ClickHouse/programs/server/Server.cpp

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

2326 lines
97 KiB
C++
Raw Normal View History

#include "Server.h"
#include <memory>
#include <sys/resource.h>
#include <sys/stat.h>
#include <sys/types.h>
#include <pwd.h>
#include <unistd.h>
2017-08-09 14:33:07 +00:00
#include <Poco/Net/HTTPServer.h>
#include <Poco/Net/NetException.h>
2019-02-02 14:05:27 +00:00
#include <Poco/Util/HelpFormatter.h>
#include <Poco/Environment.h>
2022-04-27 15:05:45 +00:00
#include <Common/scope_guard_safe.h>
#include <Common/logger_useful.h>
2021-10-02 07:13:14 +00:00
#include <base/phdr_cache.h>
2022-04-27 15:05:45 +00:00
#include <Common/ErrorHandlers.h>
2021-10-02 07:13:14 +00:00
#include <base/getMemoryAmount.h>
#include <base/getAvailableMemoryAmount.h>
2021-10-02 07:13:14 +00:00
#include <base/errnoToString.h>
#include <base/coverage.h>
#include <base/getFQDNOrHostName.h>
#include <base/safeExit.h>
2022-01-18 12:21:59 +00:00
#include <Common/MemoryTracker.h>
2017-08-09 14:33:07 +00:00
#include <Common/ClickHouseRevision.h>
2018-04-19 13:56:14 +00:00
#include <Common/DNSResolver.h>
2017-08-09 14:33:07 +00:00
#include <Common/CurrentMetrics.h>
#include <Common/ConcurrencyControl.h>
#include <Common/Macros.h>
#include <Common/ShellCommand.h>
2017-08-09 14:33:07 +00:00
#include <Common/ZooKeeper/ZooKeeper.h>
#include <Common/ZooKeeper/ZooKeeperNodeCache.h>
#include <Common/getMultipleKeysFromConfig.h>
#include <Common/getNumberOfPhysicalCPUCores.h>
2019-08-11 20:37:53 +00:00
#include <Common/getExecutablePath.h>
#include <Common/ProfileEvents.h>
#include <Common/ThreadProfileEvents.h>
#include <Common/ThreadStatus.h>
2020-09-17 12:53:52 +00:00
#include <Common/getMappedArea.h>
#include <Common/remapExecutable.h>
#include <Common/TLDListsHolder.h>
#include <Common/Config/AbstractConfigurationComparison.h>
2023-05-19 14:23:56 +00:00
#include <Common/assertProcessUserMatchesDataOwner.h>
#include <Common/makeSocketAddress.h>
#include <Server/waitServersToFinish.h>
2021-08-18 12:15:31 +00:00
#include <Core/ServerUUID.h>
#include <IO/ReadHelpers.h>
2022-03-29 16:20:20 +00:00
#include <IO/ReadBufferFromFile.h>
#include <IO/SharedThreadPools.h>
#include <IO/UseSSL.h>
#include <Interpreters/ServerAsynchronousMetrics.h>
2017-08-09 14:33:07 +00:00
#include <Interpreters/DDLWorker.h>
#include <Interpreters/DNSCacheUpdater.h>
#include <Interpreters/DatabaseCatalog.h>
#include <Interpreters/ExternalDictionariesLoader.h>
#include <Interpreters/ProcessList.h>
#include <Interpreters/loadMetadata.h>
#include <Interpreters/JIT/CompiledExpressionCache.h>
#include <Access/AccessControl.h>
#include <Storages/StorageReplicatedMergeTree.h>
#include <Storages/System/attachSystemTables.h>
#include <Storages/System/attachInformationSchemaTables.h>
2021-12-23 03:50:26 +00:00
#include <Storages/Cache/ExternalDataSourceCache.h>
2022-01-04 07:06:19 +00:00
#include <Storages/Cache/registerRemoteFileMetadatas.h>
#include <Common/NamedCollections/NamedCollectionUtils.h>
2017-08-09 14:33:07 +00:00
#include <AggregateFunctions/registerAggregateFunctions.h>
#include <Functions/UserDefined/IUserDefinedSQLObjectsLoader.h>
2017-08-09 14:33:07 +00:00
#include <Functions/registerFunctions.h>
#include <TableFunctions/registerTableFunctions.h>
2020-10-29 03:39:43 +00:00
#include <Formats/registerFormats.h>
#include <Storages/registerStorages.h>
#include <Dictionaries/registerDictionaries.h>
2019-11-27 09:39:44 +00:00
#include <Disks/registerDisks.h>
2022-09-27 13:26:41 +00:00
#include <IO/Resource/registerSchedulerNodes.h>
#include <IO/Resource/registerResourceManagers.h>
#include <Common/Config/ConfigReloader.h>
#include <Server/HTTPHandlerFactory.h>
2016-01-17 13:34:36 +00:00
#include "MetricsTransmitter.h"
2018-06-05 20:09:51 +00:00
#include <Common/StatusFile.h>
#include <Server/TCPHandlerFactory.h>
#include <Server/TCPServer.h>
#include <Common/SensitiveDataMasker.h>
2020-03-01 21:58:50 +00:00
#include <Common/ThreadFuzzer.h>
#include <Common/getHashOfLoadedBinary.h>
#include <Common/filesystemHelpers.h>
2022-06-01 20:07:26 +00:00
#include <Compression/CompressionCodecEncrypted.h>
2022-09-11 21:51:25 +00:00
#include <Server/HTTP/HTTPServerConnectionFactory.h>
#include <Server/MySQLHandlerFactory.h>
2020-05-30 20:02:11 +00:00
#include <Server/PostgreSQLHandlerFactory.h>
2022-09-11 19:00:40 +00:00
#include <Server/ProxyV1HandlerFactory.h>
#include <Server/TLSHandlerFactory.h>
2021-07-06 23:15:30 +00:00
#include <Server/CertificateReloader.h>
2020-10-08 00:23:10 +00:00
#include <Server/ProtocolServerAdapter.h>
#include <Server/HTTP/HTTPServer.h>
#include <Interpreters/AsynchronousInsertQueue.h>
2023-02-18 04:46:09 +00:00
#include <Core/ServerSettings.h>
#include <filesystem>
#include <unordered_set>
2019-07-30 14:04:18 +00:00
#include "config.h"
#include "config_version.h"
2019-09-15 10:35:12 +00:00
#if defined(OS_LINUX)
# include <cstdlib>
# include <sys/un.h>
# include <sys/mman.h>
# include <sys/ptrace.h>
# include <Common/hasLinuxCapability.h>
#endif
#if USE_SSL
# include <Poco/Net/SecureServerSocket.h>
#endif
2020-10-11 02:19:01 +00:00
#if USE_GRPC
# include <Server/GRPCServer.h>
#endif
2021-02-01 13:18:17 +00:00
#if USE_NURAFT
2021-10-27 12:26:42 +00:00
# include <Coordination/FourLetterCommand.h>
# include <Server/KeeperTCPHandlerFactory.h>
2021-02-01 13:18:17 +00:00
#endif
2020-10-11 02:19:01 +00:00
#if USE_JEMALLOC
# include <jemalloc/jemalloc.h>
#endif
2023-01-30 19:00:48 +00:00
#if USE_AZURE_BLOB_STORAGE
# include <azure/storage/common/internal/xml_wrapper.hpp>
#endif
namespace CurrentMetrics
{
extern const Metric Revision;
extern const Metric VersionInteger;
2020-04-19 21:43:06 +00:00
extern const Metric MemoryTracking;
extern const Metric MergesMutationsMemoryTracking;
2021-02-15 10:26:34 +00:00
extern const Metric MaxDDLEntryID;
2021-08-11 03:40:06 +00:00
extern const Metric MaxPushedDDLEntryID;
}
namespace ProfileEvents
{
extern const Event MainConfigLoads;
extern const Event ServerStartupMilliseconds;
}
2021-05-16 22:06:09 +00:00
namespace fs = std::filesystem;
#if USE_JEMALLOC
static bool jemallocOptionEnabled(const char *name)
{
bool value;
size_t size = sizeof(value);
if (mallctl(name, reinterpret_cast<void *>(&value), &size, /* newp= */ nullptr, /* newlen= */ 0))
throw Poco::SystemException("mallctl() failed");
return value;
}
#else
static bool jemallocOptionEnabled(const char *) { return 0; }
#endif
2020-09-17 12:39:37 +00:00
int mainEntryClickHouseServer(int argc, char ** argv)
{
DB::Server app;
if (jemallocOptionEnabled("opt.background_thread"))
{
LOG_ERROR(&app.logger(),
"jemalloc.background_thread was requested, "
"however ClickHouse uses percpu_arena and background_thread most likely will not give any benefits, "
"and also background_thread is not compatible with ClickHouse watchdog "
"(that can be disabled with CLICKHOUSE_WATCHDOG_ENABLE=0)");
}
/// Do not fork separate process from watchdog if we attached to terminal.
/// Otherwise it breaks gdb usage.
/// Can be overridden by environment variable (cannot use server config at this moment).
if (argc > 0)
{
const char * env_watchdog = getenv("CLICKHOUSE_WATCHDOG_ENABLE"); // NOLINT(concurrency-mt-unsafe)
if (env_watchdog)
{
if (0 == strcmp(env_watchdog, "1"))
app.shouldSetupWatchdog(argv[0]);
/// Other values disable watchdog explicitly.
}
else if (!isatty(STDIN_FILENO) && !isatty(STDOUT_FILENO) && !isatty(STDERR_FILENO))
app.shouldSetupWatchdog(argv[0]);
}
2020-09-17 12:39:37 +00:00
try
{
return app.run(argc, argv);
}
catch (...)
{
std::cerr << DB::getCurrentExceptionMessage(true) << "\n";
auto code = DB::getCurrentExceptionCode();
return code ? code : 1;
}
}
2012-03-09 03:06:09 +00:00
namespace DB
{
namespace ErrorCodes
{
extern const int NO_ELEMENTS_IN_CONFIG;
extern const int SUPPORT_IS_DISABLED;
extern const int ARGUMENT_OUT_OF_BOUND;
extern const int EXCESSIVE_ELEMENT_IN_CONFIG;
extern const int INVALID_CONFIG_PARAMETER;
extern const int NETWORK_ERROR;
extern const int CORRUPTED_DATA;
}
static std::string getCanonicalPath(std::string && path)
2012-03-09 03:06:09 +00:00
{
Poco::trimInPlace(path);
if (path.empty())
throw Exception(ErrorCodes::INVALID_CONFIG_PARAMETER, "path configuration parameter is empty");
if (path.back() != '/')
path += '/';
return std::move(path);
}
Poco::Net::SocketAddress Server::socketBindListen(
const Poco::Util::AbstractConfiguration & config,
Poco::Net::ServerSocket & socket,
const std::string & host,
UInt16 port,
[[maybe_unused]] bool secure) const
2020-11-11 13:07:06 +00:00
{
auto address = makeSocketAddress(host, port, &logger());
socket.bind(address, /* reuseAddress = */ true, /* reusePort = */ config.getBool("listen_reuse_port", false));
/// If caller requests any available port from the OS, discover it after binding.
if (port == 0)
{
address = socket.address();
LOG_DEBUG(&logger(), "Requested any available port (port == 0), actual port is {:d}", address.port());
}
socket.listen(/* backlog = */ config.getUInt("listen_backlog", 4096));
2020-11-11 13:07:06 +00:00
return address;
}
2022-04-05 15:04:53 +00:00
Strings getListenHosts(const Poco::Util::AbstractConfiguration & config)
{
auto listen_hosts = DB::getMultipleValuesFromConfig(config, "", "listen_host");
if (listen_hosts.empty())
{
listen_hosts.emplace_back("::1");
listen_hosts.emplace_back("127.0.0.1");
}
return listen_hosts;
}
2022-04-05 15:04:53 +00:00
Strings getInterserverListenHosts(const Poco::Util::AbstractConfiguration & config)
{
auto interserver_listen_hosts = DB::getMultipleValuesFromConfig(config, "", "interserver_listen_host");
2022-04-05 15:04:53 +00:00
if (!interserver_listen_hosts.empty())
return interserver_listen_hosts;
2022-04-05 15:04:53 +00:00
/// Use more general restriction in case of emptiness
return getListenHosts(config);
}
bool getListenTry(const Poco::Util::AbstractConfiguration & config)
{
bool listen_try = config.getBool("listen_try", false);
if (!listen_try)
{
Poco::Util::AbstractConfiguration::Keys protocols;
config.keys("protocols", protocols);
listen_try =
DB::getMultipleValuesFromConfig(config, "", "listen_host").empty() &&
std::none_of(protocols.begin(), protocols.end(), [&](const auto & protocol)
{
return config.has("protocols." + protocol + ".host") && config.has("protocols." + protocol + ".port");
});
}
return listen_try;
}
void Server::createServer(
Poco::Util::AbstractConfiguration & config,
const std::string & listen_host,
const char * port_name,
bool listen_try,
bool start_server,
std::vector<ProtocolServerAdapter> & servers,
CreateServerFunc && func) const
2020-11-11 13:07:06 +00:00
{
/// For testing purposes, user may omit tcp_port or http_port or https_port in configuration file.
if (config.getString(port_name, "").empty())
2020-11-11 13:07:06 +00:00
return;
/// If we already have an active server for this listen_host/port_name, don't create it again
for (const auto & server : servers)
Fix server reload on port change (do not wait for current connections) If you will execute 'SYSTEM RELOAD CONFIG' via, i.e., TCP protocol, then reload on port change will endlessly wait for connection from which this query had been issued, and you will see the following message in the logs: 2022.04.28 03:34:57.552513 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Debug> executeQuery: (from 127.0.0.1:11774) system reload config ... 2022.04.28 03:34:57.710640 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Information> Application: Stopped listening for http://127.0.0.1:18123 2022.04.28 03:34:57.798774 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Information> Application: Stopped listening for native protocol (tcp): 127.0.0.1:19000 ... 2022.04.28 03:34:57.901375 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Debug> Application: Server finished: http://127.0.0.1:18123 2022.04.28 03:34:57.901455 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 2022.04.28 03:34:58.001717 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 2022.04.28 03:34:58.101881 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 ... 2022.04.28 03:35:01.707951 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 But waiting for the current connection will never ends. So instead of waiting directly from the query context (SYSTEM RELOAD CONFIG) do this in background (actually not even in background, but check on server reload and on exit). v0: just don't wait for the servers v2: fix use-after-free by removing dependency from server in handlers v3: wait servers in background to avoid use-after-free of the context Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
2022-04-27 10:24:49 +00:00
{
if (!server.isStopping() && server.getListenHost() == listen_host && server.getPortName() == port_name)
return;
Fix server reload on port change (do not wait for current connections) If you will execute 'SYSTEM RELOAD CONFIG' via, i.e., TCP protocol, then reload on port change will endlessly wait for connection from which this query had been issued, and you will see the following message in the logs: 2022.04.28 03:34:57.552513 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Debug> executeQuery: (from 127.0.0.1:11774) system reload config ... 2022.04.28 03:34:57.710640 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Information> Application: Stopped listening for http://127.0.0.1:18123 2022.04.28 03:34:57.798774 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Information> Application: Stopped listening for native protocol (tcp): 127.0.0.1:19000 ... 2022.04.28 03:34:57.901375 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Debug> Application: Server finished: http://127.0.0.1:18123 2022.04.28 03:34:57.901455 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 2022.04.28 03:34:58.001717 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 2022.04.28 03:34:58.101881 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 ... 2022.04.28 03:35:01.707951 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 But waiting for the current connection will never ends. So instead of waiting directly from the query context (SYSTEM RELOAD CONFIG) do this in background (actually not even in background, but check on server reload and on exit). v0: just don't wait for the servers v2: fix use-after-free by removing dependency from server in handlers v3: wait servers in background to avoid use-after-free of the context Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
2022-04-27 10:24:49 +00:00
}
auto port = config.getInt(port_name);
2020-11-11 13:07:06 +00:00
try
{
servers.push_back(func(port));
if (start_server)
{
servers.back().start();
LOG_INFO(&logger(), "Listening for {}", servers.back().getDescription());
}
2021-08-20 04:11:47 +00:00
global_context->registerServerPort(port_name, port);
2020-11-11 13:07:06 +00:00
}
catch (const Poco::Exception &)
{
std::string message = "Listen [" + listen_host + "]:" + std::to_string(port) + " failed: " + getCurrentExceptionMessage(false);
if (listen_try)
{
LOG_WARNING(&logger(), "{}. If it is an IPv6 or IPv4 address and your host has disabled IPv6 or IPv4, then consider to "
"specify not disabled IPv4 or IPv6 address to listen in <listen_host> element of configuration "
"file. Example for disabled IPv6: <listen_host>0.0.0.0</listen_host> ."
