ClickHouse/src/Server/HTTPHandler.cpp

1169 lines
45 KiB
C++
Raw Normal View History

#include <Server/HTTPHandler.h>
2019-02-15 11:46:07 +00:00
#include <Access/Authentication.h>
#include <Access/Credentials.h>
#include <Access/ExternalAuthenticators.h>
#include <Compression/CompressedReadBuffer.h>
#include <Compression/CompressedWriteBuffer.h>
#include <Core/ExternalTable.h>
#include <Disks/StoragePolicy.h>
#include <IO/CascadeWriteBuffer.h>
#include <IO/ConcatReadBuffer.h>
#include <IO/MemoryReadWriteBuffer.h>
#include <IO/ReadBufferFromIStream.h>
#include <IO/ReadBufferFromString.h>
#include <IO/WriteBufferFromFile.h>
#include <IO/WriteBufferFromString.h>
#include <IO/WriteBufferFromTemporaryFile.h>
#include <IO/WriteHelpers.h>
#include <IO/copyData.h>
#include <Interpreters/Context.h>
#include <Interpreters/QueryParameterVisitor.h>
#include <Interpreters/executeQuery.h>
2021-08-01 14:12:34 +00:00
#include <Interpreters/Session.h>
#include <Server/HTTPHandlerFactory.h>
#include <Server/HTTPHandlerRequestFilter.h>
#include <Server/IServer.h>
2021-10-05 14:43:33 +00:00
#include <base/logger_useful.h>
#include <Common/SettingsChanges.h>
#include <Common/StringUtils/StringUtils.h>
#include <Common/escapeForFileName.h>
#include <Common/setThreadName.h>
2017-07-13 20:58:19 +00:00
#include <Common/typeid_cast.h>
2021-10-04 07:55:09 +00:00
2021-10-05 14:43:33 +00:00
#include <base/getFQDNOrHostName.h>
#include <base/scope_guard.h>
#include <Server/HTTP/HTTPResponse.h>
#if !defined(ARCADIA_BUILD)
# include <Common/config.h>
#endif
#include <Poco/Base64Decoder.h>
#include <Poco/Base64Encoder.h>
#include <Poco/Net/HTTPBasicCredentials.h>
#include <Poco/Net/HTTPStream.h>
#include <Poco/Net/NetException.h>
#include <Poco/MemoryStream.h>
#include <Poco/StreamCopier.h>
#include <Poco/String.h>
#include <chrono>
#include <iomanip>
#include <sstream>
2012-03-09 03:06:09 +00:00
namespace DB
{
namespace ErrorCodes
{
2020-02-25 18:02:41 +00:00
extern const int LOGICAL_ERROR;
extern const int CANNOT_PARSE_TEXT;
extern const int CANNOT_PARSE_ESCAPE_SEQUENCE;
extern const int CANNOT_PARSE_QUOTED_STRING;
extern const int CANNOT_PARSE_DATE;
extern const int CANNOT_PARSE_DATETIME;
extern const int CANNOT_PARSE_NUMBER;
extern const int CANNOT_PARSE_INPUT_ASSERTION_FAILED;
extern const int CANNOT_OPEN_FILE;
2020-04-26 11:06:14 +00:00
extern const int CANNOT_COMPILE_REGEXP;
extern const int UNKNOWN_ELEMENT_IN_AST;
extern const int UNKNOWN_TYPE_OF_AST_NODE;
extern const int TOO_DEEP_AST;
extern const int TOO_BIG_AST;
extern const int UNEXPECTED_AST_STRUCTURE;
extern const int SYNTAX_ERROR;
extern const int INCORRECT_DATA;
extern const int TYPE_MISMATCH;
extern const int UNKNOWN_TABLE;
extern const int UNKNOWN_FUNCTION;
extern const int UNKNOWN_IDENTIFIER;
extern const int UNKNOWN_TYPE;
extern const int UNKNOWN_STORAGE;
extern const int UNKNOWN_DATABASE;
extern const int UNKNOWN_SETTING;
extern const int UNKNOWN_DIRECTION_OF_SORTING;
extern const int UNKNOWN_AGGREGATE_FUNCTION;
extern const int UNKNOWN_FORMAT;
extern const int UNKNOWN_DATABASE_ENGINE;
extern const int UNKNOWN_TYPE_OF_QUERY;
2020-04-04 08:57:16 +00:00
extern const int NO_ELEMENTS_IN_CONFIG;
extern const int QUERY_IS_TOO_LARGE;
extern const int NOT_IMPLEMENTED;
extern const int SOCKET_TIMEOUT;
extern const int UNKNOWN_USER;
extern const int WRONG_PASSWORD;
extern const int REQUIRED_PASSWORD;
extern const int AUTHENTICATION_FAILED;
2020-08-28 01:21:08 +00:00
extern const int BAD_REQUEST_PARAMETER;
extern const int INVALID_SESSION_TIMEOUT;
extern const int HTTP_LENGTH_REQUIRED;
}
2021-09-18 09:36:02 +00:00
namespace
{
bool tryAddHeadersFromConfig(HTTPServerResponse & response, const Poco::Util::LayeredConfiguration & config)
{
if (config.has("http_options_response"))
2021-09-18 09:36:02 +00:00
{
Strings config_keys;
config.keys("http_options_response", config_keys);
for (const std::string & config_key : config_keys)
2021-09-18 09:36:02 +00:00
{
if (config_key == "header" || config_key.starts_with("header["))
2021-09-18 09:36:02 +00:00
{
/// If there is empty header name, it will not be processed and message about it will be in logs
if (config.getString("http_options_response." + config_key + ".name", "").empty())
LOG_WARNING(&Poco::Logger::get("processOptionsRequest"), "Empty header was found in config. It will not be processed.");
2021-09-24 12:40:27 +00:00
else
response.add(config.getString("http_options_response." + config_key + ".name", ""),
config.getString("http_options_response." + config_key + ".value", ""));
2021-09-18 09:36:02 +00:00
}
}
return true;
}
return false;
}
/// Process options request. Useful for CORS.
