Files
hpcore/src/hpfs/hpfs_sync.cpp
Savinda Senevirathne bed8205ca8 Config patch change detection. (#211)
* Applying patch file changes to hpcore runtime after patch file change detection.
* Removing unl sync functionality.
* Removing subjecting unl changeset to consensus.
2021-01-06 20:02:14 +05:30

649 lines
27 KiB
C++

#include "../msg/fbuf/p2pmsg_helpers.hpp"
#include "../msg/fbuf/p2pmsg_content_generated.h"
#include "../msg/fbuf/common_helpers.hpp"
#include "../p2p/p2p.hpp"
#include "../pchheader.hpp"
#include "../ledger.hpp"
#include "../hplog.hpp"
#include "../util/util.hpp"
#include "../hpfs/hpfs.hpp"
#include "../util/h32.hpp"
#include "hpfs_sync.hpp"
#include "../sc.hpp"
#include "../unl.hpp"
namespace hpfs_sync
{
// Idle loop sleep time (milliseconds).
constexpr uint16_t IDLE_WAIT = 40;
// Max number of requests that can be awaiting response at any given time.
constexpr uint16_t MAX_AWAITING_REQUESTS = 4;
// Request loop sleep time (milliseconds).
constexpr uint16_t REQUEST_LOOP_WAIT = 10;
// Max no. of repetitive reqeust resubmissions before abandoning the sync.
constexpr uint16_t ABANDON_THRESHOLD = 20;
constexpr int FILE_PERMS = 0644;
// No. of milliseconds to wait before resubmitting a request.
uint16_t REQUEST_RESUBMIT_TIMEOUT;
sync_context ctx;
bool init_success = false;
int init()
{
REQUEST_RESUBMIT_TIMEOUT = hpfs::get_request_resubmit_timeout();
ctx.target_state_hash = util::h32_empty;
ctx.target_patch_hash = util::h32_empty;
ctx.current_parent_target_hash = util::h32_empty;
// Patch file sync has the highest priority.
ctx.current_syncing_parent = hpfs::HPFS_PARENT_COMPONENTS::PATCH;
ctx.hpfs_sync_thread = std::thread(hpfs_syncer_loop);
ctx.hpfs_mount_dir = conf::ctx.hpfs_rw_dir;
init_success = true;
return 0;
}
void deinit()
{
if (init_success)
{
ctx.is_syncing = false;
ctx.is_shutting_down = true;
ctx.hpfs_sync_thread.join();
}
}
/**
* Sets a new target states for the syncing process.
* @param target_state_hash The target hpfs state which we should sync towards.
* @param target_patch_hash The target hpfs patch state which we should sync towards.
*/
void set_target(const util::h32 target_state_hash, const util::h32 target_patch_hash)
{
std::unique_lock lock(ctx.target_state_mutex);
// Do not do anything if we are already syncing towards the specified target states.
if (ctx.is_shutting_down || (ctx.is_syncing && ctx.target_state_hash == target_state_hash && ctx.target_patch_hash == target_patch_hash))
return;
ctx.target_state_hash = target_state_hash;
ctx.target_patch_hash = target_patch_hash;
if (hpfs::ctx.get_hash(hpfs::HPFS_PARENT_COMPONENTS::PATCH) != target_patch_hash)
{
ctx.current_syncing_parent = hpfs::HPFS_PARENT_COMPONENTS::PATCH;
ctx.current_parent_target_hash = ctx.target_patch_hash;
}
else
{
ctx.current_syncing_parent = hpfs::HPFS_PARENT_COMPONENTS::STATE;
ctx.current_parent_target_hash = ctx.target_state_hash;
}
ctx.is_syncing = true;
}
/**
* Runs the hpfs sync worker loop.
