249 lines
12 KiB
C++
249 lines
12 KiB
C++
|
#include "ReplicationWriteCacheClient.h"
|
||
|
#include "GlobalDataAdaptor.h"
|
||
|
|
||
|
folly::Future<PutResult> ReplicationWriteCacheClient::Put(const std::string &key,
|
||
|
size_t size,
|
||
|
const ByteBuffer &buffer,
|
||
|
const std::map <std::string, std::string> &headers,
|
||
|
size_t off) {
|
||
|
std::vector <folly::Future<PutOutput>> future_list;
|
||
|
Json::Value root, dummy_root;
|
||
|
Json::Value json_replica(Json::arrayValue), json_path(Json::arrayValue), json_headers;
|
||
|
|
||
|
butil::Timer *t = new butil::Timer();
|
||
|
t->start();
|
||
|
|
||
|
const std::vector<int> replicas = GetReplica(key);
|
||
|
for (auto server_id: replicas) {
|
||
|
json_replica.append(server_id);
|
||
|
}
|
||
|
|
||
|
auto rpc_client = parent_->GetRpcClient();
|
||
|
auto write_chunk_size = GetGlobalConfig().write_chunk_size;
|
||
|
|
||
|
for (auto iter = headers.begin(); iter != headers.end(); ++iter) {
|
||
|
json_headers[iter->first] = iter->second;
|
||
|
}
|
||
|
|
||
|
size_t aggregated_size = 0;
|
||
|
for (uint64_t offset = 0; offset < size; offset += write_chunk_size) {
|
||
|
for (auto server_id: replicas) {
|
||
|
auto region_size = std::min(size - offset, write_chunk_size);
|
||
|
ByteBuffer region_buffer(buffer.data + offset, region_size);
|
||
|
std::string partial_key = key
|
||
|
+ "-" + std::to_string((off + offset) / write_chunk_size)
|
||
|
+ "-" + std::to_string(write_chunk_size);
|
||
|
auto PutRPC = folly::via(parent_->executor_.get(), [this, server_id, partial_key, region_buffer, region_size]() -> PutOutput {
|
||
|
return parent_->GetRpcClient()->PutEntryFromWriteCache(server_id, partial_key, region_buffer, region_size).get();
|
||
|
});
|
||
|
future_list.emplace_back(std::move(PutRPC));
|
||
|
}
|
||
|
}
|
||
|
|
||
|
t->stop();
|
||
|
LOG(INFO) << "Phase 1: " << t->u_elapsed();
|
||
|
|
||
|
root["type"] = "replication";
|
||
|
root["size"] = size;
|
||
|
root["replica"] = json_replica;
|
||
|
root["headers"] = json_headers;
|
||
|
|
||
|
return folly::collectAll(future_list).via(parent_->executor_.get()).thenValue([root, t](std::vector<folly::Try<PutOutput>> &&output) -> folly::Future<PutResult> {
|
||
|
Json::Value dummy_root;
|
||
|
Json::Value json_path(Json::arrayValue);
|
||
|
for (auto &entry: output) {
|
||
|
if (!entry.hasValue())
|
||
|
return PutResult { FOLLY_ERROR, dummy_root };
|
||
|
if (entry.value().status != OK) {
|
||
|
LOG(INFO) << "Found error";
|
||
|
return PutResult { entry.value().status, dummy_root };
|
||
|
}
|
||
|
json_path.append(entry.value().internal_key);
|
||
|
}
|
||
|
Json::Value new_root = root;
|
||
|
new_root["path"] = json_path;
|
||
|
t->stop();
|
||
|
LOG(INFO) << "Duration: " << t->u_elapsed();
|
||
|
delete t;
|
||
|
return PutResult { OK, new_root };
|
||
|
});
|
||
|
}
|
||
|
|
||
|
folly::Future<int> ReplicationWriteCacheClient::Get(const std::string &key,
|
||
|
size_t start,
|
||
|
size_t size,
|
||
|
ByteBuffer &buffer,
|
||
|
Json::Value &root) {
|
||
|
std::vector<int> replicas;
|
||
|
for (auto &entry : root["replica"]) {
|
||
|
replicas.push_back(entry.asInt());
|
||
|
}
|
||
|
|
||
|
std::vector<std::string> internal_keys;
|
||
|
for (auto &entry : root["path"]) {
|
||
|
internal_keys.push_back(entry.asString());
|
||
|
}
|
||
|
|
||
|
std::vector <folly::Future<int>> future_list;
|
||
|
std::vector<GetChunkRequestV2> requests;
|
||
|