void processOptionsRequest(HTTPServerResponse & response, const Poco::Util::LayeredConfiguration & config)
{
/// If can add some headers from config
if (tryAddHeadersFromConfig(response, config))
{
response.setKeepAlive(false);
response.setStatusAndReason(HTTPResponse::HTTP_NO_CONTENT);
response.send();
2021-09-18 09:36:02 +00:00
}
}
}
2021-09-18 09:36:02 +00:00
static String base64Decode(const String & encoded)
{
String decoded;
Poco::MemoryInputStream istr(encoded.data(), encoded.size());
Poco::Base64Decoder decoder(istr);
Poco::StreamCopier::copyToString(decoder, decoded);
return decoded;
}
static String base64Encode(const String & decoded)
{
std::ostringstream ostr; // STYLE_CHECK_ALLOW_STD_STRING_STREAM
ostr.exceptions(std::ios::failbit);
Poco::Base64Encoder encoder(ostr);
encoder.rdbuf()->setLineLength(0);
encoder << decoded;
encoder.close();
return ostr.str();
}
2017-02-07 05:21:31 +00:00
static Poco::Net::HTTPResponse::HTTPStatus exceptionCodeToHTTPStatus(int exception_code)
{
using namespace Poco::Net;
if (exception_code == ErrorCodes::REQUIRED_PASSWORD)
{
return HTTPResponse::HTTP_UNAUTHORIZED;
}
else if (exception_code == ErrorCodes::UNKNOWN_USER ||
exception_code == ErrorCodes::WRONG_PASSWORD ||
exception_code == ErrorCodes::AUTHENTICATION_FAILED)
{
return HTTPResponse::HTTP_FORBIDDEN;
}
else if (exception_code == ErrorCodes::CANNOT_PARSE_TEXT ||
exception_code == ErrorCodes::CANNOT_PARSE_ESCAPE_SEQUENCE ||
exception_code == ErrorCodes::CANNOT_PARSE_QUOTED_STRING ||
exception_code == ErrorCodes::CANNOT_PARSE_DATE ||
exception_code == ErrorCodes::CANNOT_PARSE_DATETIME ||
exception_code == ErrorCodes::CANNOT_PARSE_NUMBER ||
exception_code == ErrorCodes::CANNOT_PARSE_INPUT_ASSERTION_FAILED ||
exception_code == ErrorCodes::UNKNOWN_ELEMENT_IN_AST ||
exception_code == ErrorCodes::UNKNOWN_TYPE_OF_AST_NODE ||
exception_code == ErrorCodes::TOO_DEEP_AST ||
exception_code == ErrorCodes::TOO_BIG_AST ||
exception_code == ErrorCodes::UNEXPECTED_AST_STRUCTURE ||
exception_code == ErrorCodes::SYNTAX_ERROR ||
exception_code == ErrorCodes::INCORRECT_DATA ||
exception_code == ErrorCodes::TYPE_MISMATCH)
{
return HTTPResponse::HTTP_BAD_REQUEST;
}
else if (exception_code == ErrorCodes::UNKNOWN_TABLE ||
exception_code == ErrorCodes::UNKNOWN_FUNCTION ||
exception_code == ErrorCodes::UNKNOWN_IDENTIFIER ||
exception_code == ErrorCodes::UNKNOWN_TYPE ||
exception_code == ErrorCodes::UNKNOWN_STORAGE ||
exception_code == ErrorCodes::UNKNOWN_DATABASE ||
exception_code == ErrorCodes::UNKNOWN_SETTING ||
exception_code == ErrorCodes::UNKNOWN_DIRECTION_OF_SORTING ||
exception_code == ErrorCodes::UNKNOWN_AGGREGATE_FUNCTION ||
exception_code == ErrorCodes::UNKNOWN_FORMAT ||
exception_code == ErrorCodes::UNKNOWN_DATABASE_ENGINE ||
exception_code == ErrorCodes::UNKNOWN_TYPE_OF_QUERY)
{
return HTTPResponse::HTTP_NOT_FOUND;
}
else if (exception_code == ErrorCodes::QUERY_IS_TOO_LARGE)
{
return HTTPResponse::HTTP_REQUESTENTITYTOOLARGE;
}
else if (exception_code == ErrorCodes::NOT_IMPLEMENTED)
{
return HTTPResponse::HTTP_NOT_IMPLEMENTED;
}
else if (exception_code == ErrorCodes::SOCKET_TIMEOUT ||
exception_code == ErrorCodes::CANNOT_OPEN_FILE)
{
return HTTPResponse::HTTP_SERVICE_UNAVAILABLE;
}
else if (exception_code == ErrorCodes::HTTP_LENGTH_REQUIRED)
{
return HTTPResponse::HTTP_LENGTH_REQUIRED;
}
return HTTPResponse::HTTP_INTERNAL_SERVER_ERROR;
}
static std::chrono::steady_clock::duration parseSessionTimeout(
const Poco::Util::AbstractConfiguration & config,
const HTMLForm & params)
{
unsigned session_timeout = config.getInt("default_session_timeout", 60);
if (params.has("session_timeout"))
{
unsigned max_session_timeout = config.getUInt("max_session_timeout", 3600);
std::string session_timeout_str = params.get("session_timeout");
ReadBufferFromString buf(session_timeout_str);
if (!tryReadIntText(session_timeout, buf) || !buf.eof())
throw Exception("Invalid session timeout: '" + session_timeout_str + "'", ErrorCodes::INVALID_SESSION_TIMEOUT);
if (session_timeout > max_session_timeout)
throw Exception("Session timeout '" + session_timeout_str + "' is larger than max_session_timeout: " + toString(max_session_timeout)
+ ". Maximum session timeout could be modified in configuration file.",
ErrorCodes::INVALID_SESSION_TIMEOUT);
}
return std::chrono::seconds(session_timeout);
}
void HTTPHandler::pushDelayedResults(Output & used_output)
{
std::vector<WriteBufferPtr> write_buffers;
2021-03-04 11:10:21 +00:00
ConcatReadBuffer::Buffers read_buffers;
2020-05-18 08:08:55 +00:00
auto * cascade_buffer = typeid_cast<CascadeWriteBuffer *>(used_output.out_maybe_delayed_and_compressed.get());
if (!cascade_buffer)
throw Exception("Expected CascadeWriteBuffer", ErrorCodes::LOGICAL_ERROR);
cascade_buffer->getResultBuffers(write_buffers);
if (write_buffers.empty())
throw Exception("At least one buffer is expected to overwrite result into HTTP response", ErrorCodes::LOGICAL_ERROR);
for (auto & write_buf : write_buffers)
{
IReadableWriteBuffer * write_buf_concrete;
ReadBufferPtr reread_buf;
if (write_buf
&& (write_buf_concrete = dynamic_cast<IReadableWriteBuffer *>(write_buf.get()))
&& (reread_buf = write_buf_concrete->tryGetReadBuffer()))
{
2021-03-04 11:10:21 +00:00
read_buffers.emplace_back(wrapReadBufferPointer(reread_buf));
}
}
2021-03-04 11:10:21 +00:00
if (!read_buffers.empty())
{
2021-03-04 11:10:21 +00:00
ConcatReadBuffer concat_read_buffer(std::move(read_buffers));
copyData(concat_read_buffer, *used_output.out_maybe_compressed);
}
}
HTTPHandler::HTTPHandler(IServer & server_, const std::string & name)
: server(server_)
, log(&Poco::Logger::get(name))
2021-08-01 14:12:34 +00:00
, default_settings(server.context()->getSettingsRef())
{
server_display_name = server.config().getString("display_name", getFQDNOrHostName());
}
/// We need d-tor to be present in this translation unit to make it play well with some
/// forward decls in the header. Other than that, the default d-tor would be OK.
