tyler.durden
/
icinga2
mirror of https://github.com/Icinga/icinga2.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

StartUnixWorker(): don't exit() on fork() failure 

... but let the caller handle the failure.

Not to stop working completely just because of fork() failure during a reload.
This commit is contained in:
Alexander A. Klimov 2021-01-14 13:40:18 +01:00
parent 5f548c8f89
commit 931b9307ae
1 changed files with 37 additions and 23 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

60
lib/cli/daemoncommand.cpp
View File

@ -446,7 +446,7 @@ static void NotifyWatchdog()
* @param closeConsoleLog Whether to close the console log after config loading * @param closeConsoleLog Whether to close the console log after config loading
* @param stderrFile Where to log errors * @param stderrFile Where to log errors
* *
* @return The worker's PID on success, -1 on failure (if the worker couldn't load its config) * @return The worker's PID on success, -1 on fork(2) failure, -2 if the worker couldn't load its config
*/ */
static pid_t StartUnixWorker(const std::vector<std::string>& configs, bool closeConsoleLog = false, const String& stderrFile = String()) static pid_t StartUnixWorker(const std::vector<std::string>& configs, bool closeConsoleLog = false, const String& stderrFile = String())
{ {
@ -472,7 +472,17 @@ static pid_t StartUnixWorker(const std::vector<std::string>& configs, bool close
case -1: case -1:
Log(LogCritical, "cli") Log(LogCritical, "cli")
<< "fork() failed with error code " << errno << ", \"" << Utility::FormatErrorNumber(errno) << "\""; << "fork() failed with error code " << errno << ", \"" << Utility::FormatErrorNumber(errno) << "\"";
exit(EXIT_FAILURE);
try {
Application::InitializeBase();
} catch (const std::exception& ex) {
Log(LogCritical, "cli")
<< "Failed to re-initialize thread pool after forking (parent): " << DiagnosticInformation(ex);
exit(EXIT_FAILURE);
}
(void)sigprocmask(SIG_UNBLOCK, &l_UnixWorkerSignals, nullptr);
return -1;
case 0: case 0:
try { try {
@ -550,7 +560,7 @@ static pid_t StartUnixWorker(const std::vector<std::string>& configs, bool close
NotifyWatchdog(); NotifyWatchdog();
#endif /* HAVE_SYSTEMD */ #endif /* HAVE_SYSTEMD */
} }
pid = -1; pid = -2;
break; break;
default: default:
Utility::Sleep(0.2); Utility::Sleep(0.2);
@ -710,7 +720,7 @@ int DaemonCommand::Run(const po::variables_map& vm, const std::vector<std::strin
// The PID of the current seamless worker // The PID of the current seamless worker
pid_t currentWorker = StartUnixWorker(configs, closeConsoleLog, errorLog); pid_t currentWorker = StartUnixWorker(configs, closeConsoleLog, errorLog);
if (currentWorker == -1) { if (currentWorker < 0) {
return EXIT_FAILURE; return EXIT_FAILURE;
} }
@ -769,31 +779,35 @@ int DaemonCommand::Run(const po::variables_map& vm, const std::vector<std::strin
pid_t nextWorker = StartUnixWorker(configs); pid_t nextWorker = StartUnixWorker(configs);
if (nextWorker == -1) { switch (nextWorker) {
Log(LogCritical, "Application", "Found error in config: reloading aborted"); case -1:
} else { break;
Log(LogInformation, "Application") case -2:
<< "Reload done, old process shutting down. Child process with PID '" << nextWorker << "' is taking over."; Log(LogCritical, "Application", "Found error in config: reloading aborted");
break;
default:
Log(LogInformation, "Application")
<< "Reload done, old process shutting down. Child process with PID '" << nextWorker << "' is taking over.";
(void)kill(currentWorker, SIGTERM); (void)kill(currentWorker, SIGTERM);
{ {
double start = Utility::GetTime(); double start = Utility::GetTime();
while (waitpid(currentWorker, nullptr, 0) == -1 && errno == EINTR) { while (waitpid(currentWorker, nullptr, 0) == -1 && errno == EINTR) {
#ifdef HAVE_SYSTEMD #ifdef HAVE_SYSTEMD
NotifyWatchdog(); NotifyWatchdog();
#endif /* HAVE_SYSTEMD */ #endif /* HAVE_SYSTEMD */
}
Log(LogNotice, "cli")
<< "Waited for " << Utility::FormatDuration(Utility::GetTime() - start) << " on old process to exit.";
} }
Log(LogNotice, "cli") // Old instance shut down, allow the new one to continue working beyond config validation
<< "Waited for " << Utility::FormatDuration(Utility::GetTime() - start) << " on old process to exit."; (void)kill(nextWorker, SIGUSR2);
}
// Old instance shut down, allow the new one to continue working beyond config validation currentWorker = nextWorker;
(void)kill(nextWorker, SIGUSR2);
currentWorker = nextWorker;
} }
#ifdef HAVE_SYSTEMD #ifdef HAVE_SYSTEMD