auto write_chunk_size = GetGlobalConfig().write_chunk_size;
|
||
|
GenerateGetChunkRequestsV2(key, start, size, buffer, requests, write_chunk_size);
|
||
|
if (requests.empty())
|
||
|
return folly::makeFuture(OK);
|
||
|
|
||
|
size_t aggregated_size = 0;
|
||
|
for (auto &entry: requests) {
|
||
|
int primary_replica_id = lrand48() % replicas.size();
|
||
|
int primary_server_id = replicas[primary_replica_id];
|
||
|
std::string internal_key = internal_keys[entry.chunk_id * replicas.size() + primary_replica_id];
|
||
|
future_list.emplace_back(GetChunkAsync(primary_server_id, entry, internal_key)
|
||
|
.thenValue([this, replicas, entry, primary_server_id, internal_keys](int res) -> int {
|
||
|
if (res != RPC_FAILED) {
|
||
|
return res;
|
||
|
}
|
||
|
LOG_EVERY_SECOND(WARNING) << "Unable to connect primary replicas. server_id " << primary_server_id
|
||
|
<< ", hostname: " << parent_->GetServerHostname(primary_server_id);
|
||
|
for (auto &server_id : replicas) {
|
||
|
if (server_id == primary_server_id) {
|
||
|
continue;
|
||
|
}
|
||
|
auto internal_key = internal_keys[entry.chunk_id * replicas.size() + server_id];
|
||
|
res = GetChunkAsync(server_id, entry, internal_key).get();
|
||
|
if (res != RPC_FAILED) {
|
||
|
return res;
|
||
|
}
|
||
|
LOG_EVERY_SECOND(WARNING) << "Unable to connect secondary replicas. server_id " << server_id
|
||
|
<< ", hostname: " << parent_->GetServerHostname(server_id);
|
||
|
}
|
||
|
LOG_EVERY_SECOND(ERROR) << "Unable to connect all target replicas";
|
||
|
return RPC_FAILED;
|
||
|
}));
|
||
|
|
||
|
aggregated_size += entry.chunk_len;
|
||
|
if (aggregated_size >= GetGlobalConfig().max_inflight_payload_size) {
|
||
|
auto output = folly::collectAll(future_list).get();
|
||
|
for (auto &entry: output)
|
||
|
if (entry.value_or(FOLLY_ERROR) != OK) {
|
||
|
LOG(ERROR) << "Failed to get data from write cache, key: " << key
|
||
|
<< ", start: " << start
|
||
|
<< ", size: " << size
|
||
|
<< ", buf: " << (void *) buffer.data << " " << buffer.len
|
||
|
<< ", error code: " << entry.hasValue() << " " << entry.value_or(FOLLY_ERROR);
|
||
|
return entry.value_or(FOLLY_ERROR);
|
||
|
}
|
||
|
aggregated_size = 0;
|
||
|
future_list.clear();
|
||
|
}
|
||
|
}
|
||
|
|
||
|
return folly::collectAll(future_list).via(parent_->executor_.get()).thenValue(
|
||
|
[=](std::vector <folly::Try<int>> output) -> int {
|
||
|
for (auto &entry: output)
|
||
|
if (entry.value_or(FOLLY_ERROR) != OK) {
|
||
|
LOG(ERROR) << "Failed to get data from write cache, key: " << key
|
||
|
<< ", start: " << start
|
||
|
<< ", size: " << size
|
||
|
<< ", buf: " << (void *) buffer.data << " " << buffer.len
|
||
|
<< ", error code: " << entry.hasValue() << " " << entry.value_or(FOLLY_ERROR);
|
||
|
return entry.value_or(FOLLY_ERROR);
|
||
|
}
|
||
|
return OK;
|
||
|
});
|
||
|
|
||
|
// return parent_->GetRpcClient()->GetEntryFromWriteCache(replica[primary_index], internal_keys[primary_index], start, size).thenValue(
|
||
|
// [this, &buffer, start, size, replica, internal_keys, primary_index](GetOutput &&output) -> int {
|
||
|
// if (output.status == OK) {
|
||
|
// output.buf.copy_to(buffer.data, size);
|
||
|
// }
|
||
|
// if (output.status == RPC_FAILED) {
|
||
|
// for (int index = 0; index < replica.size(); ++index) {
|
||
|
// if (index == primary_index) {
|
||
|
// continue;
|
||
|
// }
|
||
|
// auto res = parent_->GetRpcClient()->GetEntryFromWriteCache(replica[index], internal_keys[index], start, size).get();