2021-08-01 14:12:34 +00:00
HTTPHandler::~HTTPHandler() = default;
bool HTTPHandler::authenticateUser(
HTTPServerRequest & request,
HTMLForm & params,
HTTPServerResponse & response)
2012-03-09 03:06:09 +00:00
{
using namespace Poco::Net;
/// The user and password can be passed by headers (similar to X-Auth-*),
/// which is used by load balancers to pass authentication information.
std::string user = request.get("X-ClickHouse-User", "");
std::string password = request.get("X-ClickHouse-Key", "");
std::string quota_key = request.get("X-ClickHouse-Quota", "");
std::string spnego_challenge;
if (user.empty() && password.empty() && quota_key.empty())
{
/// User name and password can be passed using query parameters
/// or using HTTP Basic auth (both methods are insecure).
if (request.hasCredentials())
{
/// It is prohibited to mix different authorization schemes.
if (params.has("user") || params.has("password"))
throw Exception("Invalid authentication: it is not allowed to use Authorization HTTP header and authentication via parameters simultaneously", ErrorCodes::AUTHENTICATION_FAILED);
std::string scheme;
std::string auth_info;
request.getCredentials(scheme, auth_info);
if (Poco::icompare(scheme, "Basic") == 0)
{
HTTPBasicCredentials credentials(auth_info);
user = credentials.getUsername();
password = credentials.getPassword();
}
else if (Poco::icompare(scheme, "Negotiate") == 0)
{
spnego_challenge = auth_info;
if (spnego_challenge.empty())
throw Exception("Invalid authentication: SPNEGO challenge is empty", ErrorCodes::AUTHENTICATION_FAILED);
}
else
{
throw Exception("Invalid authentication: '" + scheme + "' HTTP Authorization scheme is not supported", ErrorCodes::AUTHENTICATION_FAILED);
}
2017-09-23 23:18:48 +00:00
}
else
{
user = params.get("user", "default");
password = params.get("password", "");
}
quota_key = params.get("quota_key", "");
}
else
{
/// It is prohibited to mix different authorization schemes.
if (request.hasCredentials() || params.has("user") || params.has("password") || params.has("quota_key"))
throw Exception("Invalid authentication: it is not allowed to use X-ClickHouse HTTP headers and other authentication methods simultaneously", ErrorCodes::AUTHENTICATION_FAILED);
}
if (spnego_challenge.empty()) // I.e., now using user name and password strings ("Basic").
{
if (!request_credentials)
request_credentials = std::make_unique<BasicCredentials>();
auto * basic_credentials = dynamic_cast<BasicCredentials *>(request_credentials.get());
if (!basic_credentials)
throw Exception("Invalid authentication: unexpected 'Basic' HTTP Authorization scheme", ErrorCodes::AUTHENTICATION_FAILED);
basic_credentials->setUserName(user);
basic_credentials->setPassword(password);
}
else
{
if (!request_credentials)
2021-08-01 14:12:34 +00:00
request_credentials = server.context()->makeGSSAcceptorContext();
auto * gss_acceptor_context = dynamic_cast<GSSAcceptorContext *>(request_credentials.get());
if (!gss_acceptor_context)
throw Exception("Invalid authentication: unexpected 'Negotiate' HTTP Authorization scheme expected", ErrorCodes::AUTHENTICATION_FAILED);
#pragma GCC diagnostic push
#pragma GCC diagnostic ignored "-Wunreachable-code"
const auto spnego_response = base64Encode(gss_acceptor_context->processToken(base64Decode(spnego_challenge), log));
#pragma GCC diagnostic pop
if (!spnego_response.empty())
response.set("WWW-Authenticate", "Negotiate " + spnego_response);
if (!gss_acceptor_context->isFailed() && !gss_acceptor_context->isReady())
{
if (spnego_response.empty())
throw Exception("Invalid authentication: 'Negotiate' HTTP Authorization failure", ErrorCodes::AUTHENTICATION_FAILED);
response.setStatusAndReason(HTTPResponse::HTTP_UNAUTHORIZED);
response.send();
return false;
}
}
2020-12-02 19:20:47 +00:00
/// Set client info. It will be used for quota accounting parameters in 'setUser' method.
2021-08-01 14:12:34 +00:00
ClientInfo & client_info = session->getClientInfo();
2020-12-02 19:20:47 +00:00
ClientInfo::HTTPMethod http_method = ClientInfo::HTTPMethod::UNKNOWN;
if (request.getMethod() == HTTPServerRequest::HTTP_GET)
2020-12-02 19:20:47 +00:00
http_method = ClientInfo::HTTPMethod::GET;
else if (request.getMethod() == HTTPServerRequest::HTTP_POST)
2020-12-02 19:20:47 +00:00
http_method = ClientInfo::HTTPMethod::POST;
client_info.http_method = http_method;
client_info.http_user_agent = request.get("User-Agent", "");
2021-01-21 22:55:45 +00:00
client_info.http_referer = request.get("Referer", "");
2020-12-02 19:20:47 +00:00
client_info.forwarded_for = request.get("X-Forwarded-For", "");
2021-08-01 14:12:34 +00:00
client_info.quota_key = quota_key;
2020-12-02 19:20:47 +00:00
try
{
2021-08-01 14:12:34 +00:00
session->authenticate(*request_credentials, request.clientAddress());
}
catch (const Authentication::Require<BasicCredentials> & required_credentials)
{
request_credentials = std::make_unique<BasicCredentials>();
if (required_credentials.getRealm().empty())
response.set("WWW-Authenticate", "Basic");
else
response.set("WWW-Authenticate", "Basic realm=\"" + required_credentials.getRealm() + "\"");
response.setStatusAndReason(HTTPResponse::HTTP_UNAUTHORIZED);
response.send();
return false;
}
catch (const Authentication::Require<GSSAcceptorContext> & required_credentials)
{
2021-08-01 14:12:34 +00:00
request_credentials = server.context()->makeGSSAcceptorContext();
if (required_credentials.getRealm().empty())
response.set("WWW-Authenticate", "Negotiate");
else
response.set("WWW-Authenticate", "Negotiate realm=\"" + required_credentials.getRealm() + "\"");
response.setStatusAndReason(HTTPResponse::HTTP_UNAUTHORIZED);
response.send();
return false;
}
request_credentials.reset();
return true;
}
void HTTPHandler::processQuery(
HTTPServerRequest & request,
HTMLForm & params,
HTTPServerResponse & response,
Output & used_output,
std::optional<CurrentThread::QueryScope> & query_scope)
{
using namespace Poco::Net;
LOG_TRACE(log, "Request URI: {}", request.getURI());
if (!authenticateUser(request, params, response))
return; // '401 Unauthorized' response with 'Negotiate' has been sent at this point.
/// The user could specify session identifier and session timeout.
/// It allows to modify settings, create temporary tables and reuse them in subsequent requests.
String session_id;
std::chrono::steady_clock::duration session_timeout;
bool session_is_set = params.has("session_id");
const auto & config = server.config();
if (session_is_set)
{
session_id = params.get("session_id");
session_timeout = parseSessionTimeout(config, params);
std::string session_check = params.get("session_check", "");
2021-08-01 14:12:34 +00:00
session->makeSessionContext(session_id, session_timeout, session_check == "1");
}
2020-10-19 21:26:10 +00:00
// Parse the OpenTelemetry traceparent header.
// Disable in Arcadia -- it interferes with the
// test_clickhouse.TestTracing.test_tracing_via_http_proxy[traceparent] test.
