[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

[Cluster-devel] Cluster Project branch, master, updated. cluster-2.99.05-67-g7c2bac8



This is an automated email from the git hooks/post-receive script. It was
generated because a ref change was pushed to the repository containing
the project "Cluster Project".

http://sources.redhat.com/git/gitweb.cgi?p=cluster.git;a=commitdiff;h=7c2bac805f3a60b320a9c95d7be6bcfdf9302bfe

The branch, master has been updated
       via  7c2bac805f3a60b320a9c95d7be6bcfdf9302bfe (commit)
      from  0dcc5e557008df4d51aef891ebfb998a34785f04 (commit)

Those revisions listed above that are new to this repository have
not appeared on any other notification email; so we list those
revisions in full, below.

- Log -----------------------------------------------------------------
commit 7c2bac805f3a60b320a9c95d7be6bcfdf9302bfe
Author: David Teigland <teigland redhat com>
Date:   Fri Jul 11 13:39:38 2008 -0500

    groupd: sync daemon setup/structure with others
    
    This daemon is defunct, but will continue to exist for back compat cases,
    so put code in sync with other daemons.
    
    Signed-off-by: David Teigland <teigland redhat com>

-----------------------------------------------------------------------

Summary of changes:
 group/daemon/Makefile      |    3 +-
 group/daemon/cman.c        |   65 +++++++----
 group/daemon/cpg.c         |    6 +-
 group/daemon/gd_internal.h |   53 ++++-----
 group/daemon/groupd.h      |    1 +
 group/daemon/main.c        |  291 +++++++++++++++++++++-----------------------
 group/tool/main.c          |    2 -
 7 files changed, 212 insertions(+), 209 deletions(-)

diff --git a/group/daemon/Makefile b/group/daemon/Makefile
index 90931f3..df75bc6 100644
--- a/group/daemon/Makefile
+++ b/group/daemon/Makefile
@@ -16,10 +16,11 @@ OBJS=	app.o \
 	joinleave.o \
 	main.o
 
-CFLAGS += -I${cmanincdir} -I${openaisincdir}
+CFLAGS += -I${ccsincdir} -I${cmanincdir} -I${openaisincdir}
 CFLAGS += -I$(S) -I$(S)/../include/ -I$(S)/../lib/
 CFLAGS += -I${incdir}
 
+LDFLAGS += -L${ccslibdir} -lccs
 LDFLAGS += -L${cmanlibdir} -L${openaislibdir} -L${libdir} -lcman -lcpg
 
 
diff --git a/group/daemon/cman.c b/group/daemon/cman.c
index 272b155..12dcd1c 100644
--- a/group/daemon/cman.c
+++ b/group/daemon/cman.c
@@ -117,49 +117,67 @@ static void cman_callback(cman_handle_t h, void *private, int reason, int arg)
 	}
 }
 
-static void close_cman(int ci)
-{
-	log_debug("cluster is down, exiting");
-	exit(1);
-}
-
-static void process_cman(int ci)
+void process_cman(int ci)
 {
 	int rv;
+
 	rv = cman_dispatch(ch, CMAN_DISPATCH_ALL);
 	if (rv == -1 && errno == EHOSTDOWN)
-		close_cman(ci);
+		cluster_dead(0);
 }
 
 int setup_cman(void)
 {
 	cman_node_t node;
 	int rv, fd;
+	int init = 0, active = 0;
 
+ retry_init:
 	ch = cman_init(NULL);
 	if (!ch) {
-		log_print("cman_init error %p %d", ch, errno);
+		if (init++ < 2) {
+			sleep(1);
+			goto retry_init;
+		}
+		log_print("cman_init error %d", errno);
+		return -ENOTCONN;
+	}
+
+ retry_active:
+	rv = cman_is_active(ch);
+	if (!rv) {
+		if (active++ < 2) {
+			sleep(1);
+			goto retry_active;
+		}
+		log_print("cman_is_active error %d", errno);
+		cman_finish(ch);
 		return -ENOTCONN;
 	}
 
