2 functions to track and manage processes
4 Copyright (C) Ronnie Sahlberg 2012
6 This program is free software; you can redistribute it and/or modify
7 it under the terms of the GNU General Public License as published by
8 the Free Software Foundation; either version 3 of the License, or
9 (at your option) any later version.
11 This program is distributed in the hope that it will be useful,
12 but WITHOUT ANY WARRANTY; without even the implied warranty of
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 GNU General Public License for more details.
16 You should have received a copy of the GNU General Public License
17 along with this program; if not, see <http://www.gnu.org/licenses/>.
21 #include "system/wait.h"
22 #include "system/network.h"
27 #include "lib/util/debug.h"
28 #include "lib/util/time.h"
30 #include "ctdb_private.h"
31 #include "ctdb_client.h"
33 #include "common/rb_tree.h"
34 #include "common/system.h"
35 #include "common/common.h"
36 #include "common/logging.h"
38 void ctdb_track_child(struct ctdb_context *ctdb, pid_t pid)
42 /* Only CTDB main daemon should track child processes */
43 if (getpid() != ctdb->ctdbd_pid) {
47 process = talloc_asprintf(ctdb->child_processes, "process:%d", (int)pid);
48 trbt_insert32(ctdb->child_processes, pid, process);
52 * This function forks a child process and drops the realtime
53 * scheduler for the child process.
55 pid_t ctdb_fork(struct ctdb_context *ctdb)
58 struct timeval before;
61 before = timeval_current();
66 (__location__ " fork() failed (%s)\n", strerror(errno)));
70 /* Close the Unix Domain socket and the TCP socket.
71 * This ensures that none of the child processes will
72 * look like the main daemon when it is not running.
73 * tevent needs to be stopped before closing sockets.
75 if (ctdb->ev != NULL) {
76 talloc_free(ctdb->ev);
79 if (ctdb->daemon.sd != -1) {
80 close(ctdb->daemon.sd);
83 if (ctdb->methods != NULL && ctdb->methods->shutdown != NULL) {
84 ctdb->methods->shutdown(ctdb);
87 /* The child does not need to be realtime */
88 if (ctdb->do_setsched) {
91 ctdb->can_send_controls = false;
96 delta_t = timeval_elapsed(&before);
98 DEBUG(DEBUG_WARNING, ("fork() took %lf seconds\n", delta_t));
101 ctdb_track_child(ctdb, pid);
108 pid_t ctdb_vfork_exec(TALLOC_CTX *mem_ctx, struct ctdb_context *ctdb,
109 const char *helper, int helper_argc,
110 const char **helper_argv)
113 struct timeval before;
118 argv = talloc_array(mem_ctx, char *, helper_argc + 1);
120 DEBUG(DEBUG_ERR, ("Memory allocation error\n"));
124 argv[0] = discard_const(helper);
125 for (i=0; i<helper_argc; i++) {
126 argv[i+1] = discard_const(helper_argv[i]);
129 before = timeval_current();
133 DEBUG(DEBUG_ERR, ("vfork() failed (%s)\n", strerror(errno)));
142 delta_t = timeval_elapsed(&before);
144 DEBUG(DEBUG_WARNING, ("vfork() took %lf seconds\n", delta_t));
147 ctdb_track_child(ctdb, pid);
151 static void ctdb_sigchld_handler(struct tevent_context *ev,
152 struct tevent_signal *te, int signum, int count,
156 struct ctdb_context *ctdb = talloc_get_type(private_data, struct ctdb_context);
161 pid = waitpid(-1, &status, WNOHANG);
163 DEBUG(DEBUG_ERR, (__location__ " waitpid() returned error. errno:%d\n", errno));
169 if (getpid() != ctdb->ctdbd_pid) {
173 process = trbt_lookup32(ctdb->child_processes, pid);
174 if (process == NULL) {
175 DEBUG(DEBUG_ERR,("Got SIGCHLD from pid:%d we didn not spawn with ctdb_fork\n", pid));
178 DEBUG(DEBUG_DEBUG, ("SIGCHLD from %d %s\n", (int)pid, process));
179 talloc_free(process);
185 struct tevent_signal *
186 ctdb_init_sigchld(struct ctdb_context *ctdb)
188 struct tevent_signal *se;
190 ctdb->child_processes = trbt_create(ctdb, 0);
192 se = tevent_add_signal(ctdb->ev, ctdb, SIGCHLD, 0, ctdb_sigchld_handler, ctdb);
197 ctdb_kill(struct ctdb_context *ctdb, pid_t pid, int signum)
202 return kill(pid, signum);
205 if (getpid() != ctdb->ctdbd_pid) {
206 return kill(pid, signum);
209 process = trbt_lookup32(ctdb->child_processes, pid);
210 if (process == NULL) {
211 DEBUG(DEBUG_ERR,("ctdb_kill: trying to kill(%d, %d) a process that does not exist\n", pid, signum));
215 return kill(pid, signum);