2021-08-01 14:12:34 +00:00
ClientInfo client_info = session->getClientInfo();
#if !defined(ARCADIA_BUILD)
2020-08-28 01:21:08 +00:00
if (request.has("traceparent"))
{
std::string opentelemetry_traceparent = request.get("traceparent");
std::string error;
if (!client_info.client_trace_context.parseTraceparentHeader(
opentelemetry_traceparent, error))
2020-08-28 01:21:08 +00:00
{
throw Exception(ErrorCodes::BAD_REQUEST_PARAMETER,
2020-08-28 01:21:08 +00:00
"Failed to parse OpenTelemetry traceparent header '{}': {}",
opentelemetry_traceparent, error);
2020-08-28 01:21:08 +00:00
}
client_info.client_trace_context.tracestate = request.get("tracestate", "");
2020-08-28 01:21:08 +00:00
}
#endif
2020-08-28 01:21:08 +00:00
2021-08-01 14:12:34 +00:00
auto context = session->makeQueryContext(std::move(client_info));
2020-12-02 19:20:47 +00:00
/// The client can pass a HTTP header indicating supported compression method (gzip or deflate).
String http_response_compression_methods = request.get("Accept-Encoding", "");
2020-01-04 22:59:08 +00:00
CompressionMethod http_response_compression_method = CompressionMethod::None;
if (!http_response_compression_methods.empty())
{
/// If client supports brotli - it's preferred.
/// Both gzip and deflate are supported. If the client supports both, gzip is preferred.
/// NOTE parsing of the list of methods is slightly incorrect.
if (std::string::npos != http_response_compression_methods.find("br"))
http_response_compression_method = CompressionMethod::Brotli;
else if (std::string::npos != http_response_compression_methods.find("gzip"))
http_response_compression_method = CompressionMethod::Gzip;
else if (std::string::npos != http_response_compression_methods.find("deflate"))
http_response_compression_method = CompressionMethod::Zlib;
else if (std::string::npos != http_response_compression_methods.find("xz"))
http_response_compression_method = CompressionMethod::Xz;
else if (std::string::npos != http_response_compression_methods.find("zstd"))
http_response_compression_method = CompressionMethod::Zstd;
}
2020-01-04 22:59:08 +00:00
bool client_supports_http_compression = http_response_compression_method != CompressionMethod::None;
/// Client can pass a 'compress' flag in the query string. In this case the query result is
/// compressed using internal algorithm. This is not reflected in HTTP headers.
bool internal_compression = params.getParsed<bool>("compress", false);
/// At least, we should postpone sending of first buffer_size result bytes
size_t buffer_size_total = std::max(
params.getParsed<size_t>("buffer_size", DBMS_DEFAULT_BUFFER_SIZE), static_cast<size_t>(DBMS_DEFAULT_BUFFER_SIZE));
/// If it is specified, the whole result will be buffered.
/// First ~buffer_size bytes will be buffered in memory, the remaining bytes will be stored in temporary file.
bool buffer_until_eof = params.getParsed<bool>("wait_end_of_query", false);
size_t buffer_size_http = DBMS_DEFAULT_BUFFER_SIZE;
size_t buffer_size_memory = (buffer_size_total > buffer_size_http) ? buffer_size_total : 0;
unsigned keep_alive_timeout = config.getUInt("keep_alive_timeout", 10);
used_output.out = std::make_shared<WriteBufferFromHTTPServerResponse>(
response,
request.getMethod() == HTTPRequest::HTTP_HEAD,
keep_alive_timeout,
client_supports_http_compression,
http_response_compression_method);
2020-01-04 22:59:08 +00:00
if (internal_compression)
used_output.out_maybe_compressed = std::make_shared<CompressedWriteBuffer>(*used_output.out);
else
used_output.out_maybe_compressed = used_output.out;
if (buffer_size_memory > 0 || buffer_until_eof)
{
CascadeWriteBuffer::WriteBufferPtrs cascade_buffer1;
CascadeWriteBuffer::WriteBufferConstructors cascade_buffer2;
if (buffer_size_memory > 0)
cascade_buffer1.emplace_back(std::make_shared<MemoryWriteBuffer>(buffer_size_memory));
if (buffer_until_eof)
{
2021-08-01 14:12:34 +00:00
const std::string tmp_path(server.context()->getTemporaryVolume()->getDisk()->getPath());
const std::string tmp_path_template(tmp_path + "http_buffers/");
auto create_tmp_disk_buffer = [tmp_path_template] (const WriteBufferPtr &)
{
return WriteBufferFromTemporaryFile::create(tmp_path_template);
};
cascade_buffer2.emplace_back(std::move(create_tmp_disk_buffer));
}
else
{
auto push_memory_buffer_and_continue = [next_buffer = used_output.out_maybe_compressed] (const WriteBufferPtr & prev_buf)
{
2020-05-18 08:08:55 +00:00
auto * prev_memory_buffer = typeid_cast<MemoryWriteBuffer *>(prev_buf.get());
if (!prev_memory_buffer)
throw Exception("Expected MemoryWriteBuffer", ErrorCodes::LOGICAL_ERROR);
auto rdbuf = prev_memory_buffer->tryGetReadBuffer();
copyData(*rdbuf , *next_buffer);
return next_buffer;
};
cascade_buffer2.emplace_back(push_memory_buffer_and_continue);
}
used_output.out_maybe_delayed_and_compressed = std::make_shared<CascadeWriteBuffer>(
std::move(cascade_buffer1), std::move(cascade_buffer2));
}
else
{
used_output.out_maybe_delayed_and_compressed = used_output.out_maybe_compressed;
}
/// Request body can be compressed using algorithm specified in the Content-Encoding header.
String http_request_compression_method_str = request.get("Content-Encoding", "");
auto in_post = wrapReadBufferWithCompressionMethod(
wrapReadBufferReference(request.getStream()), chooseCompressionMethod({}, http_request_compression_method_str));
/// The data can also be compressed using incompatible internal algorithm. This is indicated by
/// 'decompress' query parameter.
std::unique_ptr<ReadBuffer> in_post_maybe_compressed;
bool in_post_compressed = false;
if (params.getParsed<bool>("decompress", false))
{
in_post_maybe_compressed = std::make_unique<CompressedReadBuffer>(*in_post);
in_post_compressed = true;
}
else
in_post_maybe_compressed = std::move(in_post);
std::unique_ptr<ReadBuffer> in;
static const NameSet reserved_param_names{"compress", "decompress", "user", "password", "quota_key", "query_id", "stacktrace",
"buffer_size", "wait_end_of_query", "session_id", "session_timeout", "session_check"};
Names reserved_param_suffixes;
auto param_could_be_skipped = [&] (const String & name)
{
2020-06-14 00:56:13 +00:00
/// Empty parameter appears when URL like ?&a=b or a=b&&c=d. Just skip them for user's convenience.
if (name.empty())
return true;
if (reserved_param_names.count(name))
return true;
for (const String & suffix : reserved_param_suffixes)
{
if (endsWith(name, suffix))
return true;
}
return false;
};
/// Settings can be overridden in the query.
/// Some parameters (database, default_format, everything used in the code above) do not
/// belong to the Settings class.
/// 'readonly' setting values mean:
/// readonly = 0 - any query is allowed, client can change any setting.
/// readonly = 1 - only readonly queries are allowed, client can't change settings.
/// readonly = 2 - only readonly queries are allowed, client can change any setting except 'readonly'.