*/
void hpfs_syncer_loop()
{
util::mask_signal();
LOG_INFO << "hpfs sync: Worker started.";
while (!ctx.is_shutting_down)
{
util::sleep(IDLE_WAIT);
// Keep idling if we are not doing any sync activity.
if (!ctx.is_syncing)
continue;
if (hpfs::start_fs_session(ctx.hpfs_mount_dir) != -1)
{
while (!ctx.is_shutting_down)
{
{
std::shared_lock lock(ctx.target_state_mutex);
if (ctx.current_syncing_parent == hpfs::HPFS_PARENT_COMPONENTS::PATCH)
LOG_INFO << "hpfs sync: Starting sync for target patch hash: " << ctx.target_patch_hash;
else
LOG_INFO << "hpfs sync: Starting sync for target state hash: " << ctx.target_state_hash;
}
util::h32 new_state = util::h32_empty;
const int result = request_loop(ctx.current_parent_target_hash, new_state);
ctx.pending_requests.clear();
ctx.candidate_hpfs_responses.clear();
ctx.submitted_requests.clear();
if (result == -1 || ctx.is_shutting_down)
break;
{
std::shared_lock lock(ctx.target_state_mutex);
if (new_state == ctx.current_parent_target_hash)
{
if (ctx.current_syncing_parent == hpfs::HPFS_PARENT_COMPONENTS::PATCH)
{
ctx.target_patch_hash = util::h32_empty;
LOG_INFO << "hpfs sync: Target patch state achieved: " << new_state;
// Appling new patch file changes to hpcore runtime.
if (conf::validate_and_apply_patch_config(conf::cfg.contract, conf::ctx.hpfs_rw_dir) == -1)
{
LOG_ERROR << "Appling patch file changes after sync failed";
}
else
{
unl::update_unl_changes_from_patch();
// Update global hash tracker with the new patch file hash.
util::h32 updated_patch_hash;
hpfs::get_hash(updated_patch_hash, conf::ctx.hpfs_rw_dir, conf::PATCH_FILE_PATH);
hpfs::ctx.set_hash(hpfs::HPFS_PARENT_COMPONENTS::PATCH, updated_patch_hash);
}
if (ctx.target_state_hash == hpfs::ctx.get_hash(hpfs::HPFS_PARENT_COMPONENTS::STATE))
break;
ctx.current_parent_target_hash = ctx.target_state_hash;
ctx.current_syncing_parent = hpfs::HPFS_PARENT_COMPONENTS::STATE;
continue;
}
else if (ctx.current_syncing_parent == hpfs::HPFS_PARENT_COMPONENTS::STATE)
{
ctx.target_state_hash = util::h32_empty;
LOG_INFO << "hpfs sync: Target state achieved: " << new_state;
break;
}
}
else
{
LOG_INFO << "hpfs sync: Continuing sync for new target: " << ctx.current_parent_target_hash;
continue;
}
}
}
LOG_INFO << "hpfs sync: All parents synced.";
hpfs::stop_fs_session(ctx.hpfs_mount_dir);
}
else
{
LOG_ERROR << "hpfs sync: Failed to start hpfs rw session";
}
std::unique_lock lock(ctx.target_state_mutex);
ctx.current_parent_target_hash = util::h32_empty;
ctx.is_syncing = false;
}
LOG_INFO << "hpfs sync: Worker stopped.";
}
int request_loop(const util::h32 current_target, util::h32 &updated_state)
{
std::string target_parent_vpath;
BACKLOG_ITEM_TYPE target_parent_backlog_item_type;
if (ctx.current_syncing_parent == hpfs::HPFS_PARENT_COMPONENTS::STATE)
{
target_parent_vpath = sc::STATE_DIR_PATH;
target_parent_backlog_item_type = BACKLOG_ITEM_TYPE::DIR;
}
else if (ctx.current_syncing_parent == hpfs::HPFS_PARENT_COMPONENTS::PATCH)
{
target_parent_vpath = conf::PATCH_FILE_PATH;
target_parent_backlog_item_type = BACKLOG_ITEM_TYPE::FILE;
}
std::string lcl = ledger::ctx.get_lcl();
// Indicates whether any responses were processed in the previous loop iteration.
bool prev_responses_processed = false;
// No. of repetitive resubmissions so far. (This is reset whenever we receive a hpfs response)
uint16_t resubmissions_count = 0;
// Send the initial root hpfs request.
submit_request(backlog_item{target_parent_backlog_item_type, target_parent_vpath, -1, current_target}, lcl);
while (!should_stop_request_loop(current_target))
{
// Wait a small delay if there were no responses processed during previous iteration.
