Rongsong Shen
331fa02423
Thread cluster_cfg by value through xcecd_config_t -> xcecd_context_t -> cc_ctx so it's populated before controller_connector_start() fires the first join, instead of after xcecd_start() returns. checking_connection() now retries a failed post-reconnect xcecli_join() on the next liveness probe instead of leaving needs_rejoin cleared on mere attempt.
112 lines
3.5 KiB
C
112 lines
3.5 KiB
C
#ifndef XCECD_H
|
|
#define XCECD_H
|
|
|
|
#include "controller_connector.h"
|
|
#include "hardware_monitor.h"
|
|
#include "job_information.h"
|
|
#include "node_information.h"
|
|
#include "process_manager.h"
|
|
#include "resource_sync.h"
|
|
#include "static_array.h"
|
|
#include "system_monitor.h"
|
|
#include "thread_pool.h"
|
|
#include "uv_ipc.h"
|
|
#include "xcecd_db.h"
|
|
#include "xcecd_dbus_manager.h"
|
|
#include "xcecd_worker_utils.h"
|
|
#include "xcectx.h"
|
|
#include "xcelog.h"
|
|
#include <stdbool.h>
|
|
#include <stdint.h>
|
|
#include <stdlib.h>
|
|
|
|
#ifdef _WIN32
|
|
#include <pdh.h>
|
|
#include <windows.h>
|
|
#include <winsock2.h>
|
|
typedef HANDLE process_handle_t;
|
|
#else
|
|
#include <arpa/inet.h>
|
|
#include <netdb.h>
|
|
#include <netinet/in.h>
|
|
#include <sys/types.h>
|
|
#include <unistd.h>
|
|
typedef pid_t process_handle_t;
|
|
#endif
|
|
|
|
#define REMOTE_PORT 6789
|
|
#define DEFAULT_PORT 0 // disable by default
|
|
#define BUFFER_SIZE 4096
|
|
#define RESTART_FLAG_FILE "/tmp/.xcecd.restarting"
|
|
|
|
typedef struct {
|
|
int thread_pool_size;
|
|
int resource_update_interval;
|
|
int job_status_update_interval;
|
|
uint16_t port;
|
|
const char* acct_file;
|
|
char* log_file;
|
|
const char* db_path;
|
|
const char* work_dir;
|
|
const char* domain_sock_file;
|
|
const char* remote_host;
|
|
const char* virutal_name;
|
|
int remote_port;
|
|
int join_interval;
|
|
int update_interval;
|
|
int domain_sock;
|
|
int debug_level;
|
|
/* HA candidate list parsed from --peers (component-failover-adoption,
|
|
* task 2.2, Req 1.4). Threaded through xcecd_init() -> xcecd_context_t
|
|
* -> controller_connector_init() so cc_ctx->cluster_cfg is populated
|
|
* before the first join, instead of main.c assigning it after
|
|
* xcecd_start() already ran. candidates_count == 0 means no --peers was
|
|
* given (single-host default). */
|
|
xce_cluster_config_t cluster_cfg;
|
|
} xcecd_config_t;
|
|
|
|
typedef struct {
|
|
thread_pool_t* thread_pool;
|
|
network_handler_t* network;
|
|
process_manager_t* process_manager;
|
|
system_monitor_t* system_monitor;
|
|
hardware_monitor_t* hardware_monitor;
|
|
resource_sync_t* resource_sync;
|
|
xcecd_worker_t* worker_list;
|
|
xcecd_db_t* db;
|
|
xce_ctx_t* xce_ctx;
|
|
bool running;
|
|
bool normal_restart;
|
|
char* remote_host;
|
|
int local_port;
|
|
int remote_port;
|
|
int update_interval;
|
|
int join_interval;
|
|
int domain_sock;
|
|
int debug_level;
|
|
char* work_dir;
|
|
char* log_dir;
|
|
char* db_path;
|
|
char* virtual_name;
|
|
/* Copied from xcecd_config_t by xcecd_init(); consumed by
|
|
* controller_connector_init() (xcecd_start()) to populate
|
|
* cc_ctx->cluster_cfg BEFORE controller_connector_start() fires the
|
|
* first join (component-failover-adoption, task 2.2, Req 1.4). */
|
|
xce_cluster_config_t cluster_cfg;
|
|
} xcecd_context_t;
|
|
|
|
extern xcecd_context_t xcecd_ctx;
|
|
extern hostattrs_t host_attrs;
|
|
|
|
int xcecd_config_init(xcecd_config_t* config);
|
|
int xcecd_init(const xcecd_config_t* config);
|
|
int xcecd_start(void);
|
|
void xcecd_stop(void);
|
|
void xcecd_cleanup(void);
|
|
void xcecd_process_command(void* arg);
|
|
const char* handle_network_request(const char* request, void* user_data);
|
|
bool xcecd_config_is_valid(const xcecd_config_t* config);
|
|
void xcecd_prepare_restart(void);
|
|
bool xcecd_check_restart_type(void);
|
|
|
|
#endif // XCECD_H
|