/// In theory if initially readonly = 0, the client can change any setting and then set readonly
/// to some other value.
const auto & settings = context->getSettingsRef();
/// Only readonly queries are allowed for HTTP GET requests.
if (request.getMethod() == HTTPServerRequest::HTTP_GET)
{
if (settings.readonly == 0)
context->setSetting("readonly", 2);
}
2020-03-09 03:38:43 +00:00
bool has_external_data = startsWith(request.getContentType(), "multipart/form-data");
2019-06-15 18:25:27 +00:00
if (has_external_data)
{
/// Skip unneeded parameters to avoid confusing them later with context settings or query parameters.
reserved_param_suffixes.reserve(3);
/// It is a bug and ambiguity with `date_time_input_format` and `low_cardinality_allow_in_native_format` formats/settings.
reserved_param_suffixes.emplace_back("_format");
reserved_param_suffixes.emplace_back("_types");
reserved_param_suffixes.emplace_back("_structure");
}
std::string database = request.get("X-ClickHouse-Database", "");
std::string default_format = request.get("X-ClickHouse-Format", "");
SettingsChanges settings_changes;
2019-05-29 00:23:47 +00:00
for (const auto & [key, value] : params)
{
if (key == "database")
{
if (database.empty())
database = value;
}
else if (key == "default_format")
{
if (default_format.empty())
default_format = value;
}
else if (param_could_be_skipped(key))
{
}
else
{
2020-04-04 08:57:16 +00:00
/// Other than query parameters are treated as settings.
if (!customizeQueryParam(context, key, value))
settings_changes.push_back({key, value});
}
}
if (!database.empty())
context->setCurrentDatabase(database);
if (!default_format.empty())
context->setDefaultFormat(default_format);
/// For external data we also want settings
context->checkSettingsConstraints(settings_changes);
context->applySettingsChanges(settings_changes);
2021-08-01 14:12:34 +00:00
// Set the query id supplied by the user, if any, and also update the OpenTelemetry fields.
context->setCurrentQueryId(params.get("query_id", request.get("X-ClickHouse-Query-Id", "")));
2021-08-16 10:58:52 +00:00
const auto & query = getQuery(request, params, context);
2020-04-04 08:57:16 +00:00
std::unique_ptr<ReadBuffer> in_param = std::make_unique<ReadBufferFromString>(query);
in = has_external_data ? std::move(in_param) : std::make_unique<ConcatReadBuffer>(*in_param, *in_post_maybe_compressed);
/// HTTP response compression is turned on only if the client signalled that they support it
/// (using Accept-Encoding header) and 'enable_http_compression' setting is turned on.
used_output.out->setCompression(client_supports_http_compression && settings.enable_http_compression);
if (client_supports_http_compression)
used_output.out->setCompressionLevel(settings.http_zlib_compression_level);
used_output.out->setSendProgressInterval(settings.http_headers_progress_interval_ms);
/// If 'http_native_compression_disable_checksumming_on_decompress' setting is turned on,
/// checksums of client data compressed with internal algorithm are not checked.
if (in_post_compressed && settings.http_native_compression_disable_checksumming_on_decompress)
static_cast<CompressedReadBuffer &>(*in_post_maybe_compressed).disableChecksumming();
2021-10-09 12:16:37 +00:00
/// Add CORS header if 'add_http_cors_header' setting is turned on send * in Access-Control-Allow-Origin,
/// or if config has http_options_response, which means that there
/// are some headers to be sent, and the client passed Origin header.
if (config.has("http_options_response") && !request.get("Origin", "").empty())
tryAddHeadersFromConfig(response, config);
2021-10-09 12:16:37 +00:00
else if (settings.add_http_cors_header)
used_output.out->addHeaderCORS(settings.add_http_cors_header && !request.get("Origin", "").empty());
auto append_callback = [context = context] (ProgressCallback callback)
{
auto prev = context->getProgressCallback();
context->setProgressCallback([prev, callback] (const Progress & progress)
{
if (prev)
prev(progress);
callback(progress);
});
};
/// While still no data has been sent, we will report about query execution progress by sending HTTP headers.
if (settings.send_progress_in_http_headers)
append_callback([&used_output] (const Progress & progress) { used_output.out->onProgress(progress); });
if (settings.readonly > 0 && settings.cancel_http_readonly_queries_on_client_close)
{
append_callback([context = context, &request](const Progress &)
{
/// Assume that at the point this method is called no one is reading data from the socket any more:
/// should be true for read-only queries.
if (!request.checkPeerConnected())
context->killCurrentQuery();
});
}
customizeContext(request, context);
query_scope.emplace(context);
executeQuery(*in, *used_output.out_maybe_delayed_and_compressed, /* allow_into_outfile = */ false, context,
[&response] (const String & current_query_id, const String & content_type, const String & format, const String & timezone)
{
response.setContentType(content_type);
response.add("X-ClickHouse-Query-Id", current_query_id);
response.add("X-ClickHouse-Format", format);
response.add("X-ClickHouse-Timezone", timezone);
}
);
if (used_output.hasDelayed())
{
/// TODO: set Content-Length if possible
pushDelayedResults(used_output);
}
2021-04-20 12:57:23 +00:00
/// Send HTTP headers with code 200 if no exception happened and the data is still not sent to the client.
used_output.finalize();
2012-03-09 03:06:09 +00:00
}
void HTTPHandler::trySendExceptionToClient(
const std::string & s, int exception_code, HTTPServerRequest & request, HTTPServerResponse & response, Output & used_output)
try
2012-03-09 03:06:09 +00:00
{
response.set("X-ClickHouse-Exception-Code", toString<int>(exception_code));
/// FIXME: make sure that no one else is reading from the same stream at the moment.
/// If HTTP method is POST and Keep-Alive is turned on, we should read the whole request body
/// to avoid reading part of the current request body in the next request.
if (request.getMethod() == Poco::Net::HTTPRequest::HTTP_POST
&& response.getKeepAlive()
&& exception_code != ErrorCodes::HTTP_LENGTH_REQUIRED
&& !request.getStream().eof())
{
request.getStream().ignoreAll();
}
2020-01-23 13:40:16 +00:00
if (exception_code == ErrorCodes::REQUIRED_PASSWORD)
{
response.requireAuthentication("ClickHouse server HTTP API");
}
else
{
response.setStatusAndReason(exceptionCodeToHTTPStatus(exception_code));
}
if (!response.sent() && !used_output.out_maybe_compressed)
{
/// If nothing was sent yet and we don't even know if we must compress the response.
*response.send() << s << std::endl;
}
else if (used_output.out_maybe_compressed)
{
/// Destroy CascadeBuffer to actualize buffers' positions and reset extra references
if (used_output.hasDelayed())
used_output.out_maybe_delayed_and_compressed.reset();
/// Send the error message into already used (and possibly compressed) stream.
/// Note that the error message will possibly be sent after some data.
/// Also HTTP code 200 could have already been sent.
/// If buffer has data, and that data wasn't sent yet, then no need to send that data
bool data_sent = used_output.out->count() != used_output.out->offset();
if (!data_sent)
{
used_output.out_maybe_compressed->position() = used_output.out_maybe_compressed->buffer().begin();
used_output.out->position() = used_output.out->buffer().begin();
}
writeString(s, *used_output.out_maybe_compressed);
writeChar('\n', *used_output.out_maybe_compressed);
used_output.out_maybe_compressed->next();
}
else
{
assert(false);
__builtin_unreachable();
}
used_output.finalize();
}
catch (...)