if (!prev_responses_processed)
util::sleep(REQUEST_LOOP_WAIT);
// Get current lcl.
std::string lcl = ledger::ctx.get_lcl();
{
std::scoped_lock lock(p2p::ctx.collected_msgs.hpfs_responses_mutex);
// Move collected hpfs responses over to local candidate responses list.
if (!p2p::ctx.collected_msgs.hpfs_responses.empty())
ctx.candidate_hpfs_responses.splice(ctx.candidate_hpfs_responses.end(), p2p::ctx.collected_msgs.hpfs_responses);
}
prev_responses_processed = !ctx.candidate_hpfs_responses.empty();
// Reset resubmissions counter whenever we have a resposne.
if (!ctx.candidate_hpfs_responses.empty())
resubmissions_count = 0;
for (auto &response : ctx.candidate_hpfs_responses)
{
if (should_stop_request_loop(current_target))
return 0;
LOG_DEBUG << "hpfs sync: Processing hpfs response from [" << response.first.substr(2, 10) << "]";
const msg::fbuf::p2pmsg::Content *content = msg::fbuf::p2pmsg::GetContent(response.second.data());
const msg::fbuf::p2pmsg::Hpfs_Response_Message *resp_msg = content->message_as_Hpfs_Response_Message();
// Check whether we are actually waiting for this response. If not, ignore it.
std::string_view hash = msg::fbuf::flatbuff_bytes_to_sv(resp_msg->hash());
std::string_view vpath = msg::fbuf::flatbuff_str_to_sv(resp_msg->path());
const std::string key = std::string(vpath).append(hash);
const auto pending_resp_itr = ctx.submitted_requests.find(key);
if (pending_resp_itr == ctx.submitted_requests.end())
{
LOG_DEBUG << "hpfs sync: Skipping hpfs response due to hash mismatch.";
continue;
}
// Process the message based on response type.
const msg::fbuf::p2pmsg::Hpfs_Response msg_type = resp_msg->hpfs_response_type();
if (msg_type == msg::fbuf::p2pmsg::Hpfs_Response_Fs_Entry_Response)
{
const msg::fbuf::p2pmsg::Fs_Entry_Response *fs_resp = resp_msg->hpfs_response_as_Fs_Entry_Response();
// Get fs entries we have received.
std::unordered_map<std::string, p2p::hpfs_fs_hash_entry> peer_fs_entry_map;
msg::fbuf::p2pmsg::flatbuf_hpfsfshashentry_to_hpfsfshashentry(peer_fs_entry_map, fs_resp->entries());
// Validate received fs data against the hash.
if (!validate_fs_entry_hash(vpath, hash, peer_fs_entry_map))
{
LOG_INFO << "hpfs sync: Skipping hpfs response due to fs entry hash mismatch.";
continue;
}
handle_fs_entry_response(vpath, peer_fs_entry_map);
}
else if (msg_type == msg::fbuf::p2pmsg::Hpfs_Response_File_HashMap_Response)
{
const msg::fbuf::p2pmsg::File_HashMap_Response *file_resp = resp_msg->hpfs_response_as_File_HashMap_Response();
// File block hashes we received from the peer.
const util::h32 *peer_hashes = reinterpret_cast<const util::h32 *>(file_resp->hash_map()->data());
const size_t peer_hash_count = file_resp->hash_map()->size() / sizeof(util::h32);
// Validate received hashmap against the hash.
if (!validate_file_hashmap_hash(vpath, hash, peer_hashes, peer_hash_count))
{
LOG_INFO << "hpfs sync: Skipping hpfs response due to file hashmap hash mismatch.";
continue;
}
handle_file_hashmap_response(vpath, peer_hashes, peer_hash_count, file_resp->file_length());
}
else if (msg_type == msg::fbuf::p2pmsg::Hpfs_Response_Block_Response)
{
const msg::fbuf::p2pmsg::Block_Response *block_resp = resp_msg->hpfs_response_as_Block_Response();
// Get the file path of the block data we have received.
const uint32_t block_id = block_resp->block_id();
std::string_view buf = msg::fbuf::flatbuff_bytes_to_sv(block_resp->data());
// Validate received block data against the hash.
if (!validate_file_block_hash(hash, block_id, buf))
{
LOG_INFO << "hpfs sync: Skipping hpfs response due to file block hash mismatch.";
continue;
}
handle_file_block_response(vpath, block_id, buf);
}
// Now that we have received matching hash and handled it, remove it from the waiting list.
ctx.submitted_requests.erase(pending_resp_itr);
// After handling each response, check whether we have reached target hpfs state.