" Example for disabled IPv4: <listen_host>::</listen_host>",
message);
}
else
{
2023-01-23 13:16:14 +00:00
throw Exception::createDeprecated(message, ErrorCodes::NETWORK_ERROR);
2020-11-11 13:07:06 +00:00
}
}
}
2023-01-01 22:22:26 +00:00
2023-01-02 02:27:21 +00:00
#if defined(OS_LINUX)
2023-01-01 22:22:26 +00:00
namespace
{
void setOOMScore(int value, Poco::Logger * log)
{
try
{
std::string value_string = std::to_string(value);
DB::WriteBufferFromFile buf("/proc/self/oom_score_adj");
buf.write(value_string.c_str(), value_string.size());
buf.next();
buf.close();
}
catch (const Poco::Exception & e)
{
LOG_WARNING(log, "Failed to adjust OOM score: '{}'.", e.displayText());
return;
}
LOG_INFO(log, "Set OOM score adjustment to {}", value);
}
}
2023-01-02 02:27:21 +00:00
#endif
2023-01-01 22:22:26 +00:00
void Server::uninitialize()
{
logger().information("shutting down");
BaseDaemon::uninitialize();
}
2019-02-02 13:17:55 +00:00
int Server::run()
{
if (config().hasOption("help"))
{
Poco::Util::HelpFormatter help_formatter(Server::options());
2020-11-10 18:22:26 +00:00
auto header_str = fmt::format("{} [OPTION] [-- [ARG]...]\n"
"positional arguments can be used to rewrite config.xml properties, for example, --http_port=8010",
commandName());
2020-11-10 18:22:26 +00:00
help_formatter.setHeader(header_str);
help_formatter.format(std::cout);
2019-02-02 13:17:55 +00:00
return 0;
}
if (config().hasOption("version"))
{
std::cout << DBMS_NAME << " server version " << VERSION_STRING << VERSION_OFFICIAL << "." << std::endl;
return 0;
}
2020-03-18 02:02:24 +00:00
return Application::run(); // NOLINT
2019-02-02 13:17:55 +00:00
}
void Server::initialize(Poco::Util::Application & self)
{
BaseDaemon::initialize(self);
logger().information("starting up");
2021-06-22 23:02:57 +00:00
LOG_INFO(&logger(), "OS name: {}, version: {}, architecture: {}",
Poco::Environment::osName(),
Poco::Environment::osVersion(),
Poco::Environment::osArchitecture());
}
std::string Server::getDefaultCorePath() const
{
return getCanonicalPath(config().getString("path", DBMS_DEFAULT_PATH)) + "cores";
}
2019-08-03 11:02:40 +00:00
void Server::defineOptions(Poco::Util::OptionSet & options)
2019-02-02 13:17:55 +00:00
{
2019-08-03 11:02:40 +00:00
options.addOption(
2019-02-02 13:17:55 +00:00
Poco::Util::Option("help", "h", "show help and exit")
.required(false)
.repeatable(false)
.binding("help"));
2019-08-03 11:02:40 +00:00
options.addOption(
2019-02-04 12:49:54 +00:00
Poco::Util::Option("version", "V", "show version and exit")
.required(false)
.repeatable(false)
.binding("version"));
2019-08-03 11:02:40 +00:00
BaseDaemon::defineOptions(options);
2019-02-02 13:17:55 +00:00
}
2020-06-04 19:30:30 +00:00
void checkForUsersNotInMainConfig(
const Poco::Util::AbstractConfiguration & config,
const std::string & config_path,
const std::string & users_config_path,
Poco::Logger * log)
{
if (config.getBool("skip_check_for_incorrect_settings", false))
return;
if (config.has("users") || config.has("profiles") || config.has("quotas"))
{
/// We cannot throw exception here, because we have support for obsolete 'conf.d' directory
/// (that does not correspond to config.d or users.d) but substitute configuration to both of them.
LOG_ERROR(log, "The <users>, <profiles> and <quotas> elements should be located in users config file: {} not in main config {}."
" Also note that you should place configuration changes to the appropriate *.d directory like 'users.d'.",
users_config_path, config_path);
}
}
2022-03-29 12:20:46 +00:00
/// Unused in other builds
#if defined(OS_LINUX)
static String readString(const String & path)
{
ReadBufferFromFile in(path);
String contents;
readStringUntilEOF(contents, in);
return contents;
}
static int readNumber(const String & path)
{
ReadBufferFromFile in(path);
int result;
readText(result, in);
return result;
}
2022-03-29 12:20:46 +00:00
#endif
static void sanityChecks(Server & server)
{
std::string data_path = getCanonicalPath(server.config().getString("path", DBMS_DEFAULT_PATH));
std::string logs_path = server.config().getString("logger.log", "");
if (server.logger().is(Poco::Message::PRIO_TEST))
server.context()->addWarningMessage("Server logging level is set to 'test' and performance is degraded. This cannot be used in production.");
#if defined(OS_LINUX)
try
{
2022-05-31 11:50:09 +00:00
const char * filename = "/sys/devices/system/clocksource/clocksource0/current_clocksource";
2022-09-12 12:56:01 +00:00
String clocksource = readString(filename);
if (clocksource.find("tsc") == std::string::npos && clocksource.find("kvm-clock") == std::string::npos)
2022-09-13 10:44:42 +00:00
server.context()->addWarningMessage("Linux is not using a fast clock source. Performance can be degraded. Check " + String(filename));
}
catch (...)
{
}
try
{
2022-05-31 11:50:09 +00:00
const char * filename = "/proc/sys/vm/overcommit_memory";
if (readNumber(filename) == 2)
server.context()->addWarningMessage("Linux memory overcommit is disabled. Check " + String(filename));
}
catch (...)
{
}
try
{
2022-05-31 11:50:09 +00:00
const char * filename = "/sys/kernel/mm/transparent_hugepage/enabled";
if (readString(filename).find("[always]") != std::string::npos)
server.context()->addWarningMessage("Linux transparent hugepages are set to \"always\". Check " + String(filename));
}
catch (...)
{
}
try
{
2022-05-31 11:50:09 +00:00
const char * filename = "/proc/sys/kernel/pid_max";
if (readNumber(filename) < 30000)
server.context()->addWarningMessage("Linux max PID is too low. Check " + String(filename));
}
catch (...)
{
}
try
{
2022-05-31 11:50:09 +00:00
const char * filename = "/proc/sys/kernel/threads-max";
if (readNumber(filename) < 30000)
server.context()->addWarningMessage("Linux threads max count is too low. Check " + String(filename));
}
catch (...)
{
}
2022-02-15 10:39:45 +00:00
std::string dev_id = getBlockDeviceId(data_path);
if (getBlockDeviceType(dev_id) == BlockDeviceType::ROT && getBlockDeviceReadAheadBytes(dev_id) == 0)
2022-05-31 07:26:26 +00:00
server.context()->addWarningMessage("Rotational disk with disabled readahead is in use. Performance can be degraded. Used for data: " + String(data_path));
#endif
2022-03-30 22:22:36 +00:00
try
{
if (getAvailableMemoryAmount() < (2l << 30))
server.context()->addWarningMessage("Available memory at server startup is too low (2GiB).");
2022-08-13 23:49:00 +00:00
}
catch (...)
{
}
2022-08-13 23:49:00 +00:00
try
{
2022-03-30 22:22:36 +00:00
if (!enoughSpaceInDirectory(data_path, 1ull << 30))
server.context()->addWarningMessage("Available disk space for data at server startup is too low (1GiB): " + String(data_path));
2022-08-13 23:49:00 +00:00
}
catch (...)
{
}
2022-08-13 23:49:00 +00:00
try
{
2022-03-30 22:22:36 +00:00
if (!logs_path.empty())
{
2022-04-21 05:45:08 +00:00
auto logs_parent = fs::path(logs_path).parent_path();
if (!enoughSpaceInDirectory(logs_parent, 1ull << 30))
server.context()->addWarningMessage("Available disk space for logs at server startup is too low (1GiB): " + String(logs_parent));
2022-03-30 22:22:36 +00:00
}
}
catch (...)
2022-02-15 10:39:45 +00:00
{
}
2022-08-13 23:49:00 +00:00
if (server.context()->getMergeTreeSettings().allow_remote_fs_zero_copy_replication)
{
server.context()->addWarningMessage("The setting 'allow_remote_fs_zero_copy_replication' is enabled for MergeTree tables."
" But the feature of 'zero-copy replication' is under development and is not ready for production."
" The usage of this feature can lead to data corruption and loss. The setting should be disabled in production.");
}
}
2021-04-25 03:06:38 +00:00
2017-12-02 02:47:12 +00:00
int Server::main(const std::vector<std::string> & /*args*/)
try
{
Stopwatch startup_watch;
2020-05-30 21:57:37 +00:00
Poco::Logger * log = &logger();
2020-08-16 11:52:55 +00:00
UseSSL use_ssl;
MainThreadStatus::getInstance();
2023-02-18 04:46:09 +00:00
ServerSettings server_settings;
server_settings.loadSettingsFromConfig(config());
StackTrace::setShowAddresses(server_settings.show_addresses_in_stack_traces);
2022-08-20 15:09:20 +00:00
#if USE_HDFS
/// This will point libhdfs3 to the right location for its config.
/// Note: this has to be done once at server initialization, because 'setenv' is not thread-safe.
String libhdfs3_conf = config().getString("hdfs.libhdfs3_conf", "");
if (!libhdfs3_conf.empty())
{
if (std::filesystem::path{libhdfs3_conf}.is_relative() && !std::filesystem::exists(libhdfs3_conf))
{
const String config_path = config().getString("config-file", "config.xml");
const auto config_dir = std::filesystem::path{config_path}.remove_filename();
if (std::filesystem::exists(config_dir / libhdfs3_conf))
libhdfs3_conf = std::filesystem::absolute(config_dir / libhdfs3_conf);
}
setenv("LIBHDFS3_CONF", libhdfs3_conf.c_str(), true /* overwrite */); // NOLINT
}
#endif
2022-12-06 22:18:35 +00:00
#if USE_OPENSSL_INTREE
/// When building openssl into clickhouse, clickhouse owns the configuration
2022-12-06 22:44:10 +00:00
/// Therefore, the clickhouse openssl configuration should be kept separate from
/// the OS. Default to the one in the standard config directory, unless overridden
2022-12-06 22:18:35 +00:00
/// by a key in the config.
if (config().has("opensslconf"))
{
std::string opensslconf_path = config().getString("opensslconf");
setenv("OPENSSL_CONF", opensslconf_path.c_str(), true);
}
else
{
const String config_path = config().getString("config-file", "config.xml");
2022-12-16 20:31:17 +00:00
const auto config_dir = std::filesystem::path{config_path}.replace_filename("openssl.conf");
2023-02-23 14:42:06 +00:00
setenv("OPENSSL_CONF", config_dir.c_str(), true);
2022-12-06 22:18:35 +00:00
}
#endif
2017-04-21 17:47:27 +00:00
registerFunctions();
registerAggregateFunctions();
registerTableFunctions();
registerStorages();
registerDictionaries();
registerDisks(/* global_skip_access_check= */ false);
2020-10-29 03:39:43 +00:00
registerFormats();
2022-01-04 07:06:19 +00:00
registerRemoteFileMetadatas();
2022-09-27 13:26:41 +00:00
registerSchedulerNodes();
registerResourceManagers();
2017-04-21 17:47:27 +00:00
2020-09-17 12:15:05 +00:00
CurrentMetrics::set(CurrentMetrics::Revision, ClickHouseRevision::getVersionRevision());
CurrentMetrics::set(CurrentMetrics::VersionInteger, ClickHouseRevision::getVersionInteger());
/** Context contains all that query execution is dependent:
2020-07-08 10:40:02 +00:00
* settings, available functions, data types, aggregate functions, databases, ...
*/
2020-04-16 14:37:38 +00:00
auto shared_context = Context::createShared();
global_context = Context::createGlobal(shared_context.get());
2019-07-08 02:14:32 +00:00
global_context->makeGlobalContext();
global_context->setApplicationType(Context::ApplicationType::SERVER);
2021-07-09 14:40:32 +00:00
#if !defined(NDEBUG) || !defined(__OPTIMIZE__)
global_context->addWarningMessage("Server was built in debug mode. It will work slowly.");
#endif
if (ThreadFuzzer::instance().isEffective())
global_context->addWarningMessage("ThreadFuzzer is enabled. Application will run slowly and unstable.");
2021-07-12 10:57:39 +00:00
#if defined(SANITIZER)
global_context->addWarningMessage("Server was built with sanitizer. It will work slowly.");
#endif
const auto memory_amount = getMemoryAmount();
LOG_INFO(log, "Available RAM: {}; physical cores: {}; logical cores: {}.",
formatReadableSizeWithBinarySuffix(memory_amount),
getNumberOfPhysicalCPUCores(), // on ARM processors it can show only enabled at current moment cores
std::thread::hardware_concurrency());
sanityChecks(*this);
2021-07-12 10:57:39 +00:00
// Initialize global thread pool. Do it before we fetch configs from zookeeper
// nodes (`from_zk`), because ZooKeeper interface uses the pool. We will
// ignore `max_thread_pool_size` in configs we fetch from ZK, but oh well.
GlobalThreadPool::initialize(
2023-02-18 04:46:09 +00:00
server_settings.max_thread_pool_size,
server_settings.max_thread_pool_free_size,
server_settings.thread_pool_queue_size);
2023-01-30 19:00:48 +00:00
#if USE_AZURE_BLOB_STORAGE
/// It makes sense to deinitialize libxml after joining of all threads
/// in global pool because libxml uses thread-local memory allocations via
/// 'pthread_key_create' and 'pthread_setspecific' which should be deallocated
/// at 'pthread_exit'. Deinitialization of libxml leads to call of 'pthread_key_delete'
/// and if it is done before joining of threads, allocated memory will not be freed
/// and there may be memory leaks in threads that used libxml.
GlobalThreadPool::instance().addOnDestroyCallback([]
{
Azure::Storage::_internal::XmlGlobalDeinitialize();
});
#endif
getIOThreadPool().initialize(
2023-02-18 04:46:09 +00:00
server_settings.max_io_thread_pool_size,
server_settings.max_io_thread_pool_free_size,
server_settings.io_thread_pool_queue_size);
2021-11-29 08:22:43 +00:00
getBackupsIOThreadPool().initialize(
server_settings.max_backups_io_thread_pool_size,
server_settings.max_backups_io_thread_pool_free_size,
server_settings.backups_io_thread_pool_queue_size);
getActivePartsLoadingThreadPool().initialize(
server_settings.max_active_parts_loading_thread_pool_size,
0, // We don't need any threads once all the parts will be loaded
server_settings.max_active_parts_loading_thread_pool_size);
getOutdatedPartsLoadingThreadPool().initialize(
server_settings.max_outdated_parts_loading_thread_pool_size,
0, // We don't need any threads once all the parts will be loaded
server_settings.max_outdated_parts_loading_thread_pool_size);
/// It could grow if we need to synchronously wait until all the data parts will be loaded.
getOutdatedPartsLoadingThreadPool().setMaxTurboThreads(
server_settings.max_active_parts_loading_thread_pool_size
);
getPartsCleaningThreadPool().initialize(
server_settings.max_parts_cleaning_thread_pool_size,
0, // We don't need any threads one all the parts will be deleted
server_settings.max_parts_cleaning_thread_pool_size);
2021-11-29 08:22:43 +00:00
/// Initialize global local cache for remote filesystem.
if (config().has("local_cache_for_remote_fs"))
{
bool enable = config().getBool("local_cache_for_remote_fs.enable", false);
if (enable)
{
String root_dir = config().getString("local_cache_for_remote_fs.root_dir");
UInt64 limit_size = config().getUInt64("local_cache_for_remote_fs.limit_size");
UInt64 bytes_read_before_flush
= config().getUInt64("local_cache_for_remote_fs.bytes_read_before_flush", DBMS_DEFAULT_BUFFER_SIZE);
2021-12-23 03:50:26 +00:00
ExternalDataSourceCache::instance().initOnce(global_context, root_dir, limit_size, bytes_read_before_flush);
2021-11-29 08:22:43 +00:00
}
}
2023-02-18 04:46:09 +00:00
Poco::ThreadPool server_pool(3, server_settings.max_connections);
std::mutex servers_lock;
std::vector<ProtocolServerAdapter> servers;
std::vector<ProtocolServerAdapter> servers_to_start_before_tables;
/// This object will periodically calculate some metrics.
ServerAsynchronousMetrics async_metrics(
global_context,
2023-02-18 04:46:09 +00:00
server_settings.asynchronous_metrics_update_period_s,
server_settings.asynchronous_heavy_metrics_update_period_s,
[&]() -> std::vector<ProtocolServerMetrics>
{
std::vector<ProtocolServerMetrics> metrics;
2021-12-27 16:54:14 +00:00
metrics.reserve(servers_to_start_before_tables.size());
for (const auto & server : servers_to_start_before_tables)
metrics.emplace_back(ProtocolServerMetrics{server.getPortName(), server.currentThreads()});
2021-12-27 16:54:14 +00:00
std::lock_guard lock(servers_lock);
for (const auto & server : servers)
metrics.emplace_back(ProtocolServerMetrics{server.getPortName(), server.currentThreads()});
return metrics;
}
);
2023-03-23 08:58:56 +00:00
zkutil::validateZooKeeperConfig(config());
bool has_zookeeper = zkutil::hasZooKeeperConfig(config());
zkutil::ZooKeeperNodeCache main_config_zk_node_cache([&] { return global_context->getZooKeeper(); });
zkutil::EventPtr main_config_zk_changed_event = std::make_shared<Poco::Event>();
if (loaded_config.has_zk_includes)
{
auto old_configuration = loaded_config.configuration;
ConfigProcessor config_processor(config_path);
loaded_config = config_processor.loadConfigWithZooKeeperIncludes(
main_config_zk_node_cache, main_config_zk_changed_event, /* fallback_to_preprocessed = */ true);
config_processor.savePreprocessedConfig(loaded_config, config().getString("path", DBMS_DEFAULT_PATH));
config().removeConfiguration(old_configuration.get());
config().add(loaded_config.configuration.duplicate(), PRIO_DEFAULT, false);
}
Settings::checkNoSettingNamesAtTopLevel(config(), config_path);
2020-06-04 19:30:30 +00:00
2023-02-18 04:46:09 +00:00
/// We need to reload server settings because config could be updated via zookeeper.
server_settings.loadSettingsFromConfig(config());
#if defined(OS_LINUX)
2019-08-11 20:37:53 +00:00
std::string executable_path = getExecutablePath();
if (!executable_path.empty())
{
/// Integrity check based on checksum of the executable code.
