2 * Job control and execution for Storage Daemon
8 Copyright (C) 2000-2003 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 NoAttr=%d SpoolAttr=%d FileSetMD5=%127s\n";
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[] = "3914 Device \"%s\" not in SD Device resources.\n";
48 static char BAD_use[] = "3913 Bad use command: %s\n";
49 static char BAD_job[] = "3915 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, *fileset_md5;
69 int JobType, level, spool_attributes, no_attributes;
72 struct timespec timeout;
76 * Get JobId and permissions from Director
79 Dmsg1(130, "Job_cmd: %s\n", dir->msg);
80 job = get_memory(dir->msglen);
81 job_name = get_memory(dir->msglen);
82 client_name = get_memory(dir->msglen);
83 fileset_name = get_memory(dir->msglen);
84 fileset_md5 = get_memory(dir->msglen);
85 if (sscanf(dir->msg, jobcmd, &JobId, job, job_name, client_name,
86 &JobType, &level, fileset_name, &no_attributes,
87 &spool_attributes, fileset_md5) != 10) {
88 pm_strcpy(&jcr->errmsg, dir->msg);
89 bnet_fsend(dir, BAD_job, jcr->errmsg);
90 Emsg1(M_FATAL, 0, _("Bad Job Command from Director: %s\n"), jcr->errmsg);
92 free_memory(job_name);
93 free_memory(client_name);
94 free_memory(fileset_name);
95 free_memory(fileset_md5);
96 set_jcr_job_status(jcr, JS_ErrorTerminated);
100 * Since this job could be rescheduled, we
101 * check to see if we have it already. If so
102 * free the old jcr and use the new one.
104 ojcr = get_jcr_by_full_name(job);
105 if (ojcr && !ojcr->authenticated) {
106 Dmsg2(100, "Found ojcr=0x%x Job %s\n", (unsigned)ojcr, job);
110 jcr->VolSessionId = newVolSessionId();
111 jcr->VolSessionTime = VolSessionTime;
112 bstrncpy(jcr->Job, job, sizeof(jcr->Job));
113 unbash_spaces(job_name);
114 jcr->job_name = get_pool_memory(PM_NAME);
115 pm_strcpy(&jcr->job_name, job_name);
116 unbash_spaces(client_name);
117 jcr->client_name = get_pool_memory(PM_NAME);
118 pm_strcpy(&jcr->client_name, client_name);
119 unbash_spaces(fileset_name);
120 jcr->fileset_name = get_pool_memory(PM_NAME);
121 pm_strcpy(&jcr->fileset_name, fileset_name);
122 jcr->JobType = JobType;
123 jcr->JobLevel = level;
124 jcr->no_attributes = no_attributes;
125 jcr->spool_attributes = spool_attributes;
126 jcr->fileset_md5 = get_pool_memory(PM_NAME);
127 pm_strcpy(&jcr->fileset_md5, fileset_md5);
129 free_memory(job_name);
130 free_memory(client_name);
131 free_memory(fileset_name);
132 free_memory(fileset_md5);
134 /* Initialize FD start condition variable */
135 if ((errstat = pthread_cond_init(&jcr->job_start_wait, NULL)) != 0) {
136 Jmsg1(jcr, M_FATAL, 0, _("Unable to init job cond variable: ERR=%s\n"), strerror(errstat));
137 set_jcr_job_status(jcr, JS_ErrorTerminated);
140 jcr->authenticated = FALSE;
143 * Pass back an authorization key for the File daemon
145 make_session_key(auth_key, NULL, 1);
146 bnet_fsend(dir, OKjob, jcr->VolSessionId, jcr->VolSessionTime, auth_key);
147 Dmsg1(110, ">dird: %s", dir->msg);
148 jcr->sd_auth_key = bstrdup(auth_key);
149 memset(auth_key, 0, sizeof(auth_key));
152 * Wait for the device, media, and pool information
154 if (!use_device_cmd(jcr)) {
155 set_jcr_job_status(jcr, JS_ErrorTerminated);
156 memset(jcr->sd_auth_key, 0, strlen(jcr->sd_auth_key));
160 set_jcr_job_status(jcr, JS_WaitFD); /* wait for FD to connect */
161 dir_send_job_status(jcr);
163 gettimeofday(&tv, &tz);
164 timeout.tv_nsec = tv.tv_usec * 1000;
165 timeout.tv_sec = tv.tv_sec + 30 * 60; /* wait 30 minutes */
168 Dmsg1(100, "%s waiting on job_start_wait\n", jcr->Job);
169 /* Wait for the File daemon to contact us to start the Job,
170 * when he does, we will be released, unless the 30 minutes
174 for ( ;!job_canceled(jcr); ) {
175 errstat = pthread_cond_timedwait(&jcr->job_start_wait, &jcr->mutex, &timeout);
176 if (errstat == 0 || errstat == ETIMEDOUT) {
182 memset(jcr->sd_auth_key, 0, strlen(jcr->sd_auth_key));
184 if (jcr->authenticated && !job_canceled(jcr)) {
185 Dmsg1(100, "Running job %s\n", jcr->Job);
186 run_job(jcr); /* Run the job */
192 * After receiving a connection (in job.c) if it is
193 * from the File daemon, this routine is called.
195 void handle_filed_connection(BSOCK *fd, char *job_name)
199 bmicrosleep(0, 50000); /* wait 50 millisecs */
200 if (!(jcr=get_jcr_by_full_name(job_name))) {
201 Jmsg1(NULL, M_FATAL, 0, _("Job name not found: %s\n"), job_name);
205 jcr->file_bsock = fd;
206 jcr->file_bsock->jcr = jcr;
208 Dmsg1(110, "Found Job %s\n", job_name);
210 if (jcr->authenticated) {
211 Pmsg2(000, "Hey!!!! JobId %u Job %s already authenticated.\n",
212 jcr->JobId, jcr->Job);
216 * Authenticate the File daemon
218 if (jcr->authenticated || !authenticate_filed(jcr)) {
219 Dmsg1(100, "Authentication failed Job %s\n", jcr->Job);
220 Jmsg(jcr, M_FATAL, 0, _("Unable to authenticate File daemon\n"));
222 jcr->authenticated = TRUE;
223 Dmsg1(110, "OK Authentication Job %s\n", jcr->Job);
227 if (!jcr->authenticated) {
228 set_jcr_job_status(jcr, JS_ErrorTerminated);
230 pthread_cond_signal(&jcr->job_start_wait); /* wake waiting job */
238 * Use Device command from Director
239 * He tells is what Device Name to use, the Media Type,
240 * the Pool Name, and the Pool