// get_hash returns 0 incase target parent is not existing in our side.
if (hpfs::get_hash(updated_state, ctx.hpfs_mount_dir, target_parent_vpath) == -1)
{
LOG_ERROR << "hpfs sync: exiting due to hash check error.";
return -1;
}
// Update the central hpfs state tracker.
hpfs::ctx.set_hash(ctx.current_syncing_parent, updated_state);
LOG_DEBUG << "hpfs sync: current:" << updated_state << " | target:" << current_target;
if (updated_state == current_target)
return 0;
}
ctx.candidate_hpfs_responses.clear();
// Check for long-awaited responses and re-request them.
for (auto &[hash, request] : ctx.submitted_requests)
{
if (should_stop_request_loop(current_target))
return 0;
if (request.waiting_time < REQUEST_RESUBMIT_TIMEOUT)
{
// Increment wait time.
request.waiting_time += REQUEST_LOOP_WAIT;
}
else
{
if (++resubmissions_count > ABANDON_THRESHOLD)
{
LOG_INFO << "hpfs sync: Resubmission threshold exceeded. Abandoning sync.";
return -1;
}
// Reset the counter and re-submit request.
request.waiting_time = 0;
LOG_DEBUG << "hpfs sync: Resubmitting request...";
submit_request(request, lcl);
}
}
// Check whether we can submit any more requests.
if (!ctx.pending_requests.empty() && ctx.submitted_requests.size() < MAX_AWAITING_REQUESTS)
{
const uint16_t available_slots = MAX_AWAITING_REQUESTS - ctx.submitted_requests.size();
for (int i = 0; i < available_slots && !ctx.pending_requests.empty(); i++)
{
if (should_stop_request_loop(current_target))
return 0;
const backlog_item &request = ctx.pending_requests.front();
submit_request(request, lcl);
ctx.pending_requests.pop_front();
}
}
}
return 0;
}
/**
* Vadidated the received hash against the received fs entry map.
* @param vpath Virtual path of the fs.
* @param hash Received hash.
* @param fs_entry_map Received fs entry map.
* @returns true if hash is valid, otherwise false.
*/
bool validate_fs_entry_hash(std::string_view vpath, std::string_view hash, const std::unordered_map<std::string, p2p::hpfs_fs_hash_entry> &fs_entry_map)
{
util::h32 content_hash;
const std::string vpath_name = util::get_name(vpath);
// Initilal hash is vpath hash.
content_hash = crypto::get_hash(vpath_name);
// Then XOR the file hashes to the initial hash.
for (const auto &[name, fs_entry] : fs_entry_map)
{
content_hash ^= fs_entry.hash;
}
return content_hash.to_string_view() == hash;
}
/**
* Vadidated the received hash against the received file hash map.
* @param vpath Virtual path of the file.
* @param hash Received hash.
* @param hashes Received block hashes.
* @param hash_count Size of the hash list.
* @returns true if hash is valid, otherwise false.
*/
bool validate_file_hashmap_hash(std::string_view vpath, std::string_view hash, const util::h32 *hashes, const size_t hash_count)
{
util::h32 content_hash = util::h32_empty;
const std::string vpath_name = util::get_name(vpath);
// Initilal hash is vpath hash.
content_hash = crypto::get_hash(vpath_name);
// Then XOR the block hashes to the initial hash.
for (int32_t block_id = 0; block_id < hash_count; block_id++)
{
content_hash ^= hashes[block_id];
}
return content_hash.to_string_view() == hash;
}
/**
* Vadidated the received hash against the received block.
* @param hash Received hash.
* @param block_id Id of the block.
* @param buf Block buffer.
* @returns true if hash is valid, otherwise false.