|
||
|
// if (res.status == OK) {
|
||
|
// output.buf.copy_to(buffer.data, size);
|
||
|
// }
|
||
|
// if (res.status != RPC_FAILED) {
|
||
|
// return res.status;
|
||
|
// }
|
||
|
// }
|
||
|
// LOG(ERROR) << "All target replicas are crashed";
|
||
|
// return RPC_FAILED;
|
||
|
// }
|
||
|
// return output.status;
|
||
|
// }
|
||
|
// );
|
||
|
}
|
||
|
|
||
|
folly::Future<int> ReplicationWriteCacheClient::GetChunkAsync(int server_id, GetChunkRequestV2 request, std::string &internal_key) {
|
||
|
LOG_IF(INFO, FLAGS_verbose) << "GetChunkAsync server_id=" << server_id
|
||
|
<< ", internal_key=" << internal_key
|
||
|
<< ", chunk_id=" << request.chunk_id
|
||
|
<< ", chunk_start=" << request.chunk_start
|
||
|
<< ", chunk_len=" << request.chunk_len
|
||
|
<< ", buffer=" << (void *) request.buffer.data;
|
||
|
auto f = parent_->GetRpcClient()->GetEntryFromWriteCache(server_id, internal_key, request.chunk_start, request.chunk_len)
|
||
|
.then([this, server_id, request](folly::Try<GetOutput> &&output) -> folly::Future<int> {
|
||
|
if (!output.hasValue()) {
|
||
|
return folly::makeFuture(FOLLY_ERROR);
|
||
|
}
|
||
|
auto &value = output.value();
|
||
|
if (value.status == OK) {
|
||
|
value.buf.copy_to(request.buffer.data, request.buffer.len);
|
||
|
return folly::makeFuture(OK);
|
||
|
} else {
|
||
|
return folly::makeFuture(value.status);
|
||
|
}
|
||
|
}).via(parent_->executor_.get());
|
||
|
return f;
|
||
|
// memset(request.buffer.data, 'x', request.buffer.len);
|
||
|
// return folly::makeFuture(OK);
|
||
|
}
|
||
|
|
||
|
std::vector<int> ReplicationWriteCacheClient::GetReplica(const std::string &key) {
|
||
|
const int num_available = parent_->server_list_.size();
|
||
|
auto &policy = parent_->GetCachePolicy(key);
|
||
|
const int num_choose = policy.write_replication_factor;
|
||
|
uint64_t seed = std::hash < std::string > {}(key);
|
||
|
std::vector<int> output;
|
||
|
for (int i = 0; i < num_choose; ++i)
|
||
|
output.push_back((seed + i) % num_available);
|
||
|
return output;
|
||
|
}
|
||
|
|
||
|
void ReplicationWriteCacheClient::GenerateGetChunkRequestsV2(const std::string &key,
|
||
|
size_t start,
|
||
|
size_t size,
|
||
|
ByteBuffer &buffer,
|
||
|
std::vector<GetChunkRequestV2> &requests,
|
||
|
size_t chunk_size) {
|
||
|
const size_t end = start + size;
|
||
|
|
||
|
const size_t begin_chunk_id = start / chunk_size;
|
||
|
const size_t end_chunk_id = (end + chunk_size - 1) / chunk_size;
|
||
|
|
||
|
if (buffer.len < size) {
|
||
|
LOG(WARNING) << "Buffer capacity may be not enough, expect " << size << ", actual " << buffer.len;
|
||
|
}
|
||
|
|
||
|
size_t buffer_offset = 0;
|
||
|
for (size_t chunk_id = begin_chunk_id; chunk_id < end_chunk_id; ++chunk_id) {
|
||
|
size_t chunk_start = std::max(chunk_id * chunk_size, start);
|
||
|
size_t chunk_stop = std::min((chunk_id + 1) * chunk_size, end);
|
||
|
if (chunk_stop <= chunk_start)
|
||
|
return;
|
||
|
GetChunkRequestV2 item;
|
||
|
item.user_key = key;
|
||
|
item.chunk_id = chunk_id;
|
||
|
item.chunk_start = chunk_start % chunk_size;
|
||
|
item.chunk_len = chunk_stop - chunk_start;
|
||
|
item.chunk_granularity = chunk_size;
|
||
|
item.buffer.data = buffer.data + buffer_offset;
|
||
|
item.buffer.len = item.chunk_len;
|
||
|
buffer_offset += item.chunk_len;
|
||
|
requests.emplace_back(item);
|
||
|
}
|
||
|
LOG_ASSERT(buffer_offset == size);
|
||
|
}
|