2 * Job control and execution for Storage Daemon
8 Copyright (C) 2000, 2001, 2002 Kern Sibbald and John Walker
10 This program is free software; you can redistribute it and/or
11 modify it under the terms of the GNU General Public License as
12 published by the Free Software Foundation; either version 2 of
13 the License, or (at your option) any later version.
15 This program is distributed in the hope that it will be useful,
16 but WITHOUT ANY WARRANTY; without even the implied warranty of
17 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
18 General Public License for more details.
20 You should have received a copy of the GNU General Public
21 License along with this program; if not, write to the Free
22 Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
30 /* Imported variables */
31 extern uint32_t VolSessionTime;
33 /* Imported functions */
34 extern uint32_t newVolSessionId();
36 /* Forward referenced functions */
37 static int use_device_cmd(JCR *jcr);
39 /* Requests from the Director daemon */
40 static char jobcmd[] = "JobId=%d job=%127s job_name=%127s client_name=%127s \
41 type=%d level=%d FileSet=%127s Allow=";
42 static char use_device[] = "use device=%s media_type=%s pool_name=%s pool_type=%s\n";
44 /* Responses sent to Director daemon */
45 static char OKjob[] = "3000 OK Job SDid=%u SDtime=%u Authorization=%s\n";
46 static char OK_device[] = "3000 OK use device\n";
47 static char NO_device[] = "3904 Device %s not available\n";
48 static char BAD_use[] = "3903 Bad use command: %s\n";
49 static char BAD_job[] = "3905 Bad Job command: %s\n";
54 * Director requests us to start a job
55 * Basic tasks done here:
56 * - We pickup the JobId to be run from the Director.
57 * - We pickup the device, media, and pool from the Director
58 * - Wait for a connection from the File Daemon (FD)
59 * - Accept commands from the FD (i.e. run the job)
60 * - Return when the connection is terminated or
67 BSOCK *dir = jcr->dir_bsock;
68 POOLMEM *job_name, *client_name, *job, *fileset_name;
72 struct timespec timeout;
75 * Get JobId and permissions from Director
78 Dmsg1(30, "Job_cmd: %s\n", dir->msg);
79 job = get_memory(dir->msglen);
80 job_name = get_memory(dir->msglen);
81 client_name = get_memory(dir->msglen);
82 fileset_name = get_memory(dir->msglen);
83 if (sscanf(dir->msg, jobcmd, &JobId, job, job_name, client_name,
84 &JobType, &level, fileset_name) != 7) {
85 bnet_fsend(dir, BAD_job, dir->msg);
86 Emsg1(M_FATAL, 0, _("Bad Job Command from Director: %s\n"), dir->msg);
88 free_memory(job_name);
89 free_memory(client_name);
90 free_memory(fileset_name);
91 jcr->JobStatus = JS_ErrorTerminated;
95 jcr->VolSessionId = newVolSessionId();
96 jcr->VolSessionTime = VolSessionTime;
97 strcpy(jcr->Job, job);
98 unbash_spaces(job_name);
99 jcr->job_name = get_memory(strlen(job_name) + 1);
100 strcpy(jcr->job_name, job_name);
101 unbash_spaces(client_name);
102 jcr->client_name = get_memory(strlen(client_name) + 1);
103 strcpy(jcr->client_name, client_name);
104 unbash_spaces(fileset_name);
105 jcr->fileset_name = get_memory(strlen(fileset_name) + 1);
106 strcpy(jcr->fileset_name, fileset_name);
107 jcr->JobType = JobType;
108 jcr->JobLevel = level;
110 free_memory(job_name);
111 free_memory(client_name);
112 free_memory(fileset_name);
114 /* Initialize FD start condition variable */
115 if ((errstat = pthread_cond_init(&jcr->job_start_wait, NULL)) != 0) {
116 Emsg1(M_FATAL, 0, _("Unable to init job cond variable: ERR=%s\n"), strerror(errstat));
117 jcr->JobStatus = JS_ErrorTerminated;
120 jcr->authenticated = FALSE;
123 * Pass back an authorization key for the File daemon
125 gettimeofday(&tv, &tz);
126 srandom(tv.tv_usec + tv.tv_sec);
127 sprintf(auth_key, "%ld", (long)random());
128 bnet_fsend(dir, OKjob, jcr->VolSessionId, jcr->VolSessionTime, auth_key);
129 Dmsg1(10, ">dird: %s", dir->msg);
130 jcr->sd_auth_key = bstrdup(auth_key);
133 * Wait for the device, media, and pool information
135 if (!use_device_cmd(jcr)) {
136 jcr->JobStatus = JS_ErrorTerminated;
140 jcr->JobStatus = JS_WaitFD; /* wait for FD to connect */
141 dir_send_job_status(jcr);
143 gettimeofday(&tv, &tz);
144 timeout.tv_nsec = tv.tv_usec * 1000;
145 timeout.tv_sec = tv.tv_sec + 30 * 60; /* wait 30 minutes */
148 Dmsg1(200, "%s waiting on job_start_wait\n", jcr->Job);
150 for ( ;!job_cancelled(jcr); ) {
151 errstat = pthread_cond_timedwait(&jcr->job_start_wait, &jcr->mutex, &timeout);
152 if (errstat == 0 || errstat == ETIMEDOUT) {
158 if (jcr->authenticated && !job_cancelled(jcr)) {
159 run_job(jcr); /* Run the job */
165 * This entry point is only called if we have a separate
166 * Storage Daemon Data port. Otherwise, the connection
167 * is made to the main port, and if it is a File daemon
168 * calling, handl_filed_connection() is called directly.