*/
bool validate_file_block_hash(std::string_view hash, const uint32_t block_id, std::string_view buf)
{
// Calculate block offset of this block.
const off_t block_offset = block_id * hpfs::BLOCK_SIZE;
std::string_view offset = std::string_view(reinterpret_cast<const char *>(&block_offset), sizeof(off_t));
return crypto::get_hash(offset, buf) == hash;
}
/**
* Indicates whether to break out of hpfs request processing loop.
*/
bool should_stop_request_loop(const util::h32 current_target)
{
if (ctx.is_shutting_down)
return true;
// Stop request loop if the target has changed.
std::shared_lock lock(ctx.target_state_mutex);
return current_target != ctx.current_parent_target_hash;
}
/**
* Sends a hpfs request to a random peer.
* @param path Requested file or dir path.
* @param is_file Whether the requested path if a file or dir.
* @param block_id The requested block id. Only relevant if requesting a file block. Otherwise -1.
* @param expected_hash The expected hash of the requested data. The peer will ignore the request if their hash is different.
* @param target_pubkey The peer pubkey the request was submitted to.
*/
void request_state_from_peer(const std::string &path, const bool is_file, const int32_t block_id,
const util::h32 expected_hash, std::string_view lcl, std::string &target_pubkey)
{
p2p::hpfs_request hr;
hr.parent_path = path;
hr.is_file = is_file;
hr.block_id = block_id;
hr.expected_hash = expected_hash;
flatbuffers::FlatBufferBuilder fbuf(1024);
msg::fbuf::p2pmsg::create_msg_from_state_request(fbuf, hr, lcl);
p2p::send_message_to_random_peer(fbuf, target_pubkey); //todo: send to a node that hold the majority hpfs state to improve reliability of retrieving hpfs state.
}
/**
* Submits a pending hpfs request to the peer.
*/
void submit_request(const backlog_item &request, std::string_view lcl)
{
const std::string key = std::string(request.path)
.append(reinterpret_cast<const char *>(&request.expected_hash), sizeof(util::h32));
ctx.submitted_requests.try_emplace(key, request);
const bool is_file = request.type != BACKLOG_ITEM_TYPE::DIR;
std::string target_pubkey;
request_state_from_peer(request.path, is_file, request.block_id, request.expected_hash, lcl, target_pubkey);
if (!target_pubkey.empty())
LOG_DEBUG << "hpfs sync: Requesting from [" << target_pubkey.substr(2, 10) << "]. type:" << request.type
<< " path:" << request.path << " block_id:" << request.block_id
<< " hash:" << request.expected_hash;
}
/**
* Process dir children response.
* @param vpath Virtual path of the fs.
* @param fs_entry_map Received fs entry map.
* @returns 0 on success, otherwise -1.
*/
int handle_fs_entry_response(std::string_view vpath, std::unordered_map<std::string, p2p::hpfs_fs_hash_entry> &fs_entry_map)
{
// Get the parent path of the fs entries we have received.
LOG_DEBUG << "hpfs sync: Processing fs entries response for " << vpath;
// Create physical directory on our side if not exist.
std::string parent_physical_path = std::string(ctx.hpfs_mount_dir).append(vpath);
if (util::create_dir_tree_recursive(parent_physical_path) == -1)
return -1;
// Get the children hash entries and compare with what we got from peer.
std::vector<hpfs::child_hash_node> existing_fs_entries;
if (hpfs::get_dir_children_hashes(existing_fs_entries, ctx.hpfs_mount_dir, vpath) == -1)
return -1;
// Request more info on fs entries that exist on both sides but are different.
for (const auto &ex_entry : existing_fs_entries)
{
// Construct child vpath.
std::string child_vpath = std::string(vpath)
.append(vpath.back() != '/' ? "/" : "")
.append(ex_entry.name);
const auto peer_itr = fs_entry_map.find(ex_entry.name);
if (peer_itr != fs_entry_map.end())
{
// Request hpfs state if hash is different.
if (peer_itr->second.hash != ex_entry.hash)
{
// Prioritize file hpfs requests over directories.