/// Note: it is not intended to protect from malicious party,
/// because the reference checksum can be easily modified as well.
/// And we don't involve asymmetric encryption with PKI yet.
/// It's only intended to protect from faulty hardware.
/// Note: it is only based on machine code.
/// But there are other sections of the binary (e.g. exception handling tables)
/// that are interpreted (not executed) but can alter the behaviour of the program as well.
2022-06-15 22:07:59 +00:00
/// Please keep the below log messages in-sync with the ones in daemon/BaseDaemon.cpp
if (stored_binary_hash.empty())
{
LOG_WARNING(log, "Integrity check of the executable skipped because the reference checksum could not be read.");
}
else
{
String calculated_binary_hash = getHashOfLoadedBinaryHex();
if (calculated_binary_hash == stored_binary_hash)
{
LOG_INFO(log, "Integrity check of the executable successfully passed (checksum: {})", calculated_binary_hash);
}
else
{
/// If program is run under debugger, ptrace will fail.
if (ptrace(PTRACE_TRACEME, 0, nullptr, nullptr) == -1)
{
/// Program is run under debugger. Modification of it's binary image is ok for breakpoints.
global_context->addWarningMessage(fmt::format(
"Server is run under debugger and its binary image is modified (most likely with breakpoints).",
calculated_binary_hash));
}
else
{
throw Exception(
ErrorCodes::CORRUPTED_DATA,
"Calculated checksum of the executable ({0}) does not correspond"
" to the reference checksum stored in the executable ({1})."
" This may indicate one of the following:"
" - the executable {2} was changed just after startup;"
" - the executable {2} was corrupted on disk due to faulty hardware;"
" - the loaded executable was corrupted in memory due to faulty hardware;"
" - the file {2} was intentionally modified;"
" - a logical error in the code.",
calculated_binary_hash,
stored_binary_hash,
executable_path);
}
}
}
}
else
2019-08-11 20:37:53 +00:00
executable_path = "/usr/bin/clickhouse"; /// It is used for information messages.
2018-11-14 01:20:46 +00:00
/// After full config loaded
{
2020-09-14 18:08:09 +00:00
if (config().getBool("remap_executable", false))
{
LOG_DEBUG(log, "Will remap executable in memory.");
size_t size = remapExecutable();
LOG_DEBUG(log, "The code ({}) in memory has been successfully remapped.", ReadableSize(size));
2020-09-14 18:08:09 +00:00
}
2020-09-14 18:08:09 +00:00
if (config().getBool("mlock_executable", false))
{
if (hasLinuxCapability(CAP_IPC_LOCK))
{
2020-09-25 02:03:58 +00:00
try
{
/// Get the memory area with (current) code segment.
/// It's better to lock only the code segment instead of calling "mlockall",
/// because otherwise debug info will be also locked in memory, and it can be huge.
auto [addr, len] = getMappedArea(reinterpret_cast<void *>(mainEntryClickHouseServer));
LOG_TRACE(log, "Will do mlock to prevent executable memory from being paged out. It may take a few seconds.");
if (0 != mlock(addr, len))
2022-08-20 15:09:20 +00:00
LOG_WARNING(log, "Failed mlock: {}", errnoToString());
2020-09-25 02:03:58 +00:00
else
LOG_TRACE(log, "The memory map of clickhouse executable has been mlock'ed, total {}", ReadableSize(len));
}
catch (...)
{
LOG_WARNING(log, "Cannot mlock: {}", getCurrentExceptionMessage(false));
}
}
else
{
2020-05-23 22:24:01 +00:00
LOG_INFO(log, "It looks like the process has no CAP_IPC_LOCK capability, binary mlock will be disabled."
2020-05-23 22:21:29 +00:00
" It could happen due to incorrect ClickHouse package installation."
" You could resolve the problem manually with 'sudo setcap cap_ipc_lock=+ep {}'."
" Note that it will not work on 'nosuid' mounted filesystems.", executable_path);
2020-09-14 18:08:09 +00:00
}
}
}
2023-01-01 22:22:26 +00:00
int default_oom_score = 0;
#if !defined(NDEBUG)
/// In debug version on Linux, increase oom score so that clickhouse is killed
/// first, instead of some service. Use a carefully chosen random score of 555:
/// the maximum is 1000, and chromium uses 300 for its tab processes. Ignore
/// whatever errors that occur, because it's just a debugging aid and we don't
/// care if it breaks.
default_oom_score = 555;
#endif
int oom_score = config().getInt("oom_score", default_oom_score);
if (oom_score)
setOOMScore(oom_score, log);
2018-11-14 01:20:46 +00:00
#endif
global_context->setRemoteHostFilter(config());
2021-08-16 18:30:53 +00:00
std::string path_str = getCanonicalPath(config().getString("path", DBMS_DEFAULT_PATH));
fs::path path = path_str;
2023-02-18 04:46:09 +00:00
std::string default_database = server_settings.default_database.toString();
2020-08-08 01:52:28 +00:00
/// Check that the process user id matches the owner of the data.
2023-05-19 14:23:56 +00:00
assertProcessUserMatchesDataOwner(path_str, [&](const std::string & message){ global_context->addWarningMessage(message); });
2021-08-16 18:30:53 +00:00
global_context->setPath(path_str);
2021-08-16 18:30:53 +00:00
StatusFile status{path / "status", StatusFile::write_full_info};
2023-04-09 01:04:26 +00:00
ServerUUID::load(path / "uuid", log);
/// Try to increase limit on number of open files.
{
rlimit rlim;
if (getrlimit(RLIMIT_NOFILE, &rlim))
throw Poco::Exception("Cannot getrlimit");
if (rlim.rlim_cur == rlim.rlim_max)
{
2020-05-23 22:24:01 +00:00
LOG_DEBUG(log, "rlimit on number of file descriptors is {}", rlim.rlim_cur);
}
else
{
rlim_t old = rlim.rlim_cur;
rlim.rlim_cur = config().getUInt("max_open_files", static_cast<unsigned>(rlim.rlim_max));
int rc = setrlimit(RLIMIT_NOFILE, &rlim);
if (rc != 0)
2022-08-20 15:09:20 +00:00
LOG_WARNING(log, "Cannot set max number of file descriptors to {}. Try to specify max_open_files according to your system limits. error: {}", rlim.rlim_cur, errnoToString());
else
2020-05-23 22:24:01 +00:00
LOG_DEBUG(log, "Set max number of file descriptors to {} (was {}).", rlim.rlim_cur, old);
}
}
/// Try to increase limit on number of threads.
{
rlimit rlim;
if (getrlimit(RLIMIT_NPROC, &rlim))
throw Poco::Exception("Cannot getrlimit");
if (rlim.rlim_cur == rlim.rlim_max)
{
LOG_DEBUG(log, "rlimit on number of threads is {}", rlim.rlim_cur);
}
else
{
rlim_t old = rlim.rlim_cur;
rlim.rlim_cur = rlim.rlim_max;
int rc = setrlimit(RLIMIT_NPROC, &rlim);
2022-02-15 10:39:45 +00:00
if (rc != 0)
{
2022-08-20 15:09:20 +00:00
LOG_WARNING(log, "Cannot set max number of threads to {}. error: {}", rlim.rlim_cur, errnoToString());
rlim.rlim_cur = old;
}
else
{
LOG_DEBUG(log, "Set max number of threads to {} (was {}).", rlim.rlim_cur, old);
}
}
2022-02-15 10:39:45 +00:00
if (rlim.rlim_cur < 30000)
{
global_context->addWarningMessage("Maximum number of threads is lower than 30000. There could be problems with handling a lot of simultaneous queries.");
}
}
static ServerErrorHandler error_handler;
Poco::ErrorHandler::set(&error_handler);
/// Initialize DateLUT early, to not interfere with running time of first query.
2020-05-23 22:24:01 +00:00
LOG_DEBUG(log, "Initializing DateLUT.");
DateLUT::serverTimezoneInstance();
LOG_TRACE(log, "Initialized DateLUT with time zone '{}'.", DateLUT::serverTimezoneInstance().getTimeZone());
/// Storage with temporary data for processing of heavy queries.
2023-02-18 04:46:09 +00:00
if (!server_settings.tmp_policy.value.empty())
2022-12-06 10:04:15 +00:00
{
2023-02-18 04:46:09 +00:00
global_context->setTemporaryStoragePolicy(server_settings.tmp_policy, server_settings.max_temporary_data_on_disk_size);
2022-12-06 10:04:15 +00:00
}
2023-02-18 04:46:09 +00:00
else if (!server_settings.temporary_data_in_cache.value.empty())
2022-12-06 10:04:15 +00:00
{
2023-02-18 04:46:09 +00:00
global_context->setTemporaryStorageInCache(server_settings.temporary_data_in_cache, server_settings.max_temporary_data_on_disk_size);
2022-12-06 10:04:15 +00:00
}
else
{
2022-10-05 16:35:10 +00:00
std::string temporary_path = config().getString("tmp_path", path / "tmp/");
2023-02-18 04:46:09 +00:00
global_context->setTemporaryStoragePath(temporary_path, server_settings.max_temporary_data_on_disk_size);
}
/** Directory with 'flags': files indicating temporary settings for the server set by system administrator.
* Flags may be cleared automatically after being applied by the server.
* Examples: do repair of local data; clone all replicated tables from replica.
*/
{
2021-08-16 18:30:53 +00:00
auto flags_path = path / "flags/";
2021-05-16 22:06:09 +00:00
fs::create_directories(flags_path);
global_context->setFlagsPath(flags_path);
}
/** Directory with user provided files that are usable by 'file' table function.
*/
{
2021-08-16 18:30:53 +00:00
std::string user_files_path = config().getString("user_files_path", path / "user_files/");
global_context->setUserFilesPath(user_files_path);
2021-05-16 22:06:09 +00:00
fs::create_directories(user_files_path);
}
{
2021-08-16 18:30:53 +00:00
std::string dictionaries_lib_path = config().getString("dictionaries_lib_path", path / "dictionaries_lib/");
global_context->setDictionariesLibPath(dictionaries_lib_path);
2021-05-16 22:06:09 +00:00
fs::create_directories(dictionaries_lib_path);
}
2021-09-01 23:46:23 +00:00
{
std::string user_scripts_path = config().getString("user_scripts_path", path / "user_scripts/");
global_context->setUserScriptsPath(user_scripts_path);
fs::create_directories(user_scripts_path);
}
/// top_level_domains_lists
{
2021-08-16 18:30:53 +00:00
const std::string & top_level_domains_path = config().getString("top_level_domains_path", path / "top_level_domains/");
2021-05-16 22:06:09 +00:00
TLDListsHolder::getInstance().parseConfig(fs::path(top_level_domains_path) / "", config());
}
2020-03-19 21:14:52 +00:00
{
2021-08-16 18:30:53 +00:00
fs::create_directories(path / "data/");
fs::create_directories(path / "metadata/");
2020-08-12 20:40:13 +00:00
2020-03-19 21:14:52 +00:00
/// Directory with metadata of tables, which was marked as dropped by Atomic database
2021-08-16 18:30:53 +00:00
fs::create_directories(path / "metadata_dropped/");
2020-03-19 21:14:52 +00:00
}
#if USE_ROCKSDB
2021-12-28 11:29:01 +00:00
/// Initialize merge tree metadata cache
if (config().has("merge_tree_metadata_cache"))
2021-12-08 02:40:59 +00:00
{
fs::create_directories(path / "rocksdb/");
size_t size = config().getUInt64("merge_tree_metadata_cache.lru_cache_size", 256 << 20);
bool continue_if_corrupted = config().getBool("merge_tree_metadata_cache.continue_if_corrupted", false);
2022-01-04 03:44:10 +00:00
try
{
2022-12-26 12:56:09 +00:00
LOG_DEBUG(log, "Initializing MergeTree metadata cache, lru_cache_size: {} continue_if_corrupted: {}",
ReadableSize(size), continue_if_corrupted);
2022-01-04 03:44:10 +00:00
global_context->initializeMergeTreeMetadataCache(path_str + "/" + "rocksdb", size);
}
catch (...)
{
if (continue_if_corrupted)
{
/// Rename rocksdb directory and reinitialize merge tree metadata cache
time_t now = time(nullptr);
fs::rename(path / "rocksdb", path / ("rocksdb.old." + std::to_string(now)));
global_context->initializeMergeTreeMetadataCache(path_str + "/" + "rocksdb", size);
}
else
{
throw;
}
2022-01-04 03:44:10 +00:00
}
2020-03-19 21:14:52 +00:00
}
#endif
2020-03-19 21:14:52 +00:00
if (config().has("interserver_http_port") && config().has("interserver_https_port"))
throw Exception(ErrorCodes::EXCESSIVE_ELEMENT_IN_CONFIG, "Both http and https interserver ports are specified");
static const auto interserver_tags =
2014-04-03 08:47:59 +00:00
{
std::make_tuple("interserver_http_host", "interserver_http_port", "http"),
std::make_tuple("interserver_https_host", "interserver_https_port", "https")
};
for (auto [host_tag, port_tag, scheme] : interserver_tags)
{
if (config().has(port_tag))
{
String this_host = config().getString(host_tag, "");
if (this_host.empty())
{
this_host = getFQDNOrHostName();
2020-05-23 22:24:01 +00:00
LOG_DEBUG(log, "Configuration parameter '{}' doesn't exist or exists and empty. Will use '{}' as replica host.",
2020-05-23 22:21:29 +00:00
host_tag, this_host);
}
String port_str = config().getString(port_tag);
int port = parse<int>(port_str);
if (port < 0 || port > 0xFFFF)
throw Exception(ErrorCodes::ARGUMENT_OUT_OF_BOUND, "Out of range '{}': {}", String(port_tag), port);
global_context->setInterserverIOAddress(this_host, port);
global_context->setInterserverScheme(scheme);
}
2014-04-03 08:47:59 +00:00
}
2022-09-05 01:50:24 +00:00
LOG_DEBUG(log, "Initializing interserver credentials.");
global_context->updateInterserverCredentials(config());
2014-08-11 15:59:01 +00:00
if (config().has("macros"))
2020-09-26 19:18:28 +00:00
global_context->setMacros(std::make_unique<Macros>(config(), "macros", log));
/// Initialize main config reloader.
std::string include_from_path = config().getString("include_from", "/etc/metrika.xml");
2019-06-20 07:17:21 +00:00
if (config().has("query_masking_rules"))
{
SensitiveDataMasker::setInstance(std::make_unique<SensitiveDataMasker>(config(), "query_masking_rules"));
2019-06-20 07:17:21 +00:00
}
2020-06-04 19:30:30 +00:00
auto main_config_reloader = std::make_unique<ConfigReloader>(
config_path,
include_from_path,
config().getString("path", ""),
std::move(main_config_zk_node_cache),
main_config_zk_changed_event,
2021-02-19 16:22:47 +00:00
[&](ConfigurationPtr config, bool initial_loading)
{
Settings::checkNoSettingNamesAtTopLevel(*config, config_path);
2020-06-04 19:30:30 +00:00
ServerSettings server_settings_;
server_settings_.loadSettingsFromConfig(*config);
2023-02-18 04:46:09 +00:00
size_t max_server_memory_usage = server_settings_.max_server_memory_usage;
double max_server_memory_usage_to_ram_ratio = server_settings_.max_server_memory_usage_to_ram_ratio;
2022-09-11 04:05:33 +00:00
size_t default_max_server_memory_usage = static_cast<size_t>(memory_amount * max_server_memory_usage_to_ram_ratio);
if (max_server_memory_usage == 0)
{
max_server_memory_usage = default_max_server_memory_usage;
LOG_INFO(log, "Setting max_server_memory_usage was set to {}"
" ({} available * {:.2f} max_server_memory_usage_to_ram_ratio)",
formatReadableSizeWithBinarySuffix(max_server_memory_usage),
formatReadableSizeWithBinarySuffix(memory_amount),
max_server_memory_usage_to_ram_ratio);
}
else if (max_server_memory_usage > default_max_server_memory_usage)
{
max_server_memory_usage = default_max_server_memory_usage;
LOG_INFO(log, "Setting max_server_memory_usage was lowered to {}"
" because the system has low amount of memory. The amount was"
" calculated as {} available"
" * {:.2f} max_server_memory_usage_to_ram_ratio",
formatReadableSizeWithBinarySuffix(max_server_memory_usage),
formatReadableSizeWithBinarySuffix(memory_amount),
max_server_memory_usage_to_ram_ratio);
}
total_memory_tracker.setHardLimit(max_server_memory_usage);
total_memory_tracker.setDescription("(total)");
total_memory_tracker.setMetric(CurrentMetrics::MemoryTracking);
2021-10-22 15:15:33 +00:00
size_t merges_mutations_memory_usage_soft_limit = server_settings_.merges_mutations_memory_usage_soft_limit;
size_t default_merges_mutations_server_memory_usage = static_cast<size_t>(memory_amount * server_settings_.merges_mutations_memory_usage_to_ram_ratio);
if (merges_mutations_memory_usage_soft_limit == 0 || merges_mutations_memory_usage_soft_limit > default_merges_mutations_server_memory_usage)
{
merges_mutations_memory_usage_soft_limit = default_merges_mutations_server_memory_usage;
LOG_WARNING(log, "Setting merges_mutations_memory_usage_soft_limit was set to {}"
" ({} available * {:.2f} merges_mutations_memory_usage_to_ram_ratio)",
formatReadableSizeWithBinarySuffix(merges_mutations_memory_usage_soft_limit),
formatReadableSizeWithBinarySuffix(memory_amount),
server_settings_.merges_mutations_memory_usage_to_ram_ratio);
}
LOG_INFO(log, "Merges and mutations memory limit is set to {}",
formatReadableSizeWithBinarySuffix(merges_mutations_memory_usage_soft_limit));
background_memory_tracker.setSoftLimit(merges_mutations_memory_usage_soft_limit);
background_memory_tracker.setDescription("(background)");
background_memory_tracker.setMetric(CurrentMetrics::MergesMutationsMemoryTracking);
total_memory_tracker.setAllowUseJemallocMemory(server_settings_.allow_use_jemalloc_memory);
2022-11-21 16:10:47 +00:00
2021-10-22 15:15:33 +00:00
auto * global_overcommit_tracker = global_context->getGlobalOvercommitTracker();
total_memory_tracker.setOvercommitTracker(global_overcommit_tracker);
// FIXME logging-related things need synchronization -- see the 'Logger * log' saved
// in a lot of places. For now, disable updating log configuration without server restart.
