2 * Bacula File Daemon heartbeat routines
3 * Listens for heartbeats coming from the SD
4 * If configured, sends heartbeats to Dir
6 * Kern Sibbald, May MMIII
12 Copyright (C) 2000-2003 Kern Sibbald and John Walker
14 This program is free software; you can redistribute it and/or
15 modify it under the terms of the GNU General Public License as
16 published by the Free Software Foundation; either version 2 of
17 the License, or (at your option) any later version.
19 This program is distributed in the hope that it will be useful,
20 but WITHOUT ANY WARRANTY; without even the implied warranty of
21 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
22 General Public License for more details.
24 You should have received a copy of the GNU General Public
25 License along with this program; if not, write to the Free
26 Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
35 /* pthread_kill() dies on Cygwin, so disable it */
36 #define pthread_kill(x, y)
37 /* Use shorter wait interval on Cygwin because no kill */
38 #define WAIT_INTERVAL 10
40 #else /* Unix systems */
41 #define WAIT_INTERVAL 60
45 * Listen on the SD socket for heartbeat signals.
46 * Send heartbeats to the Director every HB_TIME
49 static void *sd_heartbeat_thread(void *arg)
52 JCR *jcr = (JCR *)arg;
54 time_t last_heartbeat = time(NULL);
57 pthread_detach(pthread_self());
59 /* Get our own local copy */
60 sd = dup_bsock(jcr->store_bsock);
61 dir = dup_bsock(jcr->dir_bsock);
65 /* Hang reading the socket to the SD, and every time we get
66 * a heartbeat or we get a wait timeout (1 minute), we
67 * check to see if we need to send a heartbeat to the
70 for ( ; !is_bnet_stop(sd); ) {
71 n = bnet_wait_data_intr(sd, WAIT_INTERVAL);
72 if (me->heartbeat_interval) {
74 if (now-last_heartbeat >= me->heartbeat_interval) {
75 bnet_sig(dir, BNET_HEARTBEAT);
79 if (n == 1) { /* input waiting */
80 bnet_recv(sd); /* read it -- probably heartbeat from sd */
81 Dmsg1(100, "Got %d from SD\n", sd->msglen);
90 /* Startup the heartbeat thread -- see above */
91 void start_heartbeat_monitor(JCR *jcr)
94 pthread_create(&jcr->heartbeat_id, NULL, sd_heartbeat_thread, (void *)jcr);
97 /* Terminate the heartbeat thread. Used for both SD and DIR */
98 void stop_heartbeat_monitor(JCR *jcr)
101 /* Wait max 10 secs for heartbeat thread to start */
102 while (jcr->hb_bsock == NULL && cnt++ < 200) {
103 bmicrosleep(0, 50); /* avoid race */
107 jcr->hb_bsock->timed_out = 1; /* set timed_out to terminate read */
108 jcr->hb_bsock->terminated = 1; /* set to terminate read */
111 /* Wait max 100 secs for heartbeat thread to stop */
112 while (jcr->hb_bsock && cnt++ < 200) {
113 /* Naturally, Cygwin 1.3.20 craps out on the following */
114 pthread_kill(jcr->heartbeat_id, TIMEOUT_SIGNAL); /* make heartbeat thread go away */
120 * Thread for sending heartbeats to the Director when there
121 * is no SD monitoring needed -- e.g. restore and verify Vol
122 * both do their own read() on the SD socket.
124 static void *dir_heartbeat_thread(void *arg)
126 JCR *jcr = (JCR *)arg;
128 time_t last_heartbeat = time(NULL);
130 pthread_detach(pthread_self());
132 /* Get our own local copy */
133 dir = dup_bsock(jcr->dir_bsock);
137 for ( ; !is_bnet_stop(dir); ) {
141 next = now - last_heartbeat;
142 if (next >= me->heartbeat_interval) {
143 bnet_sig(dir, BNET_HEARTBEAT);
144 last_heartbeat = now;
146 bmicrosleep(next, 0);
149 jcr->hb_bsock = NULL;
153 * Same as above but we don't listen to the SD
155 void start_dir_heartbeat(JCR *jcr)
157 if (me->heartbeat_interval) {
158 pthread_create(&jcr->heartbeat_id, NULL, dir_heartbeat_thread, (void *)jcr);
162 void stop_dir_heartbeat(JCR *jcr)
164 if (me->heartbeat_interval) {
165 stop_heartbeat_monitor(jcr);