Run for some mins, there is a segmentation error

[jchen357]

Here is my file

main.cc

#include <influxdb_raw_db_utf8.h>
#include <influxdb_simple_api.h>
#include <influxdb_line.h>
#include <influxdb_simple_async_api.h>
#include "influx.h"

#include <iostream>
#include <thread>
#include <chrono>
#include <time.h>
#include <unistd.h>
#include <sys/time.h>

using namespace influxdb::api;
using namespace std;

int main(int argc, char *argv[]) {
	try {
		int keyNum = 0;

		influxdb::async_api::simple_db async_api("http://10.11.12.17:8086", "bluewhale");

		struct timeval begin;
		gettimeofday(&begin, NULL);

		while (1) {
			struct timeval tv,td;
			gettimeofday(&tv, NULL);
			long time = (long long) (tv.tv_sec) * 1000000000 + tv.tv_usec * 1000;

			influxdb::api::key_value_pairs tags;
			influxdb::api::key_value_pairs fields;
			string g_app_name = "test";
			string rankID = "1";
			std::cout << "key num: " << keyNum << std::endl;
			tags.add("appName", g_app_name);
			tags.add("svrID", rankID);
			fields.add("value", keyNum);
			xflow::Influx::Get()->insert(line("t_lr_key", tags, fields));

			gettimeofday(&td, NULL);
			std::cout << "cost time " << (td.tv_sec - tv.tv_sec) * 1000000 + (td.tv_usec - tv.tv_usec) << std::endl;

			std::cout << "current time " << (td.tv_sec - begin.tv_sec) / 60 << std::endl << std::endl;
			keyNum++;
			sleep(1);
		}

	}
	catch (std::exception const &e) {
		std::cerr << e.what() << std::endl;
	}

	return 0;
}

[d-led]

Thanks! I'll look at it. There are important pieces of information missing: how did you compile the library and your application, e.g., which Boost version is used.

To reproduce the example, I'd probably need to know what "influx.h" and xflow::Influx::Get() are, and how the singleton is initialized. Why not try to initialize it in the scope of main?

[d-led]

as you're using the async api, are you sure, you're not just running out of memory?

[d-led]

reproduced the code in a branch. Will wait for a couple of minutes. I don't see anything suspicious. You're not pushing too many points either. I'll try shutting down the database after a while to see what happens

[d-led]

when there's no connectivity to the DB the program terminates

$ bin/macosx/gmake/x64/Debug/demo 
key num: 0
cost time 172
current time 0

Trace/BPT trap: 5

↓

Process 3267 launched: '/Users/..../src/influxdb-cpp-rest/bin/macosx/gmake/x64/Debug/demo' (x86_64)
key num: 0
cost time 206
current time 0

Process 3267 stopped
* thread #2, stop reason = signal SIGTRAP
    frame #0: 0x00007fff6eb46b66 libsystem_kernel.dylib`__pthread_kill + 10
libsystem_kernel.dylib`__pthread_kill:
->  0x7fff6eb46b66 <+10>: jae    0x7fff6eb46b70            ; <+20>
    0x7fff6eb46b68 <+12>: movq   %rax, %rdi
    0x7fff6eb46b6b <+15>: jmp    0x7fff6eb3dae9            ; cerror_nocancel
    0x7fff6eb46b70 <+20>: retq   
Target 0: (demo) stopped.
(lldb) bt
* thread #2, stop reason = signal SIGTRAP
  * frame #0: 0x00007fff6eb46b66 libsystem_kernel.dylib`__pthread_kill + 10
    frame #1: 0x00007fff6ed11080 libsystem_pthread.dylib`pthread_kill + 333
    frame #2: 0x00007fff6ea546fe libsystem_c.dylib`raise + 26
    frame #3: 0x0000000100921ddd libcpprest.2.10.dylib`pplx::details::_ExceptionHolder::~_ExceptionHolder() + 53
    frame #4: 0x00007fff6c97d8fd libc++.1.dylib`std::__1::__shared_weak_count::__release_shared() + 43
    frame #5: 0x00000001008ff066 libcpprest.2.10.dylib`pplx::details::_Task_completion_event_impl<web::http::http_response>::~_Task_completion_event_impl() + 64
    frame #6: 0x00007fff6c97d8fd libc++.1.dylib`std::__1::__shared_weak_count::__release_shared() + 43
    frame #7: 0x00000001008fd9f6 libcpprest.2.10.dylib`web::http::client::details::request_context::~request_context() + 42
    frame #8: 0x00007fff6c97d8fd libc++.1.dylib`std::__1::__shared_weak_count::__release_shared() + 43
    frame #9: 0x0000000100a0309a libcpprest.2.10.dylib`boost::asio::detail::reactive_socket_connect_op<boost::_bi::bind_t<void, boost::_mfi::mf2<void, web::http::client::details::asio_context, boost::system::error_code const&, boost::asio::ip::basic_resolver_iterator<boost::asio::ip::tcp> >, boost::_bi::list3<boost::_bi::value<std::__1::shared_ptr<web::http::client::details::asio_context> >, boost::arg<1> (*)(), boost::_bi::value<boost::asio::ip::basic_resolver_iterator<boost::asio::ip::tcp> > > > >::do_complete(void*, boost::asio::detail::scheduler_operation*, boost::system::error_code const&, unsigned long) + 160
    frame #10: 0x000000010098ad9a libcpprest.2.10.dylib`boost::asio::detail::scheduler::do_run_one(boost::asio::detail::conditionally_enabled_mutex::scoped_lock&, boost::asio::detail::scheduler_thread_info&, boost::system::error_code const&) + 410
    frame #11: 0x00000001009e3188 libcpprest.2.10.dylib`boost::asio::detail::scheduler::run(boost::system::error_code&) + 296
    frame #12: 0x00000001009e2ff7 libcpprest.2.10.dylib`boost::asio::detail::posix_thread::func<(anonymous namespace)::threadpool_impl::add_thread()::'lambda'()>::run() + 47
    frame #13: 0x00000001009e8bc8 libcpprest.2.10.dylib`boost_asio_detail_posix_thread_function + 16
    frame #14: 0x00007fff6ed0e661 libsystem_pthread.dylib`_pthread_body + 340
    frame #15: 0x00007fff6ed0e50d libsystem_pthread.dylib`_pthread_start + 377
    frame #16: 0x00007fff6ed0dbf9 libsystem_pthread.dylib`thread_start + 13

[d-led]

Now I think I can say for sure, there was an error inserting data into the DB. Added explicit waits for the cpprestsdk request tasks, catching and re-throwing errors.

Opened #18 as an open-ended follow-up reminder. Now your program should run, and in case of errors output something like:

async_api::insert failed: Failed to connect to any resolved endpoint -> Dropping 43 bytes
key num: 1
cost time 359
current time 0