//setTextLog(global_context->getTextLog());
2021-09-30 19:46:12 +00:00
updateLevels(*config, logger());
global_context->setClustersConfig(config, has_zookeeper);
2020-09-26 19:18:28 +00:00
global_context->setMacros(std::make_unique<Macros>(*config, "macros", log));
global_context->setExternalAuthenticatorsConfig(*config);
2023-05-05 13:25:18 +00:00
if (global_context->isServerCompletelyStarted())
{
/// It does not make sense to reload anything before server has started.
/// Moreover, it may break initialization order.
global_context->loadOrReloadDictionaries(*config);
global_context->loadOrReloadUserDefinedExecutableFunctions(*config);
}
global_context->setRemoteHostFilter(*config);
global_context->setMaxTableSizeToDrop(server_settings_.max_table_size_to_drop);
global_context->setMaxPartitionSizeToDrop(server_settings_.max_partition_size_to_drop);
ConcurrencyControl::SlotCount concurrent_threads_soft_limit = ConcurrencyControl::Unlimited;
if (server_settings_.concurrent_threads_soft_limit_num > 0 && server_settings_.concurrent_threads_soft_limit_num < concurrent_threads_soft_limit)
concurrent_threads_soft_limit = server_settings_.concurrent_threads_soft_limit_num;
if (server_settings_.concurrent_threads_soft_limit_ratio_to_cores > 0)
{
auto value = server_settings_.concurrent_threads_soft_limit_ratio_to_cores * std::thread::hardware_concurrency();
2022-08-18 17:30:33 +00:00
if (value > 0 && value < concurrent_threads_soft_limit)
concurrent_threads_soft_limit = value;
}
ConcurrencyControl::instance().setMaxConcurrency(concurrent_threads_soft_limit);
global_context->getProcessList().setMaxSize(server_settings_.max_concurrent_queries);
global_context->getProcessList().setMaxInsertQueriesAmount(server_settings_.max_concurrent_insert_queries);
global_context->getProcessList().setMaxSelectQueriesAmount(server_settings_.max_concurrent_select_queries);
2021-10-18 15:27:51 +00:00
if (config->has("keeper_server"))
global_context->updateKeeperConfiguration(*config);
2022-04-19 15:01:41 +00:00
/// Reload the number of threads for global pools.
/// Note: If you specified it in the top level config (not it config of default profile)
/// then ClickHouse will use it exactly.
/// This is done for backward compatibility.
2023-02-22 20:15:09 +00:00
if (global_context->areBackgroundExecutorsInitialized())
2022-04-19 15:01:41 +00:00
{
auto new_pool_size = server_settings_.background_pool_size;
auto new_ratio = server_settings_.background_merges_mutations_concurrency_ratio;
global_context->getMergeMutateExecutor()->increaseThreadsAndMaxTasksCount(new_pool_size, static_cast<size_t>(new_pool_size * new_ratio));
global_context->getMergeMutateExecutor()->updateSchedulingPolicy(server_settings_.background_merges_mutations_scheduling_policy.toString());
2022-04-19 15:01:41 +00:00
}
2023-02-22 20:15:09 +00:00
if (global_context->areBackgroundExecutorsInitialized())
2022-04-19 15:01:41 +00:00
{
auto new_pool_size = server_settings_.background_move_pool_size;
2022-04-19 15:01:41 +00:00
global_context->getMovesExecutor()->increaseThreadsAndMaxTasksCount(new_pool_size, new_pool_size);
}
2023-02-22 20:15:09 +00:00
if (global_context->areBackgroundExecutorsInitialized())
2022-04-19 15:01:41 +00:00
{
auto new_pool_size = server_settings_.background_fetches_pool_size;
2022-04-19 15:01:41 +00:00
global_context->getFetchesExecutor()->increaseThreadsAndMaxTasksCount(new_pool_size, new_pool_size);
}
2023-02-22 20:15:09 +00:00
if (global_context->areBackgroundExecutorsInitialized())
2022-04-19 15:01:41 +00:00
{
auto new_pool_size = server_settings_.background_common_pool_size;
2022-04-19 15:01:41 +00:00
global_context->getCommonExecutor()->increaseThreadsAndMaxTasksCount(new_pool_size, new_pool_size);
}
global_context->getBufferFlushSchedulePool().increaseThreadsCount(server_settings_.background_buffer_flush_schedule_pool_size);
global_context->getSchedulePool().increaseThreadsCount(server_settings_.background_schedule_pool_size);
global_context->getMessageBrokerSchedulePool().increaseThreadsCount(server_settings_.background_message_broker_schedule_pool_size);
global_context->getDistributedSchedulePool().increaseThreadsCount(server_settings_.background_distributed_schedule_pool_size);
2022-04-20 13:35:13 +00:00
getIOThreadPool().reloadConfiguration(
server_settings.max_io_thread_pool_size,
server_settings.max_io_thread_pool_free_size,
server_settings.io_thread_pool_queue_size);
getBackupsIOThreadPool().reloadConfiguration(
server_settings.max_backups_io_thread_pool_size,
server_settings.max_backups_io_thread_pool_free_size,
server_settings.backups_io_thread_pool_queue_size);
getActivePartsLoadingThreadPool().reloadConfiguration(
server_settings.max_active_parts_loading_thread_pool_size,
0, // We don't need any threads once all the parts will be loaded
server_settings.max_active_parts_loading_thread_pool_size);
getOutdatedPartsLoadingThreadPool().reloadConfiguration(
server_settings.max_outdated_parts_loading_thread_pool_size,
0, // We don't need any threads once all the parts will be loaded
server_settings.max_outdated_parts_loading_thread_pool_size);
/// It could grow if we need to synchronously wait until all the data parts will be loaded.
getOutdatedPartsLoadingThreadPool().setMaxTurboThreads(
server_settings.max_active_parts_loading_thread_pool_size
);
getPartsCleaningThreadPool().reloadConfiguration(
server_settings.max_parts_cleaning_thread_pool_size,
0, // We don't need any threads one all the parts will be deleted
server_settings.max_parts_cleaning_thread_pool_size);
2022-09-27 13:26:41 +00:00
if (config->has("resources"))
{
global_context->getResourceManager()->updateConfiguration(*config);
}
2021-02-18 14:13:23 +00:00
if (!initial_loading)
{
/// We do not load ZooKeeper configuration on the first config loading
/// because TestKeeper server is not started yet.
2023-03-23 12:58:39 +00:00
if (zkutil::hasZooKeeperConfig(*config))
2021-02-18 14:13:23 +00:00
global_context->reloadZooKeeperIfChanged(config);
2021-02-18 14:13:23 +00:00
global_context->reloadAuxiliaryZooKeepersConfigIfChanged(config);
std::lock_guard lock(servers_lock);
updateServers(*config, server_pool, async_metrics, servers);
2021-02-18 14:13:23 +00:00
}
global_context->updateStorageConfiguration(*config);
global_context->updateInterserverCredentials(*config);
global_context->updateQueryCacheConfiguration(*config);
CompressionCodecEncrypted::Configuration::instance().tryLoad(*config, "encryption_codecs");
2021-07-10 02:57:44 +00:00
#if USE_SSL
CertificateReloader::instance().tryLoad(*config);
2021-07-10 02:57:44 +00:00
#endif
2022-11-15 14:49:25 +00:00
NamedCollectionUtils::reloadFromConfig(*config);
ProfileEvents::increment(ProfileEvents::MainConfigLoads);
/// Must be the last.
latest_config = config;
},
2021-02-18 14:13:23 +00:00
/* already_loaded = */ false); /// Reload it right now (initial loading)
const auto listen_hosts = getListenHosts(config());
2022-04-05 15:04:53 +00:00
const auto interserver_listen_hosts = getInterserverListenHosts(config());
const auto listen_try = getListenTry(config());
if (config().has("keeper_server"))
{
#if USE_NURAFT
//// If we don't have configured connection probably someone trying to use clickhouse-server instead
//// of clickhouse-keeper, so start synchronously.
bool can_initialize_keeper_async = false;
if (has_zookeeper) /// We have configured connection to some zookeeper cluster
{
/// If we cannot connect to some other node from our cluster then we have to wait our Keeper start
/// synchronously.
can_initialize_keeper_async = global_context->tryCheckClientConnectionToMyKeeperCluster();
}
/// Initialize keeper RAFT.
global_context->initializeKeeperDispatcher(can_initialize_keeper_async);
FourLetterCommandFactory::registerCommands(*global_context->getKeeperDispatcher());
2022-03-03 20:27:46 +00:00
auto config_getter = [this] () -> const Poco::Util::AbstractConfiguration &
{
return global_context->getConfigRef();
};
for (const auto & listen_host : listen_hosts)
{
/// TCP Keeper
const char * port_name = "keeper_server.tcp_port";
createServer(
config(), listen_host, port_name, listen_try, /* start_server: */ false,
servers_to_start_before_tables,
[&](UInt16 port) -> ProtocolServerAdapter
{
Poco::Net::ServerSocket socket;
auto address = socketBindListen(config(), socket, listen_host, port);
2023-05-10 17:30:30 +00:00
socket.setReceiveTimeout(Poco::Timespan(config().getUInt64("keeper_server.socket_receive_timeout_sec", DBMS_DEFAULT_RECEIVE_TIMEOUT_SEC), 0));
socket.setSendTimeout(Poco::Timespan(config().getUInt64("keeper_server.socket_send_timeout_sec", DBMS_DEFAULT_SEND_TIMEOUT_SEC), 0));
return ProtocolServerAdapter(
listen_host,
port_name,
"Keeper (tcp): " + address.toString(),
std::make_unique<TCPServer>(
2022-03-03 20:27:46 +00:00
new KeeperTCPHandlerFactory(
config_getter, global_context->getKeeperDispatcher(),
2022-03-22 10:41:50 +00:00
global_context->getSettingsRef().receive_timeout.totalSeconds(),
global_context->getSettingsRef().send_timeout.totalSeconds(),
2022-03-03 20:27:46 +00:00
false), server_pool, socket));
});
const char * secure_port_name = "keeper_server.tcp_port_secure";
createServer(
config(), listen_host, secure_port_name, listen_try, /* start_server: */ false,
servers_to_start_before_tables,
[&](UInt16 port) -> ProtocolServerAdapter
{
#if USE_SSL
Poco::Net::SecureServerSocket socket;
auto address = socketBindListen(config(), socket, listen_host, port, /* secure = */ true);
2023-05-10 17:30:30 +00:00
socket.setReceiveTimeout(Poco::Timespan(config().getUInt64("keeper_server.socket_receive_timeout_sec", DBMS_DEFAULT_RECEIVE_TIMEOUT_SEC), 0));
socket.setSendTimeout(Poco::Timespan(config().getUInt64("keeper_server.socket_send_timeout_sec", DBMS_DEFAULT_SEND_TIMEOUT_SEC), 0));
return ProtocolServerAdapter(
listen_host,
secure_port_name,
"Keeper with secure protocol (tcp_secure): " + address.toString(),
std::make_unique<TCPServer>(
2022-03-03 20:27:46 +00:00
new KeeperTCPHandlerFactory(
config_getter, global_context->getKeeperDispatcher(),
2022-03-22 10:41:50 +00:00
global_context->getSettingsRef().receive_timeout.totalSeconds(),
global_context->getSettingsRef().send_timeout.totalSeconds(), true), server_pool, socket));
#else
UNUSED(port);
throw Exception(ErrorCodes::SUPPORT_IS_DISABLED, "SSL support for TCP protocol is disabled because Poco library was built without NetSSL support.");
#endif
});
}
#else
throw Exception(ErrorCodes::SUPPORT_IS_DISABLED, "ClickHouse server built without NuRaft library. Cannot use internal coordination.");
#endif
}
for (auto & server : servers_to_start_before_tables)
{
server.start();
LOG_INFO(log, "Listening for {}", server.getDescription());
}
/// Initialize access storages.
auto & access_control = global_context->getAccessControl();
try
{
access_control.setUpFromMainConfig(config(), config_path, [&] { return global_context->getZooKeeper(); });
}
catch (...)
{
2022-12-06 10:04:15 +00:00
tryLogCurrentException(log, "Caught exception while setting up access control.");
throw;
}
/// Reload config in SYSTEM RELOAD CONFIG query.
2018-05-07 02:01:11 +00:00
global_context->setConfigReloadCallback([&]()
{
main_config_reloader->reload();
2022-09-16 11:19:39 +00:00
access_control.reload(AccessControl::ReloadMode::USERS_CONFIG_ONLY);
});
2017-08-09 15:34:09 +00:00
/// Limit on total number of concurrently executed queries.
2023-02-18 04:46:09 +00:00
global_context->getProcessList().setMaxSize(server_settings.max_concurrent_queries);
2019-02-09 22:38:26 +00:00
/// Set up caches.
2023-02-18 04:46:09 +00:00
size_t max_cache_size = static_cast<size_t>(memory_amount * server_settings.cache_size_to_ram_max_ratio);
2019-02-09 22:38:26 +00:00
2023-02-18 04:46:09 +00:00
String uncompressed_cache_policy = server_settings.uncompressed_cache_policy;
2022-06-14 07:08:40 +00:00
LOG_INFO(log, "Uncompressed cache policy name {}", uncompressed_cache_policy);
2023-02-18 04:46:09 +00:00
size_t uncompressed_cache_size = server_settings.uncompressed_cache_size;
2019-02-09 22:38:26 +00:00
if (uncompressed_cache_size > max_cache_size)
{
2019-02-09 22:38:26 +00:00
uncompressed_cache_size = max_cache_size;
LOG_INFO(log, "Uncompressed cache size was lowered to {} because the system has low amount of memory",
formatReadableSizeWithBinarySuffix(uncompressed_cache_size));
}
global_context->setUncompressedCache(uncompressed_cache_policy, uncompressed_cache_size);
/// Load global settings from default_profile and system_profile.
global_context->setDefaultProfiles(config());
const Settings & settings = global_context->getSettingsRef();
/// Initialize background executors after we load default_profile config.
/// This is needed to load proper values of background_pool_size etc.
2021-10-19 08:19:43 +00:00
global_context->initializeBackgroundExecutorsIfNeeded();
if (settings.async_insert_threads)
global_context->setAsynchronousInsertQueue(std::make_shared<AsynchronousInsertQueue>(
global_context,
settings.async_insert_threads));
2021-04-19 19:16:34 +00:00
2023-02-18 04:46:09 +00:00
size_t mark_cache_size = server_settings.mark_cache_size;
String mark_cache_policy = server_settings.mark_cache_policy;
2019-02-09 22:38:26 +00:00
if (!mark_cache_size)
2020-05-23 22:24:01 +00:00
LOG_ERROR(log, "Too low mark cache size will lead to severe performance degradation.");
2019-02-09 22:38:26 +00:00
if (mark_cache_size > max_cache_size)
{
2019-02-09 22:38:26 +00:00
mark_cache_size = max_cache_size;
LOG_INFO(log, "Mark cache size was lowered to {} because the system has low amount of memory",
formatReadableSizeWithBinarySuffix(mark_cache_size));
}
global_context->setMarkCache(mark_cache_policy, mark_cache_size);
2023-02-18 04:46:09 +00:00
if (server_settings.index_uncompressed_cache_size)
global_context->setIndexUncompressedCache(server_settings.index_uncompressed_cache_size);
2023-02-18 04:46:09 +00:00
if (server_settings.index_mark_cache_size)
global_context->setIndexMarkCache(server_settings.index_mark_cache_size);
2023-02-18 04:46:09 +00:00
if (server_settings.mmap_cache_size)
global_context->setMMappedFileCache(server_settings.mmap_cache_size);
2022-11-29 13:15:28 +00:00
/// A cache for query results.
global_context->setQueryCache(config());
2022-11-29 13:15:28 +00:00
2018-09-06 19:13:49 +00:00
#if USE_EMBEDDED_COMPILER
/// 128 MB
constexpr size_t compiled_expression_cache_size_default = 1024 * 1024 * 128;
2021-05-04 20:32:43 +00:00
size_t compiled_expression_cache_size = config().getUInt64("compiled_expression_cache_size", compiled_expression_cache_size_default);
constexpr size_t compiled_expression_cache_elements_size_default = 10000;
size_t compiled_expression_cache_elements_size = config().getUInt64("compiled_expression_cache_elements_size", compiled_expression_cache_elements_size_default);
CompiledExpressionCacheFactory::instance().init(compiled_expression_cache_size, compiled_expression_cache_elements_size);
2018-09-06 19:13:49 +00:00
#endif
/// Set path for format schema files
2021-08-16 18:30:53 +00:00
fs::path format_schema_path(config().getString("format_schema_path", path / "format_schemas/"));
2021-05-16 22:06:09 +00:00
global_context->setFormatSchemaPath(format_schema_path);
fs::create_directories(format_schema_path);
2020-07-30 19:08:13 +00:00
/// Check sanity of MergeTreeSettings on server startup
{
size_t background_pool_tasks = global_context->getMergeMutateExecutor()->getMaxTasksCount();
global_context->getMergeTreeSettings().sanityCheck(background_pool_tasks);
global_context->getReplicatedMergeTreeSettings().sanityCheck(background_pool_tasks);
}
/// try set up encryption. There are some errors in config, error will be printed and server wouldn't start.
