Merge pull request #11840 from azat/fix-http-memory-accounting

Fix memory accounting via HTTP interface

(cherry picked from commit b447508)

programs/server/HTTPHandler.cpp

@@ -220,15 +220,12 @@ HTTPHandler::HTTPHandler(IServer & server_, const std::string & name)
 
 
 void HTTPHandler::processQuery(
+    Context & context,
     Poco::Net::HTTPServerRequest & request,
     HTMLForm & params,
     Poco::Net::HTTPServerResponse & response,
     Output & used_output)
 {
-    Context context = server.context();
-
-    CurrentThread::QueryScope query_scope(context);
-
     LOG_TRACE(log, "Request URI: " << request.getURI());
 
     std::istream & istr = request.stream();
@@ -667,6 +664,11 @@ void HTTPHandler::handleRequest(Poco::Net::HTTPServerRequest & request, Poco::Ne
     setThreadName("HTTPHandler");
     ThreadStatus thread_status;
 
+   /// Should be initialized before anything,
+    /// For correct memory accounting.
+    Context context = server.context();
+    CurrentThread::QueryScope query_scope(context);
+
     Output used_output;
 
     /// In case of exception, send stack trace to client.
@@ -690,7 +692,7 @@ void HTTPHandler::handleRequest(Poco::Net::HTTPServerRequest & request, Poco::Ne
             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);
+        processQuery(context, request, params, response, used_output);
         LOG_INFO(log, "Done processing query");
     }
     catch (...)

programs/server/HTTPHandler.h

@@ -71,6 +71,7 @@ class HTTPHandler : public Poco::Net::HTTPRequestHandler
 
     /// Also initializes 'used_output'.
     void processQuery(
+        Context & context,
         Poco::Net::HTTPServerRequest & request,
         HTMLForm & params,
         Poco::Net::HTTPServerResponse & response,

tests/queries/0_stateless/01091_num_threads.sql

@@ -1,44 +1,44 @@
 set log_queries=1;
 set log_query_threads=1;
 
-SELECT 1;
+WITH 01091 AS id SELECT 1;
 SYSTEM FLUSH LOGS;
 
 WITH 
     (
         SELECT query_id
         FROM system.query_log
-        WHERE (query = 'SELECT 1') AND (event_date >= (today() - 1))
+        WHERE (query = 'WITH 01091 AS id SELECT 1;\n') AND (event_date >= (today() - 1))
         ORDER BY event_time DESC
         LIMIT 1
     ) AS id
 SELECT uniqExact(thread_id)
 FROM system.query_thread_log
 WHERE (event_date >= (today() - 1)) AND (query_id = id) AND (thread_id != master_thread_id);
 
-select sum(number) from numbers(1000000);
+with 01091 as id select sum(number) from numbers(1000000);
 SYSTEM FLUSH LOGS;
 
 WITH 
     (
         SELECT query_id
         FROM system.query_log
-        WHERE (query = 'SELECT sum(number) FROM numbers(1000000)') AND (event_date >= (today() - 1))
+        WHERE (query LIKE 'with 01091 as id select sum(number) from numbers(1000000);%') AND (event_date >= (today() - 1))
         ORDER BY event_time DESC
         LIMIT 1
     ) AS id
 SELECT uniqExact(thread_id)
 FROM system.query_thread_log
 WHERE (event_date >= (today() - 1)) AND (query_id = id) AND (thread_id != master_thread_id);
 
-select sum(number) from numbers_mt(1000000);
+with 01091 as id select sum(number) from numbers_mt(1000000);
 SYSTEM FLUSH LOGS;
 
 WITH 
     (
         SELECT query_id
         FROM system.query_log
-        WHERE (query = 'SELECT sum(number) FROM numbers_mt(1000000)') AND (event_date >= (today() - 1))
+        WHERE (query LIKE 'with 01091 as id select sum(number) from numbers_mt(1000000);%') AND (event_date >= (today() - 1))
         ORDER BY event_time DESC
         LIMIT 1
     ) AS id

tests/queries/0_stateless/01238_http_memory_tracking.reference

@@ -0,0 +1 @@
+1000

tests/queries/0_stateless/01238_http_memory_tracking.sh

@@ -0,0 +1,19 @@
+#!/usr/bin/env bash
+
+CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
+. $CURDIR/../shell_config.sh
+
+set -o pipefail
+
+# This is needed to keep at least one running query for user for the time of test.
+# (1k http queries takes ~1 second, let's run for 5x more to avoid flaps)
+${CLICKHOUSE_CLIENT} --format Null -n <<<'SELECT sleepEachRow(1) FROM numbers(5)' &
+
+# ignore "yes: standard output: Broken pipe"
+yes 'SELECT 1' 2>/dev/null | {
+    head -n1000
+} | {
+    xargs -i ${CLICKHOUSE_CURL} -sS "${CLICKHOUSE_URL}&wait_end_of_query=1&max_memory_usage_for_user=$((1<<30))" -d '{}'
+} | grep -x -c 1
+
+wait