 	ch_admin = cman_admin_init(NULL);
 	if (!ch_admin) {
-		log_print("cman_admin_init error %p %d", ch_admin, errno);
-		rv = -ENOTCONN;
-		goto fail1;
+		log_print("cman_admin_init error %d", errno);
+		cman_finish(ch);
+		return -ENOTCONN;
 	}
 
 	rv = cman_start_notification(ch, cman_callback);
 	if (rv < 0) {
 		log_print("cman_start_notification error %d %d", rv, errno);
-		goto fail2;
+		cman_finish(ch);
+		cman_finish(ch_admin);
+		return rv;
 	}
 
 	memset(&node, 0, sizeof(node));
 	rv = cman_get_node(ch, CMAN_NODEID_US, &node);
 	if (rv < 0) {
 		log_print("cman_get_node us error %d %d", rv, errno);
-		goto fail3;
+		cman_stop_notification(ch);
+		cman_finish(ch);
+		cman_finish(ch_admin);
+		return rv;
 	}
 
 	cman_node_count = 0;
@@ -167,7 +185,10 @@ int setup_cman(void)
 	rv = cman_get_nodes(ch, MAX_NODES, &cman_node_count, cman_nodes);
 	if (rv < 0) {
 		log_print("cman_get_nodes error %d %d", rv, errno);
-		goto fail3;
+		cman_stop_notification(ch);
+		cman_finish(ch);
+		cman_finish(ch_admin);
+		return rv;
 	}
 
 	cman_quorate = cman_is_quorate(ch);
@@ -180,15 +201,13 @@ int setup_cman(void)
 		  our_nodeid, our_name, cman_quorate);
 
 	fd = cman_get_fd(ch);
-	client_add(fd, process_cman, close_cman);
-	return 0;
 
- fail3:
-	cman_stop_notification(ch);
- fail2:
-	cman_finish(ch_admin);
- fail1:
+	return fd;
+}
+
+void close_cman(void)
+{
 	cman_finish(ch);
-	return rv;
+	cman_finish(ch_admin);
 }
 
diff --git a/group/daemon/cpg.c b/group/daemon/cpg.c
index aa9d9db..e062f7d 100644
--- a/group/daemon/cpg.c
+++ b/group/daemon/cpg.c
@@ -325,7 +325,7 @@ void deliver_cb(cpg_handle_t handle, struct cpg_name *group_name,
 
 		g = find_group_level(name, msg->ms_level);
 		if (!g) {
-			if (groupd_debug_verbose > 1) {
+			if (daemon_debug_verbose > 1) {
 				log_print("%d:%s RECV len %d %s from %d, "
 					  "no group",
 				  	  msg->ms_level, name, data_len,
@@ -347,7 +347,7 @@ void deliver_cb(cpg_handle_t handle, struct cpg_name *group_name,
 		}
 	}
 
-	if (groupd_debug_verbose > 1)
+	if (daemon_debug_verbose > 1)
 		log_group(g, "RECV len %d %s from %d", data_len,
 			  msg_type(msg->ms_type), nodeid);
 
@@ -669,7 +669,7 @@ static int _send_message(cpg_handle_t h, group_t *g, void *buf, int len)
 
 int send_message_groupd(group_t *g, void *buf, int len, int type)
 {
-	if (groupd_debug_verbose > 1)
+	if (daemon_debug_verbose > 1)
 		log_group(g, "SEND len %d %s", len, msg_type(type));
 
 	return _send_message(groupd_handle, g, buf, len);
diff --git a/group/daemon/gd_internal.h b/group/daemon/gd_internal.h
index b842905..3c8bddf 100644
--- a/group/daemon/gd_internal.h
+++ b/group/daemon/gd_internal.h
@@ -28,37 +28,39 @@
 #include "groupd.h"
 #include "libgroup.h"
 