CompressionCodecEncrypted::Configuration::instance().load(config(), "encryption_codecs");
2020-11-23 10:55:00 +00:00
SCOPE_EXIT({
async_metrics.stop();
2020-11-23 10:55:00 +00:00
/** Ask to cancel background jobs all table engines,
* and also query_log.
* It is important to do early, not in destructor of Context, because
* table engines could use Context on destroy.
*/
LOG_INFO(log, "Shutting down storages.");
global_context->shutdown();
LOG_DEBUG(log, "Shut down storages.");
if (!servers_to_start_before_tables.empty())
{
LOG_DEBUG(log, "Waiting for current connections to servers for tables to finish.");
size_t current_connections = 0;
for (auto & server : servers_to_start_before_tables)
{
server.stop();
current_connections += server.currentConnections();
}
if (current_connections)
LOG_INFO(log, "Closed all listening sockets. Waiting for {} outstanding connections.", current_connections);
else
LOG_INFO(log, "Closed all listening sockets.");
if (current_connections > 0)
current_connections = waitServersToFinish(servers_to_start_before_tables, config().getInt("shutdown_wait_unfinished", 5));
if (current_connections)
LOG_INFO(log, "Closed connections to servers for tables. But {} remain. Probably some tables of other users cannot finish their connections after context shutdown.", current_connections);
else
LOG_INFO(log, "Closed connections to servers for tables.");
2021-01-27 17:54:25 +00:00
global_context->shutdownKeeperDispatcher();
}
2020-11-23 10:55:00 +00:00
Fix SIGSEGV by waiting servers thread pool It is easy to reproduce with shutdown_wait_unfinished=0: ================================================================= ==13442==ERROR: AddressSanitizer: heap-use-after-free on address 0x611000210f30 at pc 0x00000a8e55a0 bp 0x7fff2b83e270 sp 0x7fff2b83e268 WRITE of size 8 at 0x611000210f30 thread T2 (TCPHandler) 0 0xa8e559f in long std::__1::__cxx_atomic_fetch_add<long>(std::__1::__cxx_atomic_base_impl<long>*, long, std::__1::memory_order) obj-x86_64-linux-gnu/../contrib/libcxx/include/atomic:1050:12 1 0xa8e559f in std::__1::__atomic_base<long, true>::fetch_add(long, std::__1::memory_order) obj-x86_64-linux-gnu/../contrib/libcxx/include/atomic:1719:17 2 0xa8e559f in MemoryTracker::alloc(long) obj-x86_64-linux-gnu/../src/Common/MemoryTracker.cpp:146:35 3 0xa8e510c in MemoryTracker::alloc(long) obj-x86_64-linux-gnu/../src/Common/MemoryTracker.cpp 4 0xa90b474 in DB::ThreadStatus::~ThreadStatus() obj-x86_64-linux-gnu/../src/Common/ThreadStatus.cpp:92:28 5 0x1f90ee83 in DB::TCPHandler::runImpl() obj-x86_64-linux-gnu/../src/Server/TCPHandler.cpp:450:1 6 0x1f92dcac in DB::TCPHandler::run() obj-x86_64-linux-gnu/../src/Server/TCPHandler.cpp:1492:9 7 0x25bdc2fe in Poco::Net::TCPServerConnection::start() obj-x86_64-linux-gnu/../contrib/poco/Net/src/TCPServerConnection.cpp:43:3 8 0x25bdce1b in Poco::Net::TCPServerDispatcher::run() obj-x86_64-linux-gnu/../contrib/poco/Net/src/TCPServerDispatcher.cpp:113:19 9 0x25e9c784 in Poco::PooledThread::run() obj-x86_64-linux-gnu/../contrib/poco/Foundation/src/ThreadPool.cpp:199:14 10 0x25e96cd6 in Poco::ThreadImpl::runnableEntry(void*) obj-x86_64-linux-gnu/../contrib/poco/Foundation/src/Thread_POSIX.cpp:345:27 11 0x7ffff7f723e8 in start_thread (/usr/lib/libpthread.so.0+0x93e8) 12 0x7ffff7ea0292 in clone (/usr/lib/libc.so.6+0x100292) 0x611000210f30 is located 112 bytes inside of 216-byte region [0x611000210ec0,0x611000210f98) freed by thread T0 here: 0 0xa845d02 in operator delete(void*, unsigned long) (/src/ch/tmp/upstream/clickhouse-asan+0xa845d02) 1 0x1d38328c in void std::__1::__libcpp_operator_delete<void*, unsigned long>(void*, unsigned long) obj-x86_64-linux-gnu/../contrib/libcxx/include/new:245:3 2 0x1d38328c in void std::__1::__do_deallocate_handle_size<>(void*, unsigned long) obj-x86_64-linux-gnu/../contrib/libcxx/include/new:271:10 3 0x1d38328c in std::__1::__libcpp_deallocate(void*, unsigned long, unsigned long) obj-x86_64-linux-gnu/../contrib/libcxx/include/new:285:14 4 0x1d38328c in std::__1::allocator<std::__1::__hash_node<std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser>, void*> >::deallocate(std::__1::__hash_node<std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser>, void*>*, unsigned long) obj-x86_64-linux-gnu/../contrib/libcxx/include/memory:849:13 5 0x1d38328c in std::__1::allocator_traits<std::__1::allocator<std::__1::__hash_node<std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser>, void*> > >::deallocate(std::__1::allocator<std::__1::__hash_node<std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser>, void*> >&, std::__1::__hash_node<std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser>, void*>*, unsigned long) obj-x86_64-linux-gnu/../contrib/libcxx/include/__memory/allocator_traits.h:476:14 6 0x1d38328c in std::__1::__hash_table<std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser>, std::__1::__unordered_map_hasher<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser>, std::__1::hash<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > >, std::__1::equal_to<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > >, true>, std::__1::__unordered_map_equal<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser>, std::__1::equal_to<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > >, std::__1::hash<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > >, true>, std::__1::allocator<std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser> > >::__deallocate_node(std::__1::__hash_node_base<std::__1::__hash_node<std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser>, void*>*>*) obj-x86_64-linux-gnu/../contrib/libcxx/include/__hash_table:1581:9 7 0x1d38328c in std::__1::__hash_table<std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser>, std::__1::__unordered_map_hasher<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser>, std::__1::hash<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > >, std::__1::equal_to<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > >, true>, std::__1::__unordered_map_equal<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser>, std::__1::equal_to<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > >, std::__1::hash<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > >, true>, std::__1::allocator<std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser> > >::~__hash_table() obj-x86_64-linux-gnu/../contrib/libcxx/include/__hash_table:1519:5 8 0x1d38328c in std::__1::unordered_map<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser, std::__1::hash<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > >, std::__1::equal_to<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > >, std::__1::allocator<std::__1::pair<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const, DB::ProcessListForUser> > >::~unordered_map() obj-x86_64-linux-gnu/../contrib/libcxx/include/unordered_map:1044:5 9 0x1d38328c in DB::ProcessList::~ProcessList() obj-x86_64-linux-gnu/../src/Interpreters/ProcessList.h:263:7 10 0x1d38169c in DB::ContextShared::~ContextShared() obj-x86_64-linux-gnu/../src/Interpreters/Context.cpp:417:5 11 0x1d32f3e5 in std::__1::default_delete<DB::ContextShared>::operator()(DB::ContextShared*) const obj-x86_64-linux-gnu/../contrib/libcxx/include/memory:1397:5 12 0x1d32f3e5 in std::__1::unique_ptr<DB::ContextShared, std::__1::default_delete<DB::ContextShared> >::reset(DB::ContextShared*) obj-x86_64-linux-gnu/../contrib/libcxx/include/memory:1658:7 13 0x1d32f3e5 in DB::SharedContextHolder::reset() obj-x86_64-linux-gnu/../src/Interpreters/Context.cpp:485:44 14 0xa8863d4 in DB::Server::main(std::__1::vector<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::allocator<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > > > const&)::$_5::operator()() const obj-x86_64-linux-gnu/../programs/server/Server.cpp:880:5 15 0xa8863d4 in ext::basic_scope_guard<DB::Server::main(std::__1::vector<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::allocator<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > > > const&)::$_5>::invoke() obj-x86_64-linux-gnu/../base/common/../ext/scope_guard.h:97:9 16 0xa8863d4 in ext::basic_scope_guard<DB::Server::main(std::__1::vector<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::allocator<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > > > const&)::$_5>::~basic_scope_guard() obj-x86_64-linux-gnu/../base/common/../ext/scope_guard.h:47:28 17 0xa86d889 in DB::Server::main(std::__1::vector<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::allocator<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > > > const&) obj-x86_64-linux-gnu/../programs/server/Server.cpp:1379:1 18 0x25c0c8b5 in Poco::Util::Application::run() obj-x86_64-linux-gnu/../contrib/poco/Util/src/Application.cpp:334:8 19 0xa85070d in DB::Server::run() obj-x86_64-linux-gnu/../programs/server/Server.cpp:340:25 20 0x25c49eb7 in Poco::Util::ServerApplication::run(int, char**) obj-x86_64-linux-gnu/../contrib/poco/Util/src/ServerApplication.cpp:611:9 21 0xa84cd11 in mainEntryClickHouseServer(int, char**) obj-x86_64-linux-gnu/../programs/server/Server.cpp:132:20 22 0xa848c3a in main obj-x86_64-linux-gnu/../programs/main.cpp:368:12 23 0x7ffff7dc8151 in __libc_start_main (/usr/lib/libc.so.6+0x28151) previously allocated by thread T2 (TCPHandler) here: 0 0xa84509d in operator new(unsigned long) (/src/ch/tmp/upstream/clickhouse-asan+0xa84509d) 1 0x1e2a7aa6 in void* std::__1::__libcpp_operator_new<unsigned long>(unsigned long) obj-x86_64-linux-gnu/../contrib/libcxx/include/new:235:10 2 0x1e2a7aa6 in std::__1::__libcpp_allocate(unsigned long, unsigned long) obj-x86_64-linux-gnu/../contrib/libcxx/include/new:261:10 3 0x1e2a7aa6 in std::__1::allocator<std::__1::__hash_node<std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser>, void*> >::allocate(unsigned long) obj-x86_64-linux-gnu/../contrib/libcxx/include/memory:840:38 4 0x1e2a7aa6 in std::__1::allocator_traits<std::__1::allocator<std::__1::__hash_node<std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser>, void*> > >::allocate(std::__1::allocator<std::__1::__hash_node<std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser>, void*> >&, unsigned long) obj-x86_64-linux-gnu/../contrib/libcxx/include/__memory/allocator_traits.h:468:21 5 0x1e2a7aa6 in std::__1::unique_ptr<std::__1::__hash_node<std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser>, void*>, std::__1::__hash_node_destructor<std::__1::allocator<std::__1::__hash_node<std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser>, void*> > > > std::__1::__hash_table<std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser>, std::__1::__unordered_map_hasher<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser>, std::__1::hash<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > >, std::__1::equal_to<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > >, true>, std::__1::__unordered_map_equal<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser>, std::__1::equal_to<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > >, std::__1::hash<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > >, true>, std::__1::allocator<std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser> > >::__construct_node_hash<std::__1::piecewise_construct_t const&, std::__1::tuple<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&>, std::__1::tuple<> >(unsigned long, std::__1::piecewise_construct_t const&, std::__1::tuple<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&>&&, std::__1::tuple<>&&) obj-x86_64-linux-gnu/../contrib/libcxx/include/__hash_table:2472:23 6 0x1e2a7aa6 in std::__1::pair<std::__1::__hash_iterator<std::__1::__hash_node<std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser>, void*>*>, bool> std::__1::__hash_table<std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser>, std::__1::__unordered_map_hasher<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser>, std::__1::hash<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > >, std::__1::equal_to<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > >, true>, std::__1::__unordered_map_equal<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser>, std::__1::equal_to<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > >, std::__1::hash<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > >, true>, std::__1::allocator<std::__1::__hash_value_type<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser> > >::__emplace_unique_key_args<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::piecewise_construct_t const&, std::__1::tuple<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&>, std::__1::tuple<> >(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, std::__1::piecewise_construct_t const&, std::__1::tuple<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&>&&, std::__1::tuple<>&&) obj-x86_64-linux-gnu/../contrib/libcxx/include/__hash_table:2093:29 7 0x1e29c13d in std::__1::unordered_map<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, DB::ProcessListForUser, std::__1::hash<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > >, std::__1::equal_to<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > >, std::__1::allocator<std::__1::pair<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const, DB::ProcessListForUser> > >::operator[](std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&) obj-x86_64-linux-gnu/../contrib/libcxx/include/unordered_map:1740:21 8 0x1e29c13d in DB::ProcessList::insert(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, DB::IAST const*, DB::Context&) obj-x86_64-linux-gnu/../src/Interpreters/ProcessList.cpp:183:50 9 0x1e5a3a58 in DB::executeQueryImpl(char const*, char const*, DB::Context&, bool, DB::QueryProcessingStage::Enum, bool, DB::ReadBuffer*) obj-x86_64-linux-gnu/../src/Interpreters/executeQuery.cpp:486:59 10 0x1e5a153e in DB::executeQuery(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, DB::Context&, bool, DB::QueryProcessingStage::Enum, bool) obj-x86_64-linux-gnu/../src/Interpreters/executeQuery.cpp:904:30 11 0x1f909bdc in DB::TCPHandler::runImpl() obj-x86_64-linux-gnu/../src/Server/TCPHandler.cpp:289:24 12 0x1f92dcac in DB::TCPHandler::run() obj-x86_64-linux-gnu/../src/Server/TCPHandler.cpp:1492:9 13 0x25bdc2fe in Poco::Net::TCPServerConnection::start() obj-x86_64-linux-gnu/../contrib/poco/Net/src/TCPServerConnection.cpp:43:3 14 0x25bdce1b in Poco::Net::TCPServerDispatcher::run() obj-x86_64-linux-gnu/../contrib/poco/Net/src/TCPServerDispatcher.cpp:113:19 15 0x25e9c784 in Poco::PooledThread::run() obj-x86_64-linux-gnu/../contrib/poco/Foundation/src/ThreadPool.cpp:199:14 16 0x25e96cd6 in Poco::ThreadImpl::runnableEntry(void*) obj-x86_64-linux-gnu/../contrib/poco/Foundation/src/Thread_POSIX.cpp:345:27 17 0x7ffff7f723e8 in start_thread (/usr/lib/libpthread.so.0+0x93e8) Thread T2 (TCPHandler) created by T0 here: 0 0xa7ffe0a in pthread_create (/src/ch/tmp/upstream/clickhouse-asan+0xa7ffe0a) 1 0x25e9606f in Poco::ThreadImpl::startImpl(Poco::SharedPtr<Poco::Runnable, Poco::ReferenceCounter, Poco::ReleasePolicy<Poco::Runnable> >) obj-x86_64-linux-gnu/../contrib/poco/Foundation/src/Thread_POSIX.cpp:202:6 2 0x25e98eea in Poco::Thread::start(Poco::Runnable&) obj-x86_64-linux-gnu/../contrib/poco/Foundation/src/Thread.cpp:128:2 3 0x25e9cd28 in Poco::PooledThread::start() obj-x86_64-linux-gnu/../contrib/poco/Foundation/src/ThreadPool.cpp:85:10 4 0x25e9cd28 in Poco::ThreadPool::ThreadPool(int, int, int, int) obj-x86_64-linux-gnu/../contrib/poco/Foundation/src/ThreadPool.cpp:252:12 5 0xa865aff in DB::Server::main(std::__1::vector<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::allocator<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > > > const&) obj-x86_64-linux-gnu/../programs/server/Server.cpp:831:22 6 0x25c0c8b5 in Poco::Util::Application::run() obj-x86_64-linux-gnu/../contrib/poco/Util/src/Application.cpp:334:8 7 0xa85070d in DB::Server::run() obj-x86_64-linux-gnu/../programs/server/Server.cpp:340:25 8 0x25c49eb7 in Poco::Util::ServerApplication::run(int, char**) obj-x86_64-linux-gnu/../contrib/poco/Util/src/ServerApplication.cpp:611:9 9 0xa84cd11 in mainEntryClickHouseServer(int, char**) obj-x86_64-linux-gnu/../programs/server/Server.cpp:132:20 10 0xa848c3a in main obj-x86_64-linux-gnu/../programs/main.cpp:368:12 11 0x7ffff7dc8151 in __libc_start_main (/usr/lib/libc.so.6+0x28151) SUMMARY: AddressSanitizer: heap-use-after-free obj-x86_64-linux-gnu/../contrib/libcxx/include/atomic:1050:12 in long std::__1::__cxx_atomic_fetch_add<long>(std::__1::__cxx_atomic_base_impl<long>*, long, std::__1::memory_order) Shadow bytes around the buggy address: 0x0c228003a190: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd 0x0c228003a1a0: fd fd fd fd fd fd fa fa fa fa fa fa fa fa fa fa 0x0c228003a1b0: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd 0x0c228003a1c0: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fa fa 0x0c228003a1d0: fa fa fa fa fa fa fa fa fd fd fd fd fd fd fd fd =>0x0c228003a1e0: fd fd fd fd fd fd[fd]fd fd fd fd fd fd fd fd fd 0x0c228003a1f0: fd fd fd fa fa fa fa fa fa fa fa fa fa fa fa fa 0x0c228003a200: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd 0x0c228003a210: fd fd fd fd fd fd fd fd fd fa fa fa fa fa fa fa 0x0c228003a220: fa fa fa fa fa fa fa fa fd fd fd fd fd fd fd fd 0x0c228003a230: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd Shadow byte legend (one shadow byte represents 8 application bytes): Addressable: 00 Partially addressable: 01 02 03 04 05 06 07 Heap left redzone: fa Freed heap region: fd Stack left redzone: f1 Stack mid redzone: f2 Stack right redzone: f3 Stack after return: f5 Stack use after scope: f8 Global redzone: f9 Global init order: f6 Poisoned by user: f7 Container overflow: fc Array cookie: ac Intra object redzone: bb ASan internal: fe Left alloca redzone: ca Right alloca redzone: cb Shadow gap: cc ==13442==ABORTING 2021.02.20 16:39:50.861426 [ 13443 ] {} <Trace> BaseDaemon: Received signal -3 2021.02.20 16:39:50.861668 [ 14989 ] {} <Fatal> BaseDaemon: ######################################## 2021.02.20 16:39:50.861749 [ 14989 ] {} <Fatal> BaseDaemon: (version 21.3.1.6073 (official build), build id: AC8A516D2F60B8505FA128074527EC2C86198E64) (from thread 13874) (no query) Received signal Unknown signal (-3) 2021.02.20 16:39:50.861810 [ 14989 ] {} <Fatal> BaseDaemon: Sanitizer trap. 2021.02.20 16:39:50.861880 [ 14989 ] {} <Fatal> BaseDaemon: Stack trace: 0xa8e94a7 0xad25b1b 0xa831a16 0xa819444 0xa81aefe 0xa81bb4b 0xa8e55a0 0xa8e510d 0xa90b475 0x1f90ee84 0x1f92dcad 0x25bdc2ff 0x25bdce1c 0x25e9c785 0x25e96cd7 0x7ffff7f723e9 0x7ffff7ea0293 2021.02.20 16:39:50.903643 [ 14989 ] {} <Fatal> BaseDaemon: 0.1. inlined from ./obj-x86_64-linux-gnu/../src/Common/StackTrace.cpp:298: StackTrace::tryCapture() 2021.02.20 16:39:50.903708 [ 14989 ] {} <Fatal> BaseDaemon: 0. ../src/Common/StackTrace.cpp:259: StackTrace::StackTrace() @ 0xa8e94a7 in /src/ch/tmp/upstream/clickhouse-asan 2021.02.20 16:39:51.041733 [ 14989 ] {} <Fatal> BaseDaemon: 1.1. inlined from ./obj-x86_64-linux-gnu/../src/Common/CurrentThread.h:78: DB::CurrentThread::getQueryId() 2021.02.20 16:39:51.041768 [ 14989 ] {} <Fatal> BaseDaemon: 1. ../base/daemon/BaseDaemon.cpp:381: sanitizerDeathCallback() @ 0xad25b1b in /src/ch/tmp/upstream/clickhouse-asan 2021.02.20 16:39:52.551623 [ 13442 ] {} <Information> Application: shutting down 2021.02.20 16:39:52.551696 [ 13442 ] {} <Debug> Application: Uninitializing subsystem: Logging Subsystem 2021.02.20 16:39:52.551792 [ 13443 ] {} <Trace> BaseDaemon: Received signal -2 2021.02.20 16:39:52.551831 [ 13443 ] {} <Information> BaseDaemon: Stop SignalListener thread
2021-02-20 13:46:55 +00:00
/// Wait server pool to avoid use-after-free of destroyed context in the handlers
server_pool.joinAll();
2020-11-23 10:55:00 +00:00
/** Explicitly destroy Context. It is more convenient than in destructor of Server, because logger is still available.
