mirror of
https://github.com/ClickHouse/ClickHouse.git
synced 2024-12-16 11:22:12 +00:00
234 lines
7.4 KiB
C++
234 lines
7.4 KiB
C++
#include <Client/ConnectionEstablisher.h>
|
|
#include <Common/quoteString.h>
|
|
#include <Common/ProfileEvents.h>
|
|
|
|
namespace ProfileEvents
|
|
{
|
|
extern const Event DistributedConnectionTries;
|
|
extern const Event DistributedConnectionUsable;
|
|
extern const Event DistributedConnectionMissingTable;
|
|
extern const Event DistributedConnectionStaleReplica;
|
|
}
|
|
|
|
namespace DB
|
|
{
|
|
|
|
namespace ErrorCodes
|
|
{
|
|
extern const int ATTEMPT_TO_READ_AFTER_EOF;
|
|
extern const int DNS_ERROR;
|
|
extern const int NETWORK_ERROR;
|
|
extern const int SOCKET_TIMEOUT;
|
|
}
|
|
|
|
ConnectionEstablisher::ConnectionEstablisher(
|
|
ConnectionPoolPtr pool_,
|
|
const ConnectionTimeouts * timeouts_,
|
|
const Settings & settings_,
|
|
LoggerPtr log_,
|
|
const QualifiedTableName * table_to_check_)
|
|
: pool(std::move(pool_)), timeouts(timeouts_), settings(settings_), log(log_), table_to_check(table_to_check_)
|
|
{
|
|
}
|
|
|
|
void ConnectionEstablisher::run(ConnectionEstablisher::TryResult & result, std::string & fail_message)
|
|
{
|
|
try
|
|
{
|
|
ProfileEvents::increment(ProfileEvents::DistributedConnectionTries);
|
|
result.entry = pool->get(*timeouts, settings, /* force_connected = */ false);
|
|
AsyncCallbackSetter async_setter(&*result.entry, std::move(async_callback));
|
|
|
|
UInt64 server_revision = 0;
|
|
if (table_to_check)
|
|
server_revision = result.entry->getServerRevision(*timeouts);
|
|
|
|
if (!table_to_check || server_revision < DBMS_MIN_REVISION_WITH_TABLES_STATUS)
|
|
{
|
|
result.entry->forceConnected(*timeouts);
|
|
ProfileEvents::increment(ProfileEvents::DistributedConnectionUsable);
|
|
result.is_usable = true;
|
|
result.is_up_to_date = true;
|
|
return;
|
|
}
|
|
|
|
/// Only status of the remote table corresponding to the Distributed table is taken into account.
|
|
/// TODO: request status for joined tables also.
|
|
TablesStatusRequest status_request;
|
|
status_request.tables.emplace(*table_to_check);
|
|
|
|
TablesStatusResponse status_response = result.entry->getTablesStatus(*timeouts, status_request);
|
|
auto table_status_it = status_response.table_states_by_id.find(*table_to_check);
|
|
if (table_status_it == status_response.table_states_by_id.end())
|
|
{
|
|
LOG_WARNING(LogToStr(fail_message, log), "There is no table {}.{} on server: {}",
|
|
backQuote(table_to_check->database), backQuote(table_to_check->table), result.entry->getDescription());
|
|
ProfileEvents::increment(ProfileEvents::DistributedConnectionMissingTable);
|
|
return;
|
|
}
|
|
|
|
ProfileEvents::increment(ProfileEvents::DistributedConnectionUsable);
|
|
result.is_usable = true;
|
|
|
|
const UInt64 max_allowed_delay = settings.max_replica_delay_for_distributed_queries;
|
|
if (!max_allowed_delay)
|
|
{
|
|
result.is_up_to_date = true;
|
|
return;
|
|
}
|
|
|
|
const UInt32 delay = table_status_it->second.absolute_delay;
|
|
if (delay < max_allowed_delay)
|
|
result.is_up_to_date = true;
|
|
else
|
|
{
|
|
result.is_up_to_date = false;
|
|
result.delay = delay;
|
|
|
|
LOG_TRACE(log, "Server {} has unacceptable replica delay for table {}.{}: {}", result.entry->getDescription(), table_to_check->database, table_to_check->table, delay);
|
|
ProfileEvents::increment(ProfileEvents::DistributedConnectionStaleReplica);
|
|
}
|
|
}
|
|
catch (const Exception & e)
|
|
{
|
|
if (e.code() != ErrorCodes::NETWORK_ERROR && e.code() != ErrorCodes::SOCKET_TIMEOUT
|
|
&& e.code() != ErrorCodes::ATTEMPT_TO_READ_AFTER_EOF && e.code() != ErrorCodes::DNS_ERROR)
|
|
throw;
|
|
|
|
fail_message = getCurrentExceptionMessage(/* with_stacktrace = */ false);
|
|
|
|
if (!result.entry.isNull())