{
tryLogCurrentException(log, "Cannot send exception to client");
try
{
used_output.finalize();
}
catch (...)
{
tryLogCurrentException(log, "Cannot flush data to client (after sending exception)");
}
}
2012-06-25 05:07:34 +00:00
void HTTPHandler::handleRequest(HTTPServerRequest & request, HTTPServerResponse & response)
{
2018-08-31 00:59:48 +00:00
setThreadName("HTTPHandler");
ThreadStatus thread_status;
2018-08-31 00:59:48 +00:00
2021-08-01 14:12:34 +00:00
session = std::make_unique<Session>(server.context(), ClientInfo::Interface::HTTP);
SCOPE_EXIT({ session.reset(); });
std::optional<CurrentThread::QueryScope> query_scope;
2021-08-01 14:12:34 +00:00
Output used_output;
/// In case of exception, send stack trace to client.
bool with_stacktrace = false;
try
{
2021-09-18 09:36:02 +00:00
if (request.getMethod() == HTTPServerRequest::HTTP_OPTIONS)
{
processOptionsRequest(response, server.config());
return;
}
response.setContentType("text/plain; charset=UTF-8");
response.set("X-ClickHouse-Server-Display-Name", server_display_name);
/// For keep-alive to work.
if (request.getVersion() == HTTPServerRequest::HTTP_1_1)
response.setChunkedTransferEncoding(true);
2021-08-01 14:12:34 +00:00
HTMLForm params(default_settings, request);
with_stacktrace = params.getParsed<bool>("stacktrace", false);
/// FIXME: maybe this check is already unnecessary.
/// Workaround. Poco does not detect 411 Length Required case.
if (request.getMethod() == HTTPRequest::HTTP_POST && !request.getChunkedTransferEncoding() && !request.hasContentLength())
{
throw Exception(
"The Transfer-Encoding is not chunked and there is no Content-Length header for POST request",
ErrorCodes::HTTP_LENGTH_REQUIRED);
}
processQuery(request, params, response, used_output, query_scope);
LOG_DEBUG(log, (request_credentials ? "Authentication in progress..." : "Done processing query"));
}
catch (...)
{
SCOPE_EXIT({
request_credentials.reset(); // ...so that the next requests on the connection have to always start afresh in case of exceptions.
});
tryLogCurrentException(log);
/** If exception is received from remote server, then stack trace is embedded in message.
* If exception is thrown on local server, then stack trace is in separate field.
*/
std::string exception_message = getCurrentExceptionMessage(with_stacktrace, true);
int exception_code = getCurrentExceptionCode();
trySendExceptionToClient(exception_message, exception_code, request, response, used_output);
}
Fix abnormal server termination when http client goes away In [1] stress tests found: 2021.02.12 14:20:58.800988 [ 17728 ] {de3e7894-b401-4f7d-8530-90cd5ab06682} <Debug> executeQuery: (from [::1]:45792, using production parser) (comment: /usr/share/clickhouse-test/queries/0_stateless/01520_client_print_query_id.expect) SELECT * FROM numbers(34599) 2021.02.12 14:20:58.916484 [ 17728 ] {de3e7894-b401-4f7d-8530-90cd5ab06682} <Trace> ContextAccess (default): Access granted: CREATE TEMPORARY TABLE ON *.* 2021.02.12 14:20:59.071980 [ 17728 ] {de3e7894-b401-4f7d-8530-90cd5ab06682} <Trace> InterpreterSelectQuery: FetchColumns -> Complete 2021.02.12 14:21:10.708202 [ 17728 ] {de3e7894-b401-4f7d-8530-90cd5ab06682} <Information> executeQuery: Read 34599 rows, 270.30 KiB in 11.876294055 sec., 2913 rows/sec., 22.76 KiB/sec. 2021.02.12 14:22:10.506261 [ 17728 ] {de3e7894-b401-4f7d-8530-90cd5ab06682} <Debug> DynamicQueryHandler: Done processing query 2021.02.12 14:22:18.238037 [ 375 ] {} <Fatal> BaseDaemon: (version 21.3.1.5996, build id: 8DBCED54529C989F7AD4D991F51410774D55DE6C) (from thread 17728) Terminate called for uncaught exception: Code: 24, e.displayText() = DB::Exception: Cannot write to ostream at offset 262994, Stack trace (when copying this message, always include the lines below): 0. ./obj-x86_64-linux-gnu/../contrib/libcxx/include/exception:0: Poco::Exception::Exception(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, int) @ 0x15c976cb in /usr/bin/clickhouse 1. ./obj-x86_64-linux-gnu/../src/Common/Exception.cpp:56: DB::Exception::Exception(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, int, bool) @ 0x8c9320e in /usr/bin/clickhouse 2. ./obj-x86_64-linux-gnu/../src/IO/WriteBufferFromOStream.cpp:0: DB::WriteBufferFromOStream::nextImpl() @ 0x8d54da5 in /usr/bin/clickhouse 3. ./obj-x86_64-linux-gnu/../src/IO/BufferBase.h:39: DB::WriteBufferFromOStream::~WriteBufferFromOStream() @ 0x8d551d7 in /usr/bin/clickhouse 4. ./obj-x86_64-linux-gnu/../src/IO/WriteBufferFromOStream.cpp:44: DB::Write 2021.02.12 14:22:18.811071 [ 18134 ] {} <Fatal> BaseDaemon: ######################################## 2021.02.12 14:22:18.878935 [ 18134 ] {} <Fatal> BaseDaemon: (version 21.3.1.5996, build id: 8DBCED54529C989F7AD4D991F51410774D55DE6C) (from thread 17728) (query_id: de3e7894-b401-4f7d-8530-90cd5ab06682) Received signal Aborted (6) 2021.02.12 14:22:18.943148 [ 18134 ] {} <Fatal> BaseDaemon: 2021.02.12 14:22:19.007073 [ 18134 ] {} <Fatal> BaseDaemon: Stack trace: 0x7f109932018b 0x7f10992ff859 0x8bb33ae 0x8e301dd 0x17dac8c4 0x17dac7c7 0x8c3fe0b 0x8d552c5 0x8d552ea 0x11a29914 0x11a2a2ca 