* At this moment, no one could own shared part of Context.
*/
global_context.reset();
shared_context.reset();
LOG_DEBUG(log, "Destroyed global context.");
});
/// DNSCacheUpdater uses BackgroundSchedulePool which lives in shared context
/// and thus this object must be created after the SCOPE_EXIT object where shared
/// context is destroyed.
/// In addition this object has to be created before the loading of the tables.
std::unique_ptr<DNSCacheUpdater> dns_cache_updater;
2023-02-18 04:46:09 +00:00
if (server_settings.disable_internal_dns_cache)
{
/// Disable DNS caching at all
DNSResolver::instance().setDisableCacheFlag();
LOG_DEBUG(log, "DNS caching disabled");
}
else
{
/// Initialize a watcher periodically updating DNS cache
dns_cache_updater = std::make_unique<DNSCacheUpdater>(
2023-02-18 04:46:09 +00:00
global_context, server_settings.dns_cache_update_period, server_settings.dns_max_consecutive_failures);
}
if (dns_cache_updater)
dns_cache_updater->start();
2020-09-25 11:27:00 +00:00
/// Set current database name before loading tables and databases because
/// system logs may copy global context.
global_context->setCurrentDatabaseNameInGlobalContext(default_database);
2021-08-16 18:30:53 +00:00
LOG_INFO(log, "Loading metadata from {}", path_str);
2019-07-30 14:04:18 +00:00
2018-12-10 23:21:03 +00:00
try
{
2021-09-03 19:21:01 +00:00
auto & database_catalog = DatabaseCatalog::instance();
/// We load temporary database first, because projections need it.
database_catalog.initializeAndLoadTemporaryDatabase();
loadMetadataSystem(global_context);
maybeConvertSystemDatabase(global_context);
2018-12-10 23:21:03 +00:00
/// After attaching system databases we can initialize system log.
global_context->initializeSystemLogs();
global_context->setSystemZooKeeperLogAfterInitializationIfNeeded();
2018-12-10 23:21:03 +00:00
/// After the system database is created, attach virtual system tables (in addition to query_log and part_log)
attachSystemTablesServer(global_context, *database_catalog.getSystemDatabase(), has_zookeeper);
attachInformationSchema(global_context, *database_catalog.getDatabase(DatabaseCatalog::INFORMATION_SCHEMA));
attachInformationSchema(global_context, *database_catalog.getDatabase(DatabaseCatalog::INFORMATION_SCHEMA_UPPERCASE));
Fix UUID overlap in DROP TABLE for internal DDL from MaterializeMySQL This will fix race with DatabaseCatalog::loadMarkedAsDroppedTables(), since MaterializeMySQL, and MaterializedMySQLSyncThread in background, will be started earlier then DatabaseCatalog::loadMarkedAsDroppedTables() and will move those tables to metadata_dropped, and after loadMarkedAsDroppedTables() will start and try to load partially dropped tables and will hit UUID overlap: 12:02:51.536783 [ 3026034 ] {} <Information> Application: starting up 12:02:53.019282 [ 3026034 ] {} <Information> DatabaseMaterializeMySQL<Atomic> (mysql): Total 9 tables and 0 dictionaries. 12:02:53.041699 [ 3026200 ] {} <Debug> mysql.data (7143b65f-6982-4600-b143-b65f6982e600): Loading data parts 12:02:53.041740 [ 3026200 ] {} <Debug> mysql.data (7143b65f-6982-4600-b143-b65f6982e600): There are no data parts 12:02:53.620382 [ 3026034 ] {} <Information> DatabaseMaterializeMySQL<Atomic> (mysql): Starting up tables. 12:03:00.669730 [ 3026183 ] {} <Debug> executeQuery: (internal) /*Materialize MySQL step 1: execute MySQL DDL for dump data*/ DROP TABLE mysql.data 12:03:00.741894 [ 3026269 ] {} <Information> DatabaseCatalog: Trying load partially dropped table mysql.data (7143b65f-6982-4600-b143-b65f6982e600) from /var/lib/clickhouse/metadata_dropped/mysql.data.7143b65f-6982-4600-b143-b65f6982e600.sql 12:03:00.742582 [ 3026269 ] {} <Debug> mysql.data (7143b65f-6982-4600-b143-b65f6982e600): Loading data parts 12:03:00.742650 [ 3026269 ] {} <Debug> mysql.data (7143b65f-6982-4600-b143-b65f6982e600): There are no data parts 12:03:00.773137 [ 3026034 ] {} <Error> Application: Caught exception while loading metadata: Code: 57, e.displayText() = DB::Exception: Mapping for table with UUID=7143b65f-6982-4600-b143-b65f6982e600 already exists. It happened due to UUID collision, most likely because some not random UUIDs were manually specified in CREATE queries., Stack trace (when copying this message, always include the lines below): 12:03:01.224557 [ 3026034 ] {} <Error> Application: DB::Exception: Mapping for table with UUID=7143b65f-6982-4600-b143-b65f6982e600 already exists. It happened due to UUID collision, most likely because some not random UUIDs were manually specified in CREATE queries. Cc: @zhang2014
2021-09-02 18:05:13 +00:00
/// Firstly remove partially dropped databases, to avoid race with MaterializedMySQLSyncThread,
/// that may execute DROP before loadMarkedAsDroppedTables() in background,
/// and so loadMarkedAsDroppedTables() will find it and try to add, and UUID will overlap.
database_catalog.loadMarkedAsDroppedTables();
2023-02-17 16:57:49 +00:00
database_catalog.createBackgroundTasks();
2018-12-10 23:21:03 +00:00
/// Then, load remaining databases
loadMetadata(global_context, default_database);
convertDatabasesEnginesIfNeed(global_context);
2021-09-01 19:42:49 +00:00
startupSystemTables();
2023-02-17 16:57:49 +00:00
database_catalog.startupBackgroundCleanup();
2020-09-25 11:27:00 +00:00
/// After loading validate that default database exists
database_catalog.assertDatabaseExists(default_database);
/// Load user-defined SQL functions.
global_context->getUserDefinedSQLObjectsLoader().loadObjects();
2018-12-10 23:21:03 +00:00
}
catch (...)
{
tryLogCurrentException(log, "Caught exception while loading metadata");
throw;
}
2020-05-23 22:24:01 +00:00
LOG_DEBUG(log, "Loaded metadata.");
2019-02-03 21:30:45 +00:00
/// Init trace collector only after trace_log system table was created
/// Disable it if we collect test coverage information, because it will work extremely slow.
2022-06-17 14:44:09 +00:00
#if USE_UNWIND && !WITH_COVERAGE
2020-04-22 17:52:21 +00:00
/// Profilers cannot work reliably with any other libunwind or without PHDR cache.
2019-07-25 22:35:47 +00:00
if (hasPHDRCache())
2020-04-22 17:52:21 +00:00
{
2019-07-25 22:35:47 +00:00
global_context->initializeTraceCollector();
2020-04-22 17:52:21 +00:00
/// Set up server-wide memory profiler (for total memory tracker).
UInt64 total_memory_profiler_step = config().getUInt64("total_memory_profiler_step", 0);
if (total_memory_profiler_step)
{
total_memory_tracker.setProfilerStep(total_memory_profiler_step);
}
2020-04-30 13:25:17 +00:00
double total_memory_tracker_sample_probability = config().getDouble("total_memory_tracker_sample_probability", 0);
2022-09-11 04:05:33 +00:00
if (total_memory_tracker_sample_probability > 0.0)
2020-04-30 13:25:17 +00:00
{
total_memory_tracker.setSampleProbability(total_memory_tracker_sample_probability);
}
2020-04-22 17:52:21 +00:00
}
2019-08-02 18:51:39 +00:00
#endif
2019-02-03 21:30:45 +00:00
/// Describe multiple reasons when query profiler cannot work.
#if !USE_UNWIND
2020-05-23 22:24:01 +00:00
LOG_INFO(log, "Query Profiler and TraceCollector are disabled because they cannot work without bundled unwind (stack unwinding) library.");
#endif
#if WITH_COVERAGE
2020-05-23 22:24:01 +00:00
LOG_INFO(log, "Query Profiler and TraceCollector are disabled because they work extremely slow with test coverage.");
#endif
#if defined(SANITIZER)
LOG_INFO(log, "Query Profiler disabled because they cannot work under sanitizers"
2020-05-23 22:21:29 +00:00
" when two different stack unwinding methods will interfere with each other.");
#endif
2021-04-04 01:52:41 +00:00
#if !defined(__x86_64__)
LOG_INFO(log, "Query Profiler and TraceCollector is only tested on x86_64. It also known to not work under qemu-user.");
2021-04-04 01:52:41 +00:00
#endif
if (!hasPHDRCache())
2020-05-23 22:24:01 +00:00
LOG_INFO(log, "Query Profiler and TraceCollector are disabled because they require PHDR cache to be created"
2020-05-23 22:21:29 +00:00
" (otherwise the function 'dl_iterate_phdr' is not lock free and not async-signal safe).");
#if defined(OS_LINUX)
auto tasks_stats_provider = TasksStatsCounters::findBestAvailableProvider();
if (tasks_stats_provider == TasksStatsCounters::MetricsProvider::None)
{
2020-05-23 22:24:01 +00:00
LOG_INFO(log, "It looks like this system does not have procfs mounted at /proc location,"
2020-05-23 22:21:29 +00:00
" neither clickhouse-server process has CAP_NET_ADMIN capability."
" 'taskstats' performance statistics will be disabled."
" It could happen due to incorrect ClickHouse package installation."
" You can try to resolve the problem manually with 'sudo setcap cap_net_admin=+ep {}'."
" Note that it will not work on 'nosuid' mounted filesystems."
" It also doesn't work if you run clickhouse-server inside network namespace as it happens in some containers.",
executable_path);
2019-07-06 18:02:28 +00:00
}
else
{
LOG_INFO(log, "Tasks stats provider: {}", TasksStatsCounters::metricsProviderString(tasks_stats_provider));
}
2019-07-06 18:02:28 +00:00
if (!hasLinuxCapability(CAP_SYS_NICE))
{
2020-06-27 12:56:06 +00:00
LOG_INFO(log, "It looks like the process has no CAP_SYS_NICE capability, the setting 'os_thread_priority' will have no effect."
2020-05-23 22:21:29 +00:00
" It could happen due to incorrect ClickHouse package installation."
" You could resolve the problem manually with 'sudo setcap cap_sys_nice=+ep {}'."
" Note that it will not work on 'nosuid' mounted filesystems.",
executable_path);
}
#else
2020-05-23 22:24:01 +00:00
LOG_INFO(log, "TaskStats is not implemented for this OS. IO accounting will be disabled.");
#endif
2013-09-14 05:14:22 +00:00
{
attachSystemTablesAsync(global_context, *DatabaseCatalog::instance().getSystemDatabase(), async_metrics);
2019-11-26 20:27:24 +00:00
{
std::lock_guard lock(servers_lock);
createServers(config(), listen_hosts, interserver_listen_hosts, listen_try, server_pool, async_metrics, servers);
if (servers.empty())
throw Exception(ErrorCodes::NO_ELEMENTS_IN_CONFIG,
"No servers started (add valid listen_host and 'tcp_port' or 'http_port' "
"to configuration file.)");
2014-03-21 13:42:14 +00:00
}
2022-01-28 17:37:52 +00:00
if (servers.empty())
throw Exception(ErrorCodes::NO_ELEMENTS_IN_CONFIG,
"No servers started (add valid listen_host and 'tcp_port' or 'http_port' "
"to configuration file.)");
2021-07-10 02:59:11 +00:00
#if USE_SSL
CertificateReloader::instance().tryLoad(config());
2021-07-10 02:59:11 +00:00
#endif
2020-12-17 13:47:03 +00:00
/// Must be done after initialization of `servers`, because async_metrics will access `servers` variable from its thread.
async_metrics.start();
{
String level_str = config().getString("text_log.level", "");
int level = level_str.empty() ? INT_MAX : Poco::Logger::parseLevel(level_str);
setTextLog(global_context->getTextLog(), level);
}
buildLoggers(config(), logger());
main_config_reloader->start();
access_control.startPeriodicReloading();
/// try to load dictionaries immediately, throw on error and die
try
{
global_context->loadOrReloadDictionaries(config());
}
catch (...)
{
2021-10-01 13:44:09 +00:00
tryLogCurrentException(log, "Caught exception while loading dictionaries.");
throw;
}
/// try to load embedded dictionaries immediately, throw on error and die
try
{
global_context->tryCreateEmbeddedDictionaries(config());
}
catch (...)
{
2021-10-01 13:44:09 +00:00
tryLogCurrentException(log, "Caught exception while loading embedded dictionaries.");
throw;
}
2021-09-15 20:52:18 +00:00
/// try to load user defined executable functions, throw on error and die
try
{
global_context->loadOrReloadUserDefinedExecutableFunctions(config());
}
catch (...)
{
2021-10-01 13:44:09 +00:00
tryLogCurrentException(log, "Caught exception while loading user defined executable functions.");
throw;
}
if (has_zookeeper && config().has("distributed_ddl"))
{
/// DDL worker should be started after all tables were loaded
String ddl_zookeeper_path = config().getString("distributed_ddl.path", "/clickhouse/task_queue/ddl/");
int pool_size = config().getInt("distributed_ddl.pool_size", 1);
if (pool_size < 1)
throw Exception(ErrorCodes::ARGUMENT_OUT_OF_BOUND, "distributed_ddl.pool_size should be greater then 0");
global_context->setDDLWorker(std::make_unique<DDLWorker>(pool_size, ddl_zookeeper_path, global_context, &config(),
2021-08-11 03:40:06 +00:00
"distributed_ddl", "DDLWorker",
&CurrentMetrics::MaxDDLEntryID, &CurrentMetrics::MaxPushedDDLEntryID));
}
{
std::lock_guard lock(servers_lock);
for (auto & server : servers)
{
server.start();
LOG_INFO(log, "Listening for {}", server.getDescription());
}
2022-03-14 20:43:34 +00:00
global_context->setServerCompletelyStarted();
LOG_INFO(log, "Ready for connections.");
}
startup_watch.stop();
ProfileEvents::increment(ProfileEvents::ServerStartupMilliseconds, startup_watch.elapsedMilliseconds());
try
{
global_context->startClusterDiscovery();
}
catch (...)