-#define MAX_NAMELEN		(32)	/* should match libgroup.h */
-#define MAX_LEVELS		(4)
-#define MAX_NODES		(256)
-#define MAXARGS			(16)
-#define MAXCON			(16)
-#define NALLOC			(16)
-#define DUMP_SIZE		(1024 * 1024)
-
-extern char *prog_name;
-extern int groupd_debug_opt;
-extern int groupd_debug_verbose;
-extern char groupd_debug_buf[256];
-extern char dump_buf[DUMP_SIZE];
+#define MAX_NAMELEN		32	/* should match libgroup.h */
+#define MAX_LEVELS		4
+#define MAX_NODES		128
+
+extern int daemon_debug_opt;
+extern int daemon_debug_verbose;
+extern int daemon_quit;
+extern int cman_quorate;
+extern int our_nodeid;
+extern char *our_name;
+extern char daemon_debug_buf[256];
+extern char dump_buf[GROUPD_DUMP_SIZE];
 extern int dump_point;
 extern int dump_wrap;
+extern struct list_head gd_groups;
+extern struct list_head gd_levels[MAX_LEVELS];
+extern uint32_t gd_event_nr;
 
-extern void groupd_dump_save(void);
+void daemon_dump_save(void);
 
 #define log_debug(fmt, args...) \
 do { \
-	snprintf(groupd_debug_buf, 255, "%ld " fmt "\n", time(NULL), ##args); \
-	if (groupd_debug_opt) fprintf(stderr, "%s", groupd_debug_buf); \
-	groupd_dump_save(); \
+	snprintf(daemon_debug_buf, 255, "%ld " fmt "\n", time(NULL), ##args); \
+	if (daemon_debug_opt) fprintf(stderr, "%s", daemon_debug_buf); \
+	daemon_dump_save(); \
 } while (0)
 
 #define log_group(g, fmt, args...) \
 do { \
-	snprintf(groupd_debug_buf, 255, "%ld %d:%s " fmt "\n", time(NULL), \
+	snprintf(daemon_debug_buf, 255, "%ld %d:%s " fmt "\n", time(NULL), \
 		 (g)->level, (g)->name, ##args); \
-	if (groupd_debug_opt) fprintf(stderr, "%s", groupd_debug_buf); \
-	groupd_dump_save(); \
+	if (daemon_debug_opt) fprintf(stderr, "%s", daemon_debug_buf); \
+	daemon_dump_save(); \
 } while (0)
 
 #define log_print(fmt, args...) \
@@ -73,7 +75,6 @@ do { \
 	syslog(LOG_ERR, fmt, ##args); \
 } while (0)
 
-
 #define ASSERT(x) \
 do { \
 	if (!(x)) { \
@@ -89,13 +90,6 @@ do { \
 #define FALSE (0)
 #endif
 
-extern struct list_head		gd_groups;
-extern struct list_head		gd_levels[MAX_LEVELS];
-extern uint32_t			gd_event_nr;
-extern int			cman_quorate;
-extern int			our_nodeid;
-extern char			*our_name;
-
 struct group;
 struct app;
 struct event;
@@ -261,9 +255,12 @@ void app_terminate(app_t *a);
 void app_deliver(app_t *a, struct save_msg *save);
 int client_add(int fd, void (*workfn)(int ci), void (*deadfn)(int ci));
 void client_dead(int ci);
+void cluster_dead(int ci);
 
 /* cman.c */
 int setup_cman(void);
+void close_cman(void);
+void process_cman(int ci);
 int kill_cman(int nodeid);
 int set_cman_dirty(void);
 
diff --git a/group/daemon/groupd.h b/group/daemon/groupd.h
index e120995..69bd1fc 100644
--- a/group/daemon/groupd.h
+++ b/group/daemon/groupd.h
@@ -8,5 +8,6 @@
 #define GROUPD_MSGLEN		(2200) /* should be enough to permit
 					  group_send() of up to 2048 bytes
 					  of data plus header info */
+#define GROUPD_DUMP_SIZE	(1024 * 1024)
 
 #endif
diff --git a/group/daemon/main.c b/group/daemon/main.c
index 8aa3225..11a7ed8 100644
--- a/group/daemon/main.c
+++ b/group/daemon/main.c
@@ -2,25 +2,19 @@
 #include <time.h>
 