0x12f96092 0x12f8c65e 0x12f84300 0x15b84110 0x15bc0913 0x15bc103f 0x15d29a12 0x15d27fb0 0x15d267b8 0x8badbad 0x7f10994d5609 0x7f10993fc293 2021.02.12 14:22:19.255998 [ 18134 ] {} <Fatal> BaseDaemon: 5. raise @ 0x4618b in /usr/lib/x86_64-linux-gnu/libc-2.31.so 2021.02.12 14:22:19.270203 [ 18134 ] {} <Fatal> BaseDaemon: 6. abort @ 0x25859 in /usr/lib/x86_64-linux-gnu/libc-2.31.so 2021.02.12 14:22:50.108918 [ 370 ] {} <Fatal> Application: Child process was terminated by signal 6. [1]: https://clickhouse-test-reports.s3.yandex.net/19580/6aecb62416ece880cbb8ee3a803e14d841388dde/stress_test_(thread).html#fail1 Verified locally by commenting out->next() call in WriteBufferFromHTTPServerResponse::nextImpl(), adding a sleep(1) and canceling HTTP request before it finished, the stacktrace as follow: [ 6351 ] {} <Fatal> BaseDaemon: (version 21.3.1.1, build id: 9B40466BF3D2F5AED78A52A995A4A2FD3116787C) (from thread 6677) Terminate called for uncaught exception: Code: 24, e.displayText() = DB::Exception: Cannot write to ostream at offset 4, Stack trace (when copying this message, always include the lines below): 0. /src/ch/clickhouse/.cmake/../src/Common/StackTrace.cpp:298: StackTrace::tryCapture() @ 0x30a52a in /src/ch/clickhouse/.cmake/src/libclickhouse_common_iod.so 1. /src/ch/clickhouse/.cmake/../src/Common/StackTrace.cpp:260: StackTrace::StackTrace() @ 0x30a4e5 in /src/ch/clickhouse/.cmake/src/libclickhouse_common_iod.so 2. /src/ch/clickhouse/.cmake/../src/Common/Exception.cpp:53: DB::Exception::Exception(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, int, bool) @ 0x2a61ae in /src/ch/clickhouse/.cmake/src/libclickhouse_common_iod.so 3. /src/ch/clickhouse/.cmake/../src/IO/WriteBufferFromOStream.cpp:22: DB::WriteBufferFromOStream::nextImpl() @ 0x3b468a in /src/ch/clickhouse/.cmake/src/libclickhouse_common_iod.so 4. /src/ch/clickhouse/.cmake/../src/IO/WriteBuffer.h:47 <snip> [ 8966 ] {} <Fatal> BaseDaemon: 7. __cxxabiv1::__terminate(void (*)()) @ 0x1784ca in /src/ch/clickhouse/.cmake/contrib/replxx-cmake/libreplxxd.so <snip> [ 8966 ] {} <Fatal> BaseDaemon: 10. /src/ch/clickhouse/.cmake/../src/IO/WriteBufferFromOStream.cpp:0: DB::WriteBufferFromOStream::~WriteBufferFromOStream() @ 0x3b48c1 in /src/ch/clickhouse/.cmake/src/libclickhouse_common_iod.so [ 8966 ] {} <Fatal> BaseDaemon: 11. /src/ch/clickhouse/.cmake/../src/IO/WriteBufferFromOStream.cpp:44: DB::WriteBufferFromOStream::~WriteBufferFromOStream() @ 0x3b48ec in /src/ch/clickhouse/.cmake/src/libclickhouse_common_iod.so <snip> [ 8966 ] {} <Fatal> BaseDaemon: 14. /src/ch/clickhouse/.cmake/../src/IO/WriteBufferFromHTTPServerResponse.cpp:218: DB::WriteBufferFromHTTPServerResponse::~WriteBufferFromHTTPServerResponse() @ 0x3b33cd in /src/ch/clickhouse/.cmake/src/libclickhouse_common_iod.so <snip> [ 8966 ] {} <Fatal> BaseDaemon: 22. /src/ch/clickhouse/.cmake/../src/Server/HTTPHandler.h:43: DB::HTTPHandler::Output::~Output() @ 0x260421 in /src/ch/clickhouse/.cmake/src/libclickhouse_serverd.so [ 8966 ] {} <Fatal> BaseDaemon: 23. /src/ch/clickhouse/.cmake/../src/Server/HTTPHandler.cpp:778: DB::HTTPHandler::handleRequest(Poco::Net::HTTPServerRequest&, Poco::Net::HTTPServerResponse&) @ 0x253fd4 in /src/ch/clickhouse/.cmake/src/libclickhouse_serverd.so
2021-02-13 09:04:03 +00:00
used_output.finalize();
2012-03-09 03:06:09 +00:00
}
2020-04-04 08:57:16 +00:00
DynamicQueryHandler::DynamicQueryHandler(IServer & server_, const std::string & param_name_)
: HTTPHandler(server_, "DynamicQueryHandler"), param_name(param_name_)
{
2020-04-04 08:57:16 +00:00
}
2021-05-31 14:49:02 +00:00
bool DynamicQueryHandler::customizeQueryParam(ContextMutablePtr context, const std::string & key, const std::string & value)
2020-04-04 08:57:16 +00:00
{
if (key == param_name)
return true; /// do nothing
2020-04-04 08:57:16 +00:00
if (startsWith(key, "param_"))
{
2020-04-04 08:57:16 +00:00
/// Save name and values of substitution in dictionary.
const String parameter_name = key.substr(strlen("param_"));
2021-08-16 10:58:52 +00:00
2021-08-16 21:54:40 +00:00
if (!context->getQueryParameters().contains(parameter_name))
2021-08-16 10:58:52 +00:00
context->setQueryParameter(parameter_name, value);
2020-04-04 08:57:16 +00:00
return true;
}
2020-04-04 08:57:16 +00:00
return false;
2012-03-09 03:06:09 +00:00
}
2021-05-31 14:49:02 +00:00
std::string DynamicQueryHandler::getQuery(HTTPServerRequest & request, HTMLForm & params, ContextMutablePtr context)
2020-04-04 08:57:16 +00:00
{
if (likely(!startsWith(request.getContentType(), "multipart/form-data")))
{
/// Part of the query can be passed in the 'query' parameter and the rest in the request body
/// (http method need not necessarily be POST). In this case the entire query consists of the
/// contents of the 'query' parameter, a line break and the request body.
std::string query_param = params.get(param_name, "");
return query_param.empty() ? query_param : query_param + "\n";
}
/// Support for "external data for query processing".
/// Used in case of POST request with form-data, but it isn't expected to be deleted after that scope.