{
tryLogCurrentException(log, "Caught exception while starting cluster discovery");
}
#if defined(OS_LINUX)
/// Tell the service manager that service startup is finished.
/// NOTE: the parent clickhouse-watchdog process must do systemdNotify("MAINPID={}\n", child_pid); before
/// the child process notifies 'READY=1'.
2022-11-21 14:59:50 +00:00
systemdNotify("READY=1\n");
#endif
Fix uncaught exception during server termination Example of a stacktrace: <details> ``` [ 47463 ] {} <Trace> BaseDaemon: Received signal 15 [ 47463 ] {} <Information> Application: Received termination signal (Terminated) [ 47462 ] {} <Debug> Application: Received termination signal. [ 47462 ] {} <Debug> Application: Waiting for current connections to close. [ 47463 ] {} <Trace> BaseDaemon: Received signal 15 [ 47463 ] {} <Information> Application: Received termination signal (Terminated) ... [ 47463 ] {} <Trace> BaseDaemon: Received signal -1 [ 47463 ] {} <Fatal> BaseDaemon: (version 21.9.1.1, build id: 63945F58FC2C28ED) (from thread 47462) Terminate called for uncaught exception: [ 47463 ] {} <Fatal> BaseDaemon: Code: 210. DB::NetException: Connection reset by peer, while writing to socket (10.7.141.42:9000). (NETWORK_ERROR), Stack trace (when copying this message, always include the lines below): [ 47463 ] {} <Fatal> BaseDaemon: [ 47463 ] {} <Fatal> BaseDaemon: 0. DB::Exception::Exception(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, int, bool) @ 0x94ca99a in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug [ 47463 ] {} <Fatal> BaseDaemon: 1. DB::WriteBufferFromPocoSocket::nextImpl() @ 0x10676a3b in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug [ 47463 ] {} <Fatal> BaseDaemon: 2. DB::Connection::sendCancel() @ 0x11554701 in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug [ 47463 ] {} <Fatal> BaseDaemon: 3. DB::MultiplexedConnections::sendCancel() @ 0x1157e766 in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug [ 47463 ] {} <Fatal> BaseDaemon: 4. DB::RemoteQueryExecutor::tryCancel(char const*, std::__1::unique_ptr<DB::RemoteQueryExecutorReadContext, std::__1::default_delete<DB::RemoteQueryExecutorReadContext> >*) @ 0x10392000 in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug [ 47463 ] {} <Fatal> BaseDaemon: 5. DB::PipelineExecutor::cancel() @ 0x11697ffe in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug [ 47463 ] {} <Fatal> BaseDaemon: 6. DB::QueryStatus::cancelQuery(bool) @ 0x10c19fc8 in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug [ 47463 ] {} <Fatal> BaseDaemon: 7. DB::ProcessList::killAllQueries() @ 0x10c1a6ae in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug [ 47463 ] {} <Fatal> BaseDaemon: 8. basic_scope_guard<DB::Server::main(std::__1::vector<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::allocator<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > > > const&)::$_18>::~basic_scope_guard() @ 0x95587ad in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug [ 47463 ] {} <Fatal> BaseDaemon: 9. DB::Server::main(std::__1::vector<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::allocator<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > > > const&) @ 0x95528a2 in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug [ 47463 ] {} <Fatal> BaseDaemon: 10. Poco::Util::Application::run() @ 0x141e85a3 in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug [ 47463 ] {} <Fatal> BaseDaemon: 11. DB::Server::run() @ 0x9541dac in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug [ 47463 ] {} <Fatal> BaseDaemon: 12. mainEntryClickHouseServer(int, char**) @ 0x9540153 in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug [ 47463 ] {} <Fatal> BaseDaemon: 13. main @ 0x94c569e in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug [ 47463 ] {} <Fatal> BaseDaemon: 14. __libc_start_main @ 0x26d0a in /usr/lib/x86_64-linux-gnu/libc-2.31.so [ 47463 ] {} <Fatal> BaseDaemon: 15. _start @ 0x9490a2a in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug [ 47463 ] {} <Fatal> BaseDaemon: (version 21.9.1.1) [ 47463 ] {} <Trace> BaseDaemon: Received signal 6 [ 11858 ] {} <Fatal> BaseDaemon: ######################################## [ 11858 ] {} <Fatal> BaseDaemon: (version 21.9.1.1, build id: 63945F58FC2C28ED) (from thread 47462) (no query) Received signal Aborted (6) [ 11858 ] {} <Fatal> BaseDaemon: [ 11858 ] {} <Fatal> BaseDaemon: Stack trace: 0x7ff04c196ce1 0x7ff04c180537 0xff91f28 0x163304e3 0x1633044c 0x94c60cb 0x10c1a135 0x10c1a6ae 0x95587ad 0x95528a2 0x141e85a3 0x9541dac 0x9540153 0x94c569e 0x7ff04c181d0a 0x9490a2a [ 11858 ] {} <Fatal> BaseDaemon: 1. raise @ 0x3bce1 in /usr/lib/x86_64-linux-gnu/libc-2.31.so [ 11858 ] {} <Fatal> BaseDaemon: 2. abort @ 0x25537 in /usr/lib/x86_64-linux-gnu/libc-2.31.so [ 11858 ] {} <Fatal> BaseDaemon: 3. terminate_handler() @ 0xff91f28 in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug [ 11858 ] {} <Fatal> BaseDaemon: 4. std::__terminate(void (*)()) @ 0x163304e3 in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug [ 11858 ] {} <Fatal> BaseDaemon: 5. std::terminate() @ 0x1633044c in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug [ 11858 ] {} <Fatal> BaseDaemon: 6. ? @ 0x94c60cb in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug [ 11858 ] {} <Fatal> BaseDaemon: 7. ? @ 0x10c1a135 in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug [ 11858 ] {} <Fatal> BaseDaemon: 8. DB::ProcessList::killAllQueries() @ 0x10c1a6ae in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug [ 11858 ] {} <Fatal> BaseDaemon: 9. basic_scope_guard<DB::Server::main(std::__1::vector<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::allocator<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > > > const&)::$_18>::~basic_scope_guard() @ 0x95587ad in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug [ 11858 ] {} <Fatal> BaseDaemon: 10. DB::Server::main(std::__1::vector<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::allocator<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > > > const&) @ 0x95528a2 in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug [ 11858 ] {} <Fatal> BaseDaemon: 11. Poco::Util::Application::run() @ 0x141e85a3 in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug [ 11858 ] {} <Fatal> BaseDaemon: 12. DB::Server::run() @ 0x9541dac in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug [ 11858 ] {} <Fatal> BaseDaemon: 13. mainEntryClickHouseServer(int, char**) @ 0x9540153 in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug [ 11858 ] {} <Fatal> BaseDaemon: 14. main @ 0x94c569e in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug [ 11858 ] {} <Fatal> BaseDaemon: 15. __libc_start_main @ 0x26d0a in /usr/lib/x86_64-linux-gnu/libc-2.31.so [ 11858 ] {} <Fatal> BaseDaemon: 16. _start @ 0x9490a2a in /usr/lib/debug/.build-id/63/945f58fc2c28ed.debug ``` </details>
2021-09-08 18:10:34 +00:00
SCOPE_EXIT_SAFE({
2020-05-23 22:24:01 +00:00
LOG_DEBUG(log, "Received termination signal.");
/// Stop reloading of the main config. This must be done before everything else because it
/// can try to access/modify already deleted objects.
/// E.g. it can recreate new servers or it may pass a changed config to some destroyed parts of ContextSharedPart.
main_config_reloader.reset();
access_control.stopPeriodicReloading();
2017-03-03 21:10:41 +00:00
is_cancelled = true;
LOG_DEBUG(log, "Waiting for current connections to close.");
size_t current_connections = 0;
2017-03-28 18:38:12 +00:00
{
std::lock_guard lock(servers_lock);
for (auto & server : servers)
{
server.stop();
current_connections += server.currentConnections();
}
2017-03-28 18:38:12 +00:00
}
2020-05-23 22:21:29 +00:00
if (current_connections)
LOG_WARNING(log, "Closed all listening sockets. Waiting for {} outstanding connections.", current_connections);
2020-05-23 22:21:29 +00:00
else
2020-05-23 22:24:01 +00:00
LOG_INFO(log, "Closed all listening sockets.");
2019-07-30 23:12:04 +00:00
/// Killing remaining queries.
if (!server_settings.shutdown_wait_unfinished_queries)
global_context->getProcessList().killAllQueries();
2019-07-30 23:12:04 +00:00
2017-03-28 18:38:12 +00:00
if (current_connections)
current_connections = waitServersToFinish(servers, config().getInt("shutdown_wait_unfinished", 5));
2020-05-23 22:21:29 +00:00
if (current_connections)
LOG_WARNING(log, "Closed connections. But {} remain."
2020-05-23 22:21:29 +00:00
" Tip: To increase wait time add to config: <shutdown_wait_unfinished>60</shutdown_wait_unfinished>", current_connections);
else
2020-05-23 22:24:01 +00:00
LOG_INFO(log, "Closed connections.");
dns_cache_updater.reset();
2019-07-30 23:12:04 +00:00
if (current_connections)
{
/// There is no better way to force connections to close in Poco.
/// Otherwise connection handlers will continue to live
/// (they are effectively dangling objects, but they use global thread pool
/// and global thread pool destructor will wait for threads, preventing server shutdown).
/// Dump coverage here, because std::atexit callback would not be called.
dumpCoverageReportIfPossible();
LOG_WARNING(log, "Will shutdown forcefully.");
safeExit(0);
2019-07-30 23:12:04 +00:00
}
2017-03-03 21:10:41 +00:00
});
std::vector<std::unique_ptr<MetricsTransmitter>> metrics_transmitters;
for (const auto & graphite_key : DB::getMultipleKeysFromConfig(config(), "", "graphite"))
{
metrics_transmitters.emplace_back(std::make_unique<MetricsTransmitter>(
2019-07-04 19:08:37 +00:00
global_context->getConfigRef(), graphite_key, async_metrics));
}
waitForTerminationRequest();
}
2012-03-09 03:06:09 +00:00
return Application::EXIT_OK;
}
catch (...)
{
/// Poco does not provide stacktrace.
tryLogCurrentException("Application");
throw;
}
std::unique_ptr<TCPProtocolStackFactory> Server::buildProtocolStackFromConfig(
const Poco::Util::AbstractConfiguration & config,
const std::string & protocol,
Poco::Net::HTTPServerParams::Ptr http_params,
AsynchronousMetrics & async_metrics,
bool & is_secure)
{
auto create_factory = [&](const std::string & type, const std::string & conf_name) -> TCPServerConnectionFactory::Ptr
2022-09-08 06:12:33 +00:00
{
if (type == "tcp")
return TCPServerConnectionFactory::Ptr(new TCPHandlerFactory(*this, false, false));
2022-09-11 21:35:38 +00:00
2022-09-08 06:12:33 +00:00
if (type == "tls")
2022-09-11 22:44:27 +00:00
#if USE_SSL
2022-09-11 00:40:40 +00:00
return TCPServerConnectionFactory::Ptr(new TLSHandlerFactory(*this, conf_name));
2022-09-11 21:35:38 +00:00
#else
throw Exception(ErrorCodes::SUPPORT_IS_DISABLED, "SSL support for TCP protocol is disabled because Poco library was built without NetSSL support.");
2022-09-11 21:35:38 +00:00
#endif
if (type == "proxy1")
2022-09-11 00:40:40 +00:00
return TCPServerConnectionFactory::Ptr(new ProxyV1HandlerFactory(*this, conf_name));
2022-09-08 06:12:33 +00:00
if (type == "mysql")
return TCPServerConnectionFactory::Ptr(new MySQLHandlerFactory(*this));
if (type == "postgres")
return TCPServerConnectionFactory::Ptr(new PostgreSQLHandlerFactory(*this));
if (type == "http")
return TCPServerConnectionFactory::Ptr(
2022-11-09 08:02:04 +00:00
new HTTPServerConnectionFactory(httpContext(), http_params, createHandlerFactory(*this, config, async_metrics, "HTTPHandler-factory"))
2022-09-08 06:12:33 +00:00
);
if (type == "prometheus")
return TCPServerConnectionFactory::Ptr(
2022-11-09 08:02:04 +00:00
new HTTPServerConnectionFactory(httpContext(), http_params, createHandlerFactory(*this, config, async_metrics, "PrometheusHandler-factory"))
2022-09-08 06:12:33 +00:00
);
if (type == "interserver")
return TCPServerConnectionFactory::Ptr(
2022-11-09 08:02:04 +00:00
new HTTPServerConnectionFactory(httpContext(), http_params, createHandlerFactory(*this, config, async_metrics, "InterserverIOHTTPHandler-factory"))
2022-09-08 06:12:33 +00:00
);
2022-09-11 04:40:58 +00:00
throw Exception(ErrorCodes::INVALID_CONFIG_PARAMETER, "Protocol configuration error, unknown protocol name '{}'", type);
2022-09-08 06:12:33 +00:00
};
std::string conf_name = "protocols." + protocol;
std::string prefix = conf_name + ".";
std::unordered_set<std::string> pset {conf_name};
auto stack = std::make_unique<TCPProtocolStackFactory>(*this, conf_name);
while (true)
{
// if there is no "type" - it's a reference to another protocol and this is just an endpoint
if (config.has(prefix + "type"))
{
std::string type = config.getString(prefix + "type");
if (type == "tls")
{
if (is_secure)
throw Exception(ErrorCodes::INVALID_CONFIG_PARAMETER, "Protocol '{}' contains more than one TLS layer", protocol);
is_secure = true;
}
stack->append(create_factory(type, conf_name));
}
if (!config.has(prefix + "impl"))
break;
conf_name = "protocols." + config.getString(prefix + "impl");
prefix = conf_name + ".";
if (!pset.insert(conf_name).second)
throw Exception(ErrorCodes::INVALID_CONFIG_PARAMETER, "Protocol '{}' configuration contains a loop on '{}'", protocol, conf_name);
}
return stack;
}
2022-04-15 20:54:08 +00:00
2022-11-09 08:02:04 +00:00
HTTPContextPtr Server::httpContext() const
{
2022-11-09 14:51:41 +00:00
return std::make_shared<HTTPContext>(context());
2022-11-09 08:02:04 +00:00
}
void Server::createServers(
Poco::Util::AbstractConfiguration & config,
2022-04-05 15:04:53 +00:00
const Strings & listen_hosts,
const Strings & interserver_listen_hosts,
bool listen_try,
Poco::ThreadPool & server_pool,
AsynchronousMetrics & async_metrics,
std::vector<ProtocolServerAdapter> & servers,
bool start_servers)
{
const Settings & settings = global_context->getSettingsRef();
Poco::Timespan keep_alive_timeout(config.getUInt("keep_alive_timeout", 10), 0);
Poco::Net::HTTPServerParams::Ptr http_params = new Poco::Net::HTTPServerParams;
http_params->setTimeout(settings.http_receive_timeout);
http_params->setKeepAliveTimeout(keep_alive_timeout);
Poco::Util::AbstractConfiguration::Keys protocols;
config.keys("protocols", protocols);
2022-09-08 06:12:33 +00:00
for (const auto & protocol : protocols)
{
std::vector<std::string> hosts;
if (config.has("protocols." + protocol + ".host"))
hosts.push_back(config.getString("protocols." + protocol + ".host"));
else
hosts = listen_hosts;
2022-09-14 16:29:26 +00:00
for (const auto & host : hosts)
2022-09-08 06:12:33 +00:00
{
std::string conf_name = "protocols." + protocol;
std::string prefix = conf_name + ".";
2022-09-29 00:13:40 +00:00
if (!config.has(prefix + "port"))
continue;
std::string description {"<undefined> protocol"};
if (config.has(prefix + "description"))
description = config.getString(prefix + "description");
std::string port_name = prefix + "port";
bool is_secure = false;
auto stack = buildProtocolStackFromConfig(config, protocol, http_params, async_metrics, is_secure);
2022-09-29 00:13:40 +00:00
if (stack->empty())
throw Exception(ErrorCodes::INVALID_CONFIG_PARAMETER, "Protocol '{}' stack empty", protocol);
createServer(config, host, port_name.c_str(), listen_try, start_servers, servers, [&](UInt16 port) -> ProtocolServerAdapter
{
Poco::Net::ServerSocket socket;
auto address = socketBindListen(config, socket, host, port, is_secure);
socket.setReceiveTimeout(settings.receive_timeout);
socket.setSendTimeout(settings.send_timeout);
return ProtocolServerAdapter(
host,
port_name.c_str(),
description + ": " + address.toString(),
std::make_unique<TCPServer>(
stack.release(),
server_pool,
socket,
new Poco::Net::TCPServerParams));
});
2022-09-08 06:12:33 +00:00
}
}
2022-09-11 22:44:27 +00:00
for (const auto & listen_host : listen_hosts)
{
/// HTTP
const char * port_name = "http_port";
createServer(config, listen_host, port_name, listen_try, start_servers, servers, [&](UInt16 port) -> ProtocolServerAdapter
{
Poco::Net::ServerSocket socket;
auto address = socketBindListen(config, socket, listen_host, port);
socket.setReceiveTimeout(settings.http_receive_timeout);
socket.setSendTimeout(settings.http_send_timeout);
return ProtocolServerAdapter(
listen_host,
port_name,
"http://" + address.toString(),
std::make_unique<HTTPServer>(
2022-11-09 08:02:04 +00:00
httpContext(), createHandlerFactory(*this, config, async_metrics, "HTTPHandler-factory"), server_pool, socket, http_params));
});
/// HTTPS
port_name = "https_port";
createServer(config, listen_host, port_name, listen_try, start_servers, servers, [&](UInt16 port) -> ProtocolServerAdapter
{
#if USE_SSL
Poco::Net::SecureServerSocket socket;
auto address = socketBindListen(config, socket, listen_host, port, /* secure = */ true);
socket.setReceiveTimeout(settings.http_receive_timeout);
socket.setSendTimeout(settings.http_send_timeout);
return ProtocolServerAdapter(
listen_host,
port_name,
"https://" + address.toString(),
std::make_unique<HTTPServer>(
2022-11-09 08:02:04 +00:00
httpContext(), createHandlerFactory(*this, config, async_metrics, "HTTPSHandler-factory"), server_pool, socket, http_params));
#else
UNUSED(port);
throw Exception(ErrorCodes::SUPPORT_IS_DISABLED, "HTTPS protocol is disabled because Poco library was built without NetSSL support.");
#endif
});
/// TCP
port_name = "tcp_port";
createServer(config, listen_host, port_name, listen_try, start_servers, servers, [&](UInt16 port) -> ProtocolServerAdapter
{
Poco::Net::ServerSocket socket;
auto address = socketBindListen(config, socket, listen_host, port);
socket.setReceiveTimeout(settings.receive_timeout);
socket.setSendTimeout(settings.send_timeout);
return ProtocolServerAdapter(
listen_host,
port_name,
"native protocol (tcp): " + address.toString(),
std::make_unique<TCPServer>(
new TCPHandlerFactory(*this, /* secure */ false, /* proxy protocol */ false),
server_pool,
socket,