 #include "gd_internal.h"
+#include "ccs.h"
 
-#define OPTION_STRING			"DhVv"
-#define LOCKFILE_NAME			"/var/run/groupd.pid"
-#define LOG_FILE				"/var/log/groupd.log"
+#define LOCKFILE_NAME	"/var/run/groupd.pid"
+#define CLIENT_NALLOC	32
 
 extern struct list_head recovery_sets;
 
-struct list_head	gd_groups;
-struct list_head	gd_levels[MAX_LEVELS];
-uint32_t		gd_event_nr;
-char			*our_name;
-int			our_nodeid;
-int			cman_quorate;
-
 static int client_maxi;
 static int client_size = 0;
 static struct client *client = NULL;
-static struct pollfd *pollfd;
+static struct pollfd *pollfd = NULL;
 static char last_action[16];
+static int ccs_handle;
 
 struct client {
 	int fd;
@@ -30,6 +24,23 @@ struct client {
 	void *deadfn;
 };
 
+static int do_read(int fd, void *buf, size_t count)
+{
+	int rv, off = 0;
+
+	while (off < count) {
+		rv = read(fd, buf + off, count - off);
+		if (rv == 0)
+			return -1;
+		if (rv == -1 && errno == EINTR)
+			continue;
+		if (rv == -1)
+			return -1;
+		off += rv;
+	}
+	return 0;
+}
+
 static int do_write(int fd, void *buf, size_t count)
 {
 	int rv, off = 0;
@@ -51,23 +62,28 @@ static int do_write(int fd, void *buf, size_t count)
 	return 0;
 }
 
-static int do_read(int fd, void *buf, size_t count)
+int setup_ccs(void)
 {
-	int rv, off = 0;
+	int i = 0, cd;
 
-	while (off < count) {
-		rv = read(fd, buf + off, count - off);
-		if (rv == 0)
-			return -1;
-		if (rv == -1 && errno == EINTR)
-			continue;
-		if (rv == -1)
-			return -1;
-		off += rv;
+	while ((cd = ccs_connect()) < 0) {
+		sleep(1);
+		if (++i > 9 && !(i % 10))
+			log_print("connect to ccs error %d, "
+				  "check cluster status", cd);
 	}
+
+	ccs_handle = cd;
+
 	return 0;
 }
 
+void close_ccs(void)
+{
+	ccs_disconnect(ccs_handle);
+}
+
+
 /* Look for any instances of gfs or dlm in the kernel, if we find any, it
    means they're uncontrolled by us (via gfs_controld/dlm_controld/groupd).
    We need to be rebooted to clear out this uncontrolled kernel state.  Most
@@ -173,7 +189,8 @@ void app_deliver(app_t *a, struct save_msg *save)
 	int rv;
 
 	rv = snprintf(buf, sizeof(buf), "deliver %s %d %d",
-		      a->g->name, save->nodeid, save->msg_len - sizeof(msg_t));
+		      a->g->name, save->nodeid,
+		      (int)(save->msg_len - sizeof(msg_t)));
 
 	log_group(a->g, "deliver to app: %s", buf);
 
@@ -251,25 +268,7 @@ void app_finish(app_t *a)
 	app_action(a, buf);
 }
 
-#if 0
-static void make_args(char *buf, int *argc, char **argv, char sep)
-{
-	char *p = buf;
-	int i;
-
-	argv[0] = p;
-
-	for (i = 1; i < MAXARGS; i++) {
-		p = strchr(buf, sep);
-		if (!p)
-			break;
-		*p = '\0';
-		argv[i] = p + 1;
-		buf = p + 1;
-	}
-	*argc = i;
-}
-#endif
+#define MAXARGS 16
 
 static char *get_args(char *buf, int *argc, char **argv, char sep, int want)
 {
@@ -369,20 +368,21 @@ static void client_alloc(void)
 {
 	int i;
 