if (ex_entry.is_file)
ctx.pending_requests.push_front(backlog_item{BACKLOG_ITEM_TYPE::FILE, child_vpath, -1, peer_itr->second.hash});
else
ctx.pending_requests.push_back(backlog_item{BACKLOG_ITEM_TYPE::DIR, child_vpath, -1, peer_itr->second.hash});
}
fs_entry_map.erase(peer_itr);
}
else
{
// If there was an entry that does not exist on other side, delete it.
std::string child_physical_path = std::string(ctx.hpfs_mount_dir).append(child_vpath);
if ((ex_entry.is_file && unlink(child_physical_path.c_str()) == -1) ||
!ex_entry.is_file && util::remove_directory_recursively(child_physical_path.c_str()) == -1)
return -1;
LOG_DEBUG << "hpfs sync: Deleted " << (ex_entry.is_file ? "file" : "dir") << " path " << child_vpath;
}
}
// Queue the remaining peer fs entries (that our side does not have at all) to request.
for (const auto &[name, fs_entry] : fs_entry_map)
{
// Construct child vpath.
std::string child_vpath = std::string(vpath)
.append(vpath.back() != '/' ? "/" : "")
.append(name);
// Prioritize file hpfs requests over directories.
if (fs_entry.is_file)
ctx.pending_requests.push_front(backlog_item{BACKLOG_ITEM_TYPE::FILE, child_vpath, -1, fs_entry.hash});
else
ctx.pending_requests.push_back(backlog_item{BACKLOG_ITEM_TYPE::DIR, child_vpath, -1, fs_entry.hash});
}
return 0;
}
/**
* Process file block hash map response.
* @param vpath Virtual path of the file.
* @param hash Received hash.
* @param hashes Received block hashes.
* @param file_length Size of the file.
* @returns 0 on success, otherwise -1.
*/
int handle_file_hashmap_response(std::string_view vpath, const util::h32 *hashes, const size_t hash_count, const uint64_t file_length)
{
// Get the file path of the block hashes we have received.
LOG_DEBUG << "hpfs sync: Processing file block hashes response for " << vpath;
// File block hashes on our side (file might not exist on our side).
std::vector<util::h32> existing_hashes;
if (hpfs::get_file_block_hashes(existing_hashes, ctx.hpfs_mount_dir, vpath) == -1 && errno != ENOENT)
return -1;
const size_t existing_hash_count = existing_hashes.size();
// Compare the block hashes and request any differences.
auto insert_itr = ctx.pending_requests.begin();
const int32_t max_block_id = MAX(existing_hash_count, hash_count) - 1;
for (int32_t block_id = 0; block_id <= max_block_id; block_id++)
{
// Insert at front to give priority to block requests while preserving block order.
if (block_id >= existing_hash_count || existing_hashes[block_id] != hashes[block_id])
ctx.pending_requests.insert(insert_itr, backlog_item{BACKLOG_ITEM_TYPE::BLOCK, std::string(vpath), block_id, hashes[block_id]});
}
if (existing_hashes.size() >= hash_count)
{
// If peer file might be smaller, truncate our file to match with peer file.
std::string file_physical_path = std::string(ctx.hpfs_mount_dir).append(vpath);
if (truncate(file_physical_path.c_str(), file_length) == -1)
return -1;
}
return 0;
}
/**
* Process file block response.
* @param vpath Virtual path of the file.
* @param block_id Id of the block.
* @param buf Block buffer.
* @returns 0 on success, otherwise -1.
*/
int handle_file_block_response(std::string_view vpath, const uint32_t block_id, std::string_view buf)
{
LOG_DEBUG << "hpfs sync: Writing block_id " << block_id
<< " (len:" << buf.length()
<< ") of " << vpath;
std::string file_physical_path = std::string(ctx.hpfs_mount_dir).append(vpath);
const int fd = open(file_physical_path.c_str(), O_WRONLY | O_CREAT | O_CLOEXEC, FILE_PERMS);
if (fd == -1)
{
LOG_ERROR << errno << " Open failed " << file_physical_path;
return -1;
}
const off_t offset = block_id * hpfs::BLOCK_SIZE;
const int res = pwrite(fd, buf.data(), buf.length(), offset);
close(fd);
if (res < buf.length())
{
LOG_ERROR << errno << " Write failed " << file_physical_path;
return -1;
}
return 0;
}
} // namespace hpfs_sync