ExternalTablesHandler handler(context, params);
params.load(request, request.getStream(), handler);
2020-04-04 08:57:16 +00:00
std::string full_query;
/// Params are of both form params POST and uri (GET params)
for (const auto & it : params)
2021-08-16 10:58:52 +00:00
{
2020-04-04 08:57:16 +00:00
if (it.first == param_name)
2021-08-16 10:58:52 +00:00
{
2020-04-04 08:57:16 +00:00
full_query += it.second;
2021-08-16 10:58:52 +00:00
}
else
{
customizeQueryParam(context, it.first, it.second);
}
}
2020-04-04 08:57:16 +00:00
return full_query;
}
2020-04-27 20:51:39 +00:00
PredefinedQueryHandler::PredefinedQueryHandler(
IServer & server_,
const NameSet & receive_params_,
const std::string & predefined_query_,
const CompiledRegexPtr & url_regex_,
const std::unordered_map<String, CompiledRegexPtr> & header_name_with_regex_)
: HTTPHandler(server_, "PredefinedQueryHandler")
, receive_params(receive_params_)
, predefined_query(predefined_query_)
, url_regex(url_regex_)
, header_name_with_capture_regex(header_name_with_regex_)
2020-04-04 08:57:16 +00:00
{
}
2021-05-31 14:49:02 +00:00
bool PredefinedQueryHandler::customizeQueryParam(ContextMutablePtr context, const std::string & key, const std::string & value)
2020-04-04 08:57:16 +00:00
{
if (receive_params.count(key))
{
context->setQueryParameter(key, value);
2020-04-04 08:57:16 +00:00
return true;
}
return false;
}
2021-05-31 14:49:02 +00:00
void PredefinedQueryHandler::customizeContext(HTTPServerRequest & request, ContextMutablePtr context)
{
/// If in the configuration file, the handler's header is regex and contains named capture group
/// We will extract regex named capture groups as query parameters
const auto & set_query_params = [&](const char * begin, const char * end, const CompiledRegexPtr & compiled_regex)
{
int num_captures = compiled_regex->NumberOfCapturingGroups() + 1;
2020-04-28 07:19:37 +00:00
re2::StringPiece matches[num_captures];
re2::StringPiece input(begin, end - begin);
if (compiled_regex->Match(input, 0, end - begin, re2::RE2::Anchor::ANCHOR_BOTH, matches, num_captures))
{
for (const auto & [capturing_name, capturing_index] : compiled_regex->NamedCapturingGroups())
{
const auto & capturing_value = matches[capturing_index];
if (capturing_value.data())
context->setQueryParameter(capturing_name, String(capturing_value.data(), capturing_value.size()));
}
}
};
2020-04-26 11:06:14 +00:00
if (url_regex)
{
const auto & uri = request.getURI();
2020-04-27 22:19:24 +00:00
set_query_params(uri.data(), find_first_symbols<'?'>(uri.data(), uri.data() + uri.size()), url_regex);
2020-04-26 11:06:14 +00:00
}
for (const auto & [header_name, regex] : header_name_with_capture_regex)
{
2020-04-26 11:06:14 +00:00
const auto & header_value = request.get(header_name);
set_query_params(header_value.data(), header_value.data() + header_value.size(), regex);
}
}
2021-05-31 14:49:02 +00:00
std::string PredefinedQueryHandler::getQuery(HTTPServerRequest & request, HTMLForm & params, ContextMutablePtr context)
2020-04-04 08:57:16 +00:00
{
if (unlikely(startsWith(request.getContentType(), "multipart/form-data")))
{
/// Support for "external data for query processing".
ExternalTablesHandler handler(context, params);
params.load(request, request.getStream(), handler);
2020-04-04 08:57:16 +00:00
}
2020-04-27 20:51:39 +00:00
return predefined_query;
2020-04-04 08:57:16 +00:00
}
HTTPRequestHandlerFactoryPtr createDynamicHandlerFactory(IServer & server, const std::string & config_prefix)
2020-04-04 08:57:16 +00:00
{
const auto & query_param_name = server.config().getString(config_prefix + ".handler.query_param_name", "query");
auto factory = std::make_shared<HandlingRuleHTTPHandlerFactory<DynamicQueryHandler>>(server, std::move(query_param_name));
factory->addFiltersFromConfig(server.config(), config_prefix);
return factory;
2020-04-04 08:57:16 +00:00
}
2020-04-27 22:19:24 +00:00
static inline bool capturingNamedQueryParam(NameSet receive_params, const CompiledRegexPtr & compiled_regex)
2020-04-26 11:06:14 +00:00
{
const auto & capturing_names = compiled_regex->NamedCapturingGroups();
return std::count_if(capturing_names.begin(), capturing_names.end(), [&](const auto & iterator)
{
return std::count_if(receive_params.begin(), receive_params.end(),
[&](const auto & param_name) { return param_name == iterator.first; });
2020-04-26 11:06:14 +00:00
});
}
2020-04-04 08:57:16 +00:00
2020-04-27 22:19:24 +00:00
static inline CompiledRegexPtr getCompiledRegex(const std::string & expression)
{
2020-04-28 07:19:37 +00:00
auto compiled_regex = std::make_shared<const re2::RE2>(expression);
2020-04-27 22:19:24 +00:00
if (!compiled_regex->ok())
throw Exception(
"Cannot compile re2: " + expression + " for http handling rule, error: " + compiled_regex->error()
+ ". Look at https://github.com/google/re2/wiki/Syntax for reference.",
ErrorCodes::CANNOT_COMPILE_REGEXP);
2020-04-27 22:19:24 +00:00
return compiled_regex;
}
HTTPRequestHandlerFactoryPtr createPredefinedHandlerFactory(IServer & server, const std::string & config_prefix)
2020-04-04 08:57:16 +00:00
{
2020-04-26 11:06:14 +00:00
Poco::Util::AbstractConfiguration & configuration = server.config();
if (!configuration.has(config_prefix + ".handler.query"))
throw Exception("There is no path '" + config_prefix + ".handler.query' in configuration file.", ErrorCodes::NO_ELEMENTS_IN_CONFIG);
2020-04-04 08:57:16 +00:00
2020-04-27 20:51:39 +00:00
std::string predefined_query = configuration.getString(config_prefix + ".handler.query");
NameSet analyze_receive_params = analyzeReceiveQueryParams(predefined_query);
2020-04-27 22:19:24 +00:00
std::unordered_map<String, CompiledRegexPtr> headers_name_with_regex;
2020-04-26 11:06:14 +00:00
Poco::Util::AbstractConfiguration::Keys headers_name;
configuration.keys(config_prefix + ".headers", headers_name);
2020-04-26 11:06:14 +00:00
for (const auto & header_name : headers_name)
{
2020-04-26 11:06:14 +00:00
auto expression = configuration.getString(config_prefix + ".headers." + header_name);
2020-04-04 08:57:16 +00:00
2020-04-26 11:06:14 +00:00
if (!startsWith(expression, "regex:"))
continue;
expression = expression.substr(6);
2020-04-27 22:19:24 +00:00
auto regex = getCompiledRegex(expression);
if (capturingNamedQueryParam(analyze_receive_params, regex))
headers_name_with_regex.emplace(std::make_pair(header_name, regex));
}
2020-04-26 11:06:14 +00:00
std::shared_ptr<HandlingRuleHTTPHandlerFactory<PredefinedQueryHandler>> factory;
2020-04-26 11:06:14 +00:00
if (configuration.has(config_prefix + ".url"))
{
auto url_expression = configuration.getString(config_prefix + ".url");
if (startsWith(url_expression, "regex:"))
url_expression = url_expression.substr(6);
2020-04-27 22:19:24 +00:00
auto regex = getCompiledRegex(url_expression);
if (capturingNamedQueryParam(analyze_receive_params, regex))
{
factory = std::make_shared<HandlingRuleHTTPHandlerFactory<PredefinedQueryHandler>>(
server,
std::move(analyze_receive_params),
std::move(predefined_query),
std::move(regex),
std::move(headers_name_with_regex));
factory->addFiltersFromConfig(configuration, config_prefix);
return factory;
}
}
factory = std::make_shared<HandlingRuleHTTPHandlerFactory<PredefinedQueryHandler>>(
server, std::move(analyze_receive_params), std::move(predefined_query), CompiledRegexPtr{}, std::move(headers_name_with_regex));
factory->addFiltersFromConfig(configuration, config_prefix);
return factory;
2020-04-26 11:06:14 +00:00
}
2012-03-09 03:06:09 +00:00
}