-	if (!client)
-		client = malloc(NALLOC * sizeof(struct client));
-	else {
-		client = realloc(client, (client_size + NALLOC) *
+	if (!client) {
+		client = malloc(CLIENT_NALLOC * sizeof(struct client));
+		pollfd = malloc(CLIENT_NALLOC * sizeof(struct pollfd));
+	} else {
+		client = realloc(client, (client_size + CLIENT_NALLOC) *
 					 sizeof(struct client));
-		pollfd = realloc(pollfd, (client_size + NALLOC) *
+		pollfd = realloc(pollfd, (client_size + CLIENT_NALLOC) *
 					 sizeof(struct pollfd));
 		if (!pollfd)
 			log_print("can't alloc for pollfd");
 	}
-	if (!client)
+	if (!client || !pollfd)
 		log_print("can't alloc for client array");
 
-	for (i = client_size; i < client_size + NALLOC; i++) {
+	for (i = client_size; i < client_size + CLIENT_NALLOC; i++) {
 		client[i].workfn = NULL;
 		client[i].deadfn = NULL;
 		client[i].fd = -1;
@@ -391,7 +391,7 @@ static void client_alloc(void)
 		pollfd[i].fd = -1;
 		pollfd[i].revents = 0;
 	}
-	client_size += NALLOC;
+	client_size += CLIENT_NALLOC;
 }
 
 void client_dead(int ci)
@@ -429,6 +429,11 @@ int client_add(int fd, void (*workfn)(int ci), void (*deadfn)(int ci))
 	goto again;
 }
 
+static void sigterm_handler(int sig)
+{
+	daemon_quit = 1;
+}
+
 static void do_setup(int ci, int argc, char **argv)
 {
 	log_debug("setup %s %s", argv[1], argv[2]);
@@ -544,7 +549,7 @@ static int do_dump(int fd)
 	int len;
 
 	if (dump_wrap) {
-		len = DUMP_SIZE - dump_point;
+		len = GROUPD_DUMP_SIZE - dump_point;
 		do_write(fd, dump_buf + dump_point, len);
 		len = dump_point;
 	} else
@@ -602,8 +607,6 @@ static void process_connection(int ci)
 
 	rv = do_read(client[ci].fd, buf, GROUPD_MSGLEN);
 	if (rv < 0) {
-		log_print("client %d fd %d read error %d %d", ci,
-			  client[ci].fd, rv, errno);
 		client_dead(ci);
 		return;
 	}
@@ -681,10 +684,10 @@ static void process_listener(int ci)
 	
 	i = client_add(fd, process_connection, NULL);
 
-	log_debug("client connection %d", i);
+	log_debug("client connection %d fd %d", i, fd);
 }
 
-static int setup_listener(void)
+static int setup_listener(char *sock_path)
 {
 	struct sockaddr_un addr;
 	socklen_t addrlen;
@@ -700,7 +703,7 @@ static int setup_listener(void)
 
 	memset(&addr, 0, sizeof(addr));
 	addr.sun_family = AF_LOCAL;
-	strcpy(&addr.sun_path[1], GROUPD_SOCK_PATH);
+	strcpy(&addr.sun_path[1], sock_path);
 	addrlen = sizeof(sa_family_t) + strlen(addr.sun_path+1) + 1;
 
 	rv = bind(s, (struct sockaddr *) &addr, addrlen);
@@ -716,51 +719,73 @@ static int setup_listener(void)
 		close(s);
 		return rv;
 	}
+	return s;
+}
 
-	client_add(s, process_listener, NULL);
-
-	return 0;
+void cluster_dead(int ci)
+{
+	log_print("cluster is down, exiting");
+	daemon_quit = 1;
 }
 
-static int loop(void)
+static void loop(void)
 {
-	int rv, i, timeout = -1;
+	int poll_timeout = -1;	
+	int rv, i;
 	void (*workfn) (int ci);
 	void (*deadfn) (int ci);
 