new Poco::Net::TCPServerParams));
});
/// TCP with PROXY protocol, see https://github.com/wolfeidau/proxyv2/blob/master/docs/proxy-protocol.txt
port_name = "tcp_with_proxy_port";
createServer(config, listen_host, port_name, listen_try, start_servers, servers, [&](UInt16 port) -> ProtocolServerAdapter
{
Poco::Net::ServerSocket socket;
auto address = socketBindListen(config, socket, listen_host, port);
socket.setReceiveTimeout(settings.receive_timeout);
socket.setSendTimeout(settings.send_timeout);
return ProtocolServerAdapter(
listen_host,
port_name,
"native protocol (tcp) with PROXY: " + address.toString(),
std::make_unique<TCPServer>(
new TCPHandlerFactory(*this, /* secure */ false, /* proxy protocol */ true),
server_pool,
socket,
new Poco::Net::TCPServerParams));
});
/// TCP with SSL
port_name = "tcp_port_secure";
createServer(config, listen_host, port_name, listen_try, start_servers, servers, [&](UInt16 port) -> ProtocolServerAdapter
{
#if USE_SSL
Poco::Net::SecureServerSocket socket;
auto address = socketBindListen(config, socket, listen_host, port, /* secure = */ true);
socket.setReceiveTimeout(settings.receive_timeout);
socket.setSendTimeout(settings.send_timeout);
return ProtocolServerAdapter(
listen_host,
port_name,
"secure native protocol (tcp_secure): " + address.toString(),
std::make_unique<TCPServer>(
new TCPHandlerFactory(*this, /* secure */ true, /* proxy protocol */ false),
server_pool,
socket,
new Poco::Net::TCPServerParams));
#else
UNUSED(port);
throw Exception(ErrorCodes::SUPPORT_IS_DISABLED, "SSL support for TCP protocol is disabled because Poco library was built without NetSSL support.");
#endif
});
port_name = "mysql_port";
createServer(config, listen_host, port_name, listen_try, start_servers, servers, [&](UInt16 port) -> ProtocolServerAdapter
{
Poco::Net::ServerSocket socket;
auto address = socketBindListen(config, socket, listen_host, port, /* secure = */ true);
socket.setReceiveTimeout(Poco::Timespan());
socket.setSendTimeout(settings.send_timeout);
return ProtocolServerAdapter(
listen_host,
port_name,
"MySQL compatibility protocol: " + address.toString(),
std::make_unique<TCPServer>(new MySQLHandlerFactory(*this), server_pool, socket, new Poco::Net::TCPServerParams));
});
port_name = "postgresql_port";
createServer(config, listen_host, port_name, listen_try, start_servers, servers, [&](UInt16 port) -> ProtocolServerAdapter
{
Poco::Net::ServerSocket socket;
auto address = socketBindListen(config, socket, listen_host, port, /* secure = */ true);
socket.setReceiveTimeout(Poco::Timespan());
socket.setSendTimeout(settings.send_timeout);
return ProtocolServerAdapter(
listen_host,
port_name,
"PostgreSQL compatibility protocol: " + address.toString(),
std::make_unique<TCPServer>(new PostgreSQLHandlerFactory(*this), server_pool, socket, new Poco::Net::TCPServerParams));
});
#if USE_GRPC
port_name = "grpc_port";
createServer(config, listen_host, port_name, listen_try, start_servers, servers, [&](UInt16 port) -> ProtocolServerAdapter
{
Poco::Net::SocketAddress server_address(listen_host, port);
return ProtocolServerAdapter(
listen_host,
port_name,
"gRPC protocol: " + server_address.toString(),
std::make_unique<GRPCServer>(*this, makeSocketAddress(listen_host, port, &logger())));
});
#endif
/// Prometheus (if defined and not setup yet with http_port)
port_name = "prometheus.port";
createServer(config, listen_host, port_name, listen_try, start_servers, servers, [&](UInt16 port) -> ProtocolServerAdapter
{
Poco::Net::ServerSocket socket;
auto address = socketBindListen(config, socket, listen_host, port);
socket.setReceiveTimeout(settings.http_receive_timeout);
socket.setSendTimeout(settings.http_send_timeout);
return ProtocolServerAdapter(
listen_host,
port_name,
"Prometheus: http://" + address.toString(),
std::make_unique<HTTPServer>(
2022-11-09 08:02:04 +00:00
httpContext(), createHandlerFactory(*this, config, async_metrics, "PrometheusHandler-factory"), server_pool, socket, http_params));
});
}
2022-04-05 15:04:53 +00:00
/// Now iterate over interserver_listen_hosts
for (const auto & interserver_listen_host : interserver_listen_hosts)
{
/// Interserver IO HTTP
const char * port_name = "interserver_http_port";
createServer(config, interserver_listen_host, port_name, listen_try, start_servers, servers, [&](UInt16 port) -> ProtocolServerAdapter
{
Poco::Net::ServerSocket socket;
auto address = socketBindListen(config, socket, interserver_listen_host, port);
socket.setReceiveTimeout(settings.http_receive_timeout);
socket.setSendTimeout(settings.http_send_timeout);
return ProtocolServerAdapter(
interserver_listen_host,
port_name,
"replica communication (interserver): http://" + address.toString(),
std::make_unique<HTTPServer>(
2022-11-09 08:02:04 +00:00
httpContext(),
createHandlerFactory(*this, config, async_metrics, "InterserverIOHTTPHandler-factory"),
server_pool,
socket,
http_params));
});
port_name = "interserver_https_port";
createServer(config, interserver_listen_host, port_name, listen_try, start_servers, servers, [&](UInt16 port) -> ProtocolServerAdapter
{
#if USE_SSL
Poco::Net::SecureServerSocket socket;
auto address = socketBindListen(config, socket, interserver_listen_host, port, /* secure = */ true);
socket.setReceiveTimeout(settings.http_receive_timeout);
socket.setSendTimeout(settings.http_send_timeout);
return ProtocolServerAdapter(
interserver_listen_host,
port_name,
"secure replica communication (interserver): https://" + address.toString(),
std::make_unique<HTTPServer>(
2022-11-09 08:02:04 +00:00
httpContext(),
createHandlerFactory(*this, config, async_metrics, "InterserverIOHTTPSHandler-factory"),
server_pool,
socket,
http_params));
#else
UNUSED(port);
throw Exception(ErrorCodes::SUPPORT_IS_DISABLED, "SSL support for TCP protocol is disabled because Poco library was built without NetSSL support.");
#endif
});
}
}
void Server::updateServers(
Poco::Util::AbstractConfiguration & config,
Poco::ThreadPool & server_pool,
AsynchronousMetrics & async_metrics,
std::vector<ProtocolServerAdapter> & servers)
{
Poco::Logger * log = &logger();
Fix server reload on port change (do not wait for current connections) If you will execute 'SYSTEM RELOAD CONFIG' via, i.e., TCP protocol, then reload on port change will endlessly wait for connection from which this query had been issued, and you will see the following message in the logs: 2022.04.28 03:34:57.552513 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Debug> executeQuery: (from 127.0.0.1:11774) system reload config ... 2022.04.28 03:34:57.710640 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Information> Application: Stopped listening for http://127.0.0.1:18123 2022.04.28 03:34:57.798774 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Information> Application: Stopped listening for native protocol (tcp): 127.0.0.1:19000 ... 2022.04.28 03:34:57.901375 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Debug> Application: Server finished: http://127.0.0.1:18123 2022.04.28 03:34:57.901455 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 2022.04.28 03:34:58.001717 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 2022.04.28 03:34:58.101881 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 ... 2022.04.28 03:35:01.707951 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 But waiting for the current connection will never ends. So instead of waiting directly from the query context (SYSTEM RELOAD CONFIG) do this in background (actually not even in background, but check on server reload and on exit). v0: just don't wait for the servers v2: fix use-after-free by removing dependency from server in handlers v3: wait servers in background to avoid use-after-free of the context Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
2022-04-27 10:24:49 +00:00
const auto listen_hosts = getListenHosts(config);
2022-04-05 15:04:53 +00:00
const auto interserver_listen_hosts = getInterserverListenHosts(config);
const auto listen_try = getListenTry(config);
/// Remove servers once all their connections are closed
auto check_server = [&log](const char prefix[], auto & server)
Fix server reload on port change (do not wait for current connections) If you will execute 'SYSTEM RELOAD CONFIG' via, i.e., TCP protocol, then reload on port change will endlessly wait for connection from which this query had been issued, and you will see the following message in the logs: 2022.04.28 03:34:57.552513 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Debug> executeQuery: (from 127.0.0.1:11774) system reload config ... 2022.04.28 03:34:57.710640 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Information> Application: Stopped listening for http://127.0.0.1:18123 2022.04.28 03:34:57.798774 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Information> Application: Stopped listening for native protocol (tcp): 127.0.0.1:19000 ... 2022.04.28 03:34:57.901375 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Debug> Application: Server finished: http://127.0.0.1:18123 2022.04.28 03:34:57.901455 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 2022.04.28 03:34:58.001717 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 2022.04.28 03:34:58.101881 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 ... 2022.04.28 03:35:01.707951 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 But waiting for the current connection will never ends. So instead of waiting directly from the query context (SYSTEM RELOAD CONFIG) do this in background (actually not even in background, but check on server reload and on exit). v0: just don't wait for the servers v2: fix use-after-free by removing dependency from server in handlers v3: wait servers in background to avoid use-after-free of the context Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
2022-04-27 10:24:49 +00:00
{
if (!server.isStopping())
return false;
size_t current_connections = server.currentConnections();
LOG_DEBUG(log, "Server {}{}: {} ({} connections)",
server.getDescription(),
prefix,
!current_connections ? "finished" : "waiting",
current_connections);
return !current_connections;
};
std::erase_if(servers, std::bind_front(check_server, " (from one of previous reload)"));
Fix server reload on port change (do not wait for current connections) If you will execute 'SYSTEM RELOAD CONFIG' via, i.e., TCP protocol, then reload on port change will endlessly wait for connection from which this query had been issued, and you will see the following message in the logs: 2022.04.28 03:34:57.552513 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Debug> executeQuery: (from 127.0.0.1:11774) system reload config ... 2022.04.28 03:34:57.710640 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Information> Application: Stopped listening for http://127.0.0.1:18123 2022.04.28 03:34:57.798774 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Information> Application: Stopped listening for native protocol (tcp): 127.0.0.1:19000 ... 2022.04.28 03:34:57.901375 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Debug> Application: Server finished: http://127.0.0.1:18123 2022.04.28 03:34:57.901455 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 2022.04.28 03:34:58.001717 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 2022.04.28 03:34:58.101881 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 ... 2022.04.28 03:35:01.707951 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 But waiting for the current connection will never ends. So instead of waiting directly from the query context (SYSTEM RELOAD CONFIG) do this in background (actually not even in background, but check on server reload and on exit). v0: just don't wait for the servers v2: fix use-after-free by removing dependency from server in handlers v3: wait servers in background to avoid use-after-free of the context Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
2022-04-27 10:24:49 +00:00
Poco::Util::AbstractConfiguration & previous_config = latest_config ? *latest_config : this->config();
for (auto & server : servers)
Fix server reload on port change (do not wait for current connections) If you will execute 'SYSTEM RELOAD CONFIG' via, i.e., TCP protocol, then reload on port change will endlessly wait for connection from which this query had been issued, and you will see the following message in the logs: 2022.04.28 03:34:57.552513 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Debug> executeQuery: (from 127.0.0.1:11774) system reload config ... 2022.04.28 03:34:57.710640 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Information> Application: Stopped listening for http://127.0.0.1:18123 2022.04.28 03:34:57.798774 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Information> Application: Stopped listening for native protocol (tcp): 127.0.0.1:19000 ... 2022.04.28 03:34:57.901375 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Debug> Application: Server finished: http://127.0.0.1:18123 2022.04.28 03:34:57.901455 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 2022.04.28 03:34:58.001717 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 2022.04.28 03:34:58.101881 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 ... 2022.04.28 03:35:01.707951 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 But waiting for the current connection will never ends. So instead of waiting directly from the query context (SYSTEM RELOAD CONFIG) do this in background (actually not even in background, but check on server reload and on exit). v0: just don't wait for the servers v2: fix use-after-free by removing dependency from server in handlers v3: wait servers in background to avoid use-after-free of the context Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
2022-04-27 10:24:49 +00:00
{
if (!server.isStopping())
{
std::string port_name = server.getPortName();
bool has_host = false;
bool is_http = false;
if (port_name.starts_with("protocols."))
{
std::string protocol = port_name.substr(0, port_name.find_last_of('.'));
has_host = config.has(protocol + ".host");
std::string conf_name = protocol;
std::string prefix = protocol + ".";
std::unordered_set<std::string> pset {conf_name};
while (true)
{
if (config.has(prefix + "type"))
{
std::string type = config.getString(prefix + "type");
if (type == "http")
{
is_http = true;
break;
}
}
if (!config.has(prefix + "impl"))
break;
conf_name = "protocols." + config.getString(prefix + "impl");
prefix = conf_name + ".";
if (!pset.insert(conf_name).second)
throw Exception(ErrorCodes::INVALID_CONFIG_PARAMETER, "Protocol '{}' configuration contains a loop on '{}'", protocol, conf_name);
}
}
else
{
/// NOTE: better to compare using getPortName() over using
/// dynamic_cast<> since HTTPServer is also used for prometheus and
/// internal replication communications.
is_http = server.getPortName() == "http_port" || server.getPortName() == "https_port";
}
if (!has_host)
has_host = std::find(listen_hosts.begin(), listen_hosts.end(), server.getListenHost()) != listen_hosts.end();
bool has_port = !config.getString(port_name, "").empty();
bool force_restart = is_http && !isSameConfiguration(previous_config, config, "http_handlers");
if (force_restart)
LOG_TRACE(log, "<http_handlers> had been changed, will reload {}", server.getDescription());
if (!has_host || !has_port || config.getInt(server.getPortName()) != server.portNumber() || force_restart)
{
server.stop();
LOG_INFO(log, "Stopped listening for {}", server.getDescription());
}
}
Fix server reload on port change (do not wait for current connections) If you will execute 'SYSTEM RELOAD CONFIG' via, i.e., TCP protocol, then reload on port change will endlessly wait for connection from which this query had been issued, and you will see the following message in the logs: 2022.04.28 03:34:57.552513 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Debug> executeQuery: (from 127.0.0.1:11774) system reload config ... 2022.04.28 03:34:57.710640 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Information> Application: Stopped listening for http://127.0.0.1:18123 2022.04.28 03:34:57.798774 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Information> Application: Stopped listening for native protocol (tcp): 127.0.0.1:19000 ... 2022.04.28 03:34:57.901375 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Debug> Application: Server finished: http://127.0.0.1:18123 2022.04.28 03:34:57.901455 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 2022.04.28 03:34:58.001717 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 2022.04.28 03:34:58.101881 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 ... 2022.04.28 03:35:01.707951 [ 37101 ] {b41d855c-4dbf-470a-a144-c6ae5a1abda8} <Trace> Application: Waiting server to finish: native protocol (tcp): 127.0.0.1:19000 But waiting for the current connection will never ends. So instead of waiting directly from the query context (SYSTEM RELOAD CONFIG) do this in background (actually not even in background, but check on server reload and on exit). v0: just don't wait for the servers v2: fix use-after-free by removing dependency from server in handlers v3: wait servers in background to avoid use-after-free of the context Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
2022-04-27 10:24:49 +00:00
}
createServers(config, listen_hosts, interserver_listen_hosts, listen_try, server_pool, async_metrics, servers, /* start_servers= */ true);
std::erase_if(servers, std::bind_front(check_server, ""));
}
2012-03-09 03:06:09 +00:00
}