170 void connection_from_filed(void *arg)
172 BSOCK *fd = (BSOCK *)arg;
173 char job_name[MAX_NAME_LENGTH];
175 Dmsg0(200, "enter connection_from_filed\n");
176 if (bnet_recv(fd) <= 0) {
177 Emsg0(M_FATAL, 0, _("Unable to authenticate Client.\n"));
180 Dmsg1(100, "got: %s\n", fd->msg);
182 if (sscanf(fd->msg, "Hello Start Job %127s\n", job_name) != 1) {
183 Emsg1(M_FATAL, 0, _("Authentication failure: %s\n"), fd->msg);
186 handle_filed_connection(fd, job_name);
190 void handle_filed_connection(BSOCK *fd, char *job_name)
194 if (!(jcr=get_jcr_by_full_name(job_name))) {
195 Emsg1(M_FATAL, 0, _("Job name not found: %s\n"), job_name);
199 jcr->file_bsock = fd;
201 Dmsg1(100, "Found Job %s\n", job_name);
203 if (jcr->authenticated) {
204 Pmsg2(000, "Hey!!!! JobId %d Job %s already authenticated.\n",
205 jcr->JobId, jcr->Job);
209 * Authenticate the File daemon
211 if (jcr->authenticated || !authenticate_filed(jcr)) {
212 Dmsg1(100, "Authentication failed Job %s\n", jcr->Job);
213 Jmsg(jcr, M_FATAL, 0, _("Unable to authenticate File daemon\n"));
215 jcr->authenticated = TRUE;
216 Dmsg1(100, "OK Authentication Job %s\n", jcr->Job);
220 if (!jcr->authenticated) {
221 jcr->JobStatus = JS_ErrorTerminated;
223 pthread_cond_signal(&jcr->job_start_wait); /* wake waiting job */
231 * Use Device command from Director
232 * He tells is what Device Name to use, the Media Type,
233 * the Pool Name, and the Pool Type.
235 * Ensure that the device exists and is opened, then store
236 * the media and pool info in the JCR.
238 static int use_device_cmd(JCR *jcr)
240 POOLMEM *dev_name, *media_type, *pool_name, *pool_type;
241 BSOCK *dir = jcr->dir_bsock;
244 if (bnet_recv(dir) <= 0) {
245 Emsg0(M_FATAL, 0, "No Device from Director\n");
249 Dmsg1(20, "Use device: %s", dir->msg);
250 dev_name = get_memory(dir->msglen);
251 media_type = get_memory(dir->msglen);
252 pool_name = get_memory(dir->msglen);
253 pool_type = get_memory(dir->msglen);
254 if (sscanf(dir->msg, use_device, dev_name, media_type, pool_name, pool_type) == 4) {
255 unbash_spaces(dev_name);
256 unbash_spaces(media_type);
257 unbash_spaces(pool_name);
258 unbash_spaces(pool_type);
261 while ((device=(DEVRES *)GetNextRes(R_DEVICE, (RES *)device))) {
262 /* Find resource, and make sure we were able to open it */
263 if (strcmp(device->hdr.name, dev_name) == 0 && device->dev) {
264 Dmsg1(20, "Found device %s\n", device->hdr.name);
265 jcr->pool_name = get_memory(strlen(pool_name) + 1);
266 strcpy(jcr->pool_name, pool_name);
267 jcr->pool_type = get_memory(strlen(pool_type) + 1);
268 strcpy(jcr->pool_type, pool_type);
269 jcr->media_type = get_memory(strlen(media_type) + 1);
270 strcpy(jcr->media_type, media_type);
271 jcr->dev_name = get_memory(strlen(dev_name) + 1);
272 strcpy(jcr->dev_name, dev_name);
273 jcr->device = device;
274 Dmsg4(20, use_device, dev_name, media_type, pool_name, pool_type);
275 free_memory(dev_name);
276 free_memory(media_type);
277 free_memory(pool_name);
278 free_memory(pool_type);
280 return bnet_fsend(dir, OK_device);
284 Jmsg(jcr, M_FATAL, 0, _("Requested device %s not found. Cannot continue.\n"),
286 bnet_fsend(dir, NO_device, dev_name);
288 Jmsg(jcr, M_FATAL, 0, _("store<dir: Bad Use Device command: %s\n"), dir->msg);
289 bnet_fsend(dir, BAD_use, dir->msg);
292 free_memory(dev_name);
293 free_memory(media_type);
294 free_memory(pool_name);
295 free_memory(pool_type);
296 return 0; /* ERROR return */
300 * Destroy the Job Control Record and associated
301 * resources (sockets).
303 void stored_free_jcr(JCR *jcr)
305 if (jcr->file_bsock) {
306 bnet_close(jcr->file_bsock);
308 if (jcr->pool_name) {
309 free_memory(jcr->pool_name);
311 if (jcr->pool_type) {
312 free_memory(jcr->pool_type);
314 if (jcr->media_type) {
315 free_memory(jcr->media_type);
318 free_memory(jcr->dev_name);
321 free_pool_memory(jcr->job_name);
323 if (jcr->client_name) {
324 free_memory(jcr->client_name);
325 jcr->client_name = NULL;
327 if (jcr->fileset_name) {
328 free_memory(jcr->fileset_name);
334 if (jcr->RestoreBootstrap) {
335 unlink(jcr->RestoreBootstrap);
336 free_pool_memory(jcr->RestoreBootstrap);
337 jcr->RestoreBootstrap = NULL;
339 if (jcr->next_dev && jcr->prev_dev) {
340 Emsg0(M_FATAL, 0, _("In free_jcr(), but still attached to device!!!!\n"));