-	rv = setup_listener();
+	rv = setup_listener(GROUPD_SOCK_PATH);
 	if (rv < 0)
-		return rv;
+		goto out;
+	client_add(rv, process_listener, NULL);
 
 	rv = setup_cman();
 	if (rv < 0)
-		return rv;
+		goto out;
+	client_add(rv, process_cman, cluster_dead);
+
+	rv = setup_ccs();
+	if (rv < 0)
+		goto out;
 
 	rv = check_uncontrolled_groups();
 	if (rv < 0)
-		return rv;
+		goto out;
 
 	rv = setup_cpg();
 	if (rv < 0)
-		return rv;
+		goto out;
 
-	while (1) {
-		rv = poll(pollfd, client_maxi + 1, timeout);
-		if (rv < 0)
-			log_debug("poll error %d %d", rv, errno);
+	for (;;) {
+		rv = poll(pollfd, client_maxi + 1, poll_timeout);
+		if (rv == -1 && errno == EINTR) {
+			if (daemon_quit && list_empty(&gd_groups))
+				goto out;
+			daemon_quit = 0;
+			continue;
+		}
+		if (rv < 0) {
+			log_print("poll errno %d", errno);
+			goto out;
+		}
 
 		for (i = 0; i <= client_maxi; i++) {
 			if (client[i].fd < 0)
 				continue;
+			if (pollfd[i].revents & POLLIN) {
+				workfn = client[i].workfn;
+				workfn(i);
+			}
 			if (pollfd[i].revents & (POLLERR | POLLHUP | POLLNVAL)) {
 				deadfn = client[i].deadfn;
 				deadfn(i);
-			} else if (pollfd[i].revents & POLLIN) {
-				workfn = client[i].workfn;
-				workfn(i);
 			}
 		}
 
+		if (daemon_quit)
+			break;
+
 		/* process_apps() returns non-zero if there may be
 		   more work to do */
 
@@ -769,8 +794,12 @@ static int loop(void)
 			rv += process_apps();
 		} while (rv);
 	}
+ out:
+	close_ccs();
+	close_cman();
 
-	return 0;
+	if (!list_empty(&gd_groups))
+		log_print("groups abandoned");
 }
 
 static void lockfile(void)
@@ -815,34 +844,11 @@ static void lockfile(void)
 	}
 }
 
-void daemonize(void)
-{
-	pid_t pid = fork();
-	if (pid < 0) {
-		perror("main: cannot fork");
-		exit(EXIT_FAILURE);
-	}
-	if (pid)
-		exit(EXIT_SUCCESS);
-	setsid();
-	if(chdir("/") < 0) {
-		perror("main: unable to chdir");
-		exit(EXIT_FAILURE);
-	}
-	umask(0);
-	close(0);
-	close(1);
-	close(2);
-	openlog("groupd", LOG_PID, LOG_DAEMON);
-
-	lockfile();
-}
-
 static void print_usage(void)
 {
 	printf("Usage:\n");
 	printf("\n");
-	printf("%s [options]\n", prog_name);
+	printf("groupd [options]\n");
 	printf("\n");
 	printf("Options:\n");
 	printf("\n");
@@ -851,9 +857,11 @@ static void print_usage(void)
 	printf("  -V	       Print program version information, then exit\n");
 }
 
-static void decode_arguments(int argc, char **argv)
+#define OPTION_STRING "DhVv"
+
+static void read_arguments(int argc, char **argv)
 {
-	int cont = TRUE;
+	int cont = 1;
 	int optchar;
 
 	while (cont) {
@@ -862,7 +870,7 @@ static void decode_arguments(int argc, char **argv)
 		switch (optchar) {
 
 		case 'D':
-			groupd_debug_opt = 1;
+			daemon_debug_opt = 1;
 			break;
 
 		case 'h':
@@ -871,7 +879,7 @@ static void decode_arguments(int argc, char **argv)
 			break;
 
 		case 'v':
-			groupd_debug_verbose++;
+			daemon_debug_verbose++;
 			break;
 
 		case 'V':
@@ -899,7 +907,7 @@ static void decode_arguments(int argc, char **argv)
 	}
 }
 
-void set_oom_adj(int val)
+static void set_oom_adj(int val)
 {
 	FILE *fp;
 