|
|
{
|
|
result.entry->disconnect();
|
|
result.reset();
|
|
}
|
|
}
|
|
}
|
|
|
|
#if defined(OS_LINUX)
|
|
|
|
ConnectionEstablisherAsync::ConnectionEstablisherAsync(
|
|
ConnectionPoolPtr pool_,
|
|
const ConnectionTimeouts * timeouts_,
|
|
const Settings & settings_,
|
|
LoggerPtr log_,
|
|
const QualifiedTableName * table_to_check_)
|
|
: AsyncTaskExecutor(std::make_unique<Task>(*this))
|
|
, connection_establisher(std::move(pool_), timeouts_, settings_, log_, table_to_check_)
|
|
{
|
|
epoll.add(timeout_descriptor.getDescriptor());
|
|
}
|
|
|
|
void ConnectionEstablisherAsync::Task::run(AsyncCallback async_callback, SuspendCallback)
|
|
{
|
|
connection_establisher_async.reset();
|
|
connection_establisher_async.connection_establisher.setAsyncCallback(async_callback);
|
|
connection_establisher_async.connection_establisher.run(connection_establisher_async.result, connection_establisher_async.fail_message);
|
|
connection_establisher_async.is_finished = true;
|
|
}
|
|
|
|
void ConnectionEstablisherAsync::processAsyncEvent(int fd, Poco::Timespan socket_timeout, AsyncEventTimeoutType type, const std::string & description, uint32_t events)
|
|
{
|
|
socket_fd = fd;
|
|
socket_description = description;
|
|
epoll.add(fd, events);
|
|
timeout_descriptor.setRelative(socket_timeout);
|
|
timeout = socket_timeout;
|
|
timeout_type = type;
|
|
}
|
|
|
|
void ConnectionEstablisherAsync::clearAsyncEvent()
|
|
{
|
|
timeout_descriptor.reset();
|
|
epoll.remove(socket_fd);
|
|
}
|
|
|
|
bool ConnectionEstablisherAsync::checkBeforeTaskResume()
|
|
{
|
|
/// If we just restarted the task, no need to check timeout.
|
|
if (restarted)
|
|
{
|
|
restarted = false;
|
|
return true;
|
|
}
|
|
|
|
return checkTimeout();
|
|
}
|
|
|
|
void ConnectionEstablisherAsync::cancelAfter()
|
|
{
|
|
if (!is_finished)
|
|
reset();
|
|
}
|
|
|
|
bool ConnectionEstablisherAsync::checkTimeout()
|
|
{
|
|
bool is_socket_ready = false;
|
|
bool is_timeout_alarmed = false;
|
|
|
|
epoll_event events[2];
|
|
events[0].data.fd = events[1].data.fd = -1;
|
|
size_t ready_count = epoll.getManyReady(2, events, 0);
|
|
for (size_t i = 0; i != ready_count; ++i)
|
|
{
|
|
if (events[i].data.fd == socket_fd)
|
|
is_socket_ready = true;
|
|
if (events[i].data.fd == timeout_descriptor.getDescriptor())
|
|
is_timeout_alarmed = true;
|
|
}
|
|
|
|
if (is_timeout_alarmed && !is_socket_ready && !haveMoreAddressesToConnect())
|
|
{
|
|
/// In not async case timeout exception would be thrown and caught in ConnectionEstablisher::run,
|
|
/// but in async case we process timeout outside and cannot throw exception. So, we just save fail message.
|
|
fail_message = getSocketTimeoutExceededMessageByTimeoutType(timeout_type, timeout, socket_description);
|
|
|
|
epoll.remove(socket_fd);
|
|
/// Restart task, so the connection process will start from the beginning in the next resume().
|
|
restart();
|
|
/// The result should be Null in case of timeout.
|
|
resetResult();
|
|
restarted = true;
|
|
/// Mark that current connection process is finished.
|
|
is_finished = true;
|
|
return false;
|
|
}
|
|
|
|
return true;
|
|
}
|
|
|
|
void ConnectionEstablisherAsync::afterTaskResume()
|
|
{
|
|
if (is_finished)
|
|
{
|
|
restart();
|
|
restarted = true;
|
|
}
|
|
}
|
|
|
|
void ConnectionEstablisherAsync::reset()
|
|
{
|
|
resetResult();
|
|
fail_message.clear();
|
|
socket_fd = -1;
|
|
is_finished = false;
|
|
}
|
|
|
|
void ConnectionEstablisherAsync::resetResult()
|
|
{
|
|
if (!result.entry.isNull())
|
|
{
|
|
result.entry->disconnect();
|
|
result.reset();
|
|
}
|
|
}
|
|
|
|
bool ConnectionEstablisherAsync::haveMoreAddressesToConnect()
|
|
{
|
|
return !result.entry.isNull() && result.entry->haveMoreAddressesToConnect();
|
|
}
|
|
|
|
#endif
|
|
|
|
}
|