int64_t plan_id);
private:
Coordination *coordination_;
};
} // namespace distributed