@@ -911,7 +919,7 @@ void set_oom_adj(int val)
 	fclose(fp);
 }
 
-void set_scheduler(void)
+static void set_scheduler(void)
 {
 	struct sched_param sched_param;
 	int rv;
@@ -929,38 +937,8 @@ void set_scheduler(void)
 	}
 }
 
-void bail_with_log(int sig)
-{
-	int fd;
-	time_t now;
-
-	unlink(LOG_FILE);
-	fd = creat(LOG_FILE, S_IRUSR | S_IWUSR);
-	if (fd > 0) {
-		char now_ascii[32];
-
-		do_dump(fd);
-		memset(now_ascii, 0, sizeof(now_ascii));
-		time(&now);
-		sprintf(now_ascii, "%ld", now);
-		if (write(fd, now_ascii, strlen(now_ascii)) < 0) {
-			perror("Unable to write");
-			exit(1);
-		}
-		if (write(fd, " groupd segfault log follows:\n", 30) < 0) {
-			perror("Unable to write");
-			exit(1);
-		}
-		close(fd);
-	} else
-		perror(LOG_FILE);
-	if (sig == SIGSEGV)
-		exit(0);
-}
-
 int main(int argc, char *argv[])
 {
-	prog_name = argv[0];
 	int i;
 
 	INIT_LIST_HEAD(&recovery_sets);
@@ -968,45 +946,54 @@ int main(int argc, char *argv[])
 	for (i = 0; i < MAX_LEVELS; i++)
 		INIT_LIST_HEAD(&gd_levels[i]);
 
-	decode_arguments(argc, argv);
+	read_arguments(argc, argv);
 
-	signal(SIGSEGV, bail_with_log);
-	signal(SIGUSR1, bail_with_log);
+	lockfile();
 
-	if (!groupd_debug_opt)
-		daemonize();
+	if (!daemon_debug_opt) {
+		if (daemon(0, 0) < 0) {
+			perror("daemon error");
+			exit(EXIT_FAILURE);
+		}
+	}
+	openlog("groupd", LOG_PID, LOG_DAEMON);
+	signal(SIGTERM, sigterm_handler);
 
 	set_scheduler();
 	set_oom_adj(-16);
 
-	pollfd = malloc(NALLOC * sizeof(struct pollfd));
-	if (!pollfd)
-		return -1;
+	loop();
 
-	return loop();
+	return 0;
 }
 
-void groupd_dump_save(void)
+void daemon_dump_save(void)
 {
 	int len, i;
 
-	len = strlen(groupd_debug_buf);
+	len = strlen(daemon_debug_buf);
 
 	for (i = 0; i < len; i++) {
-		dump_buf[dump_point++] = groupd_debug_buf[i];
+		dump_buf[dump_point++] = daemon_debug_buf[i];
 
-		if (dump_point == DUMP_SIZE) {
+		if (dump_point == GROUPD_DUMP_SIZE) {
 			dump_point = 0;
 			dump_wrap = 1;
 		}
 	}
 }
 
-char *prog_name;
-int groupd_debug_opt;
-int groupd_debug_verbose;
-char groupd_debug_buf[256];
-char dump_buf[DUMP_SIZE];
+int daemon_debug_opt;
+int daemon_debug_verbose;
+int daemon_quit;
+int cman_quorate;
+int our_nodeid;
+char *our_name;
+char daemon_debug_buf[256];
+char dump_buf[GROUPD_DUMP_SIZE];
 int dump_point;
 int dump_wrap;
+struct list_head gd_groups;
+struct list_head gd_levels[MAX_LEVELS];
+uint32_t gd_event_nr;
 
diff --git a/group/tool/main.c b/group/tool/main.c
index 33e7bd1..bbdf52e 100644
--- a/group/tool/main.c
+++ b/group/tool/main.c
@@ -23,8 +23,6 @@
 #define MAX_MG				128
 #define MAX_GROUPS			128
 
-#define GROUPD_DUMP_SIZE		(1024 * 1024)
-
 #define OP_LIST				1
 #define OP_DUMP				2
 #define OP_LOG				3


hooks/post-receive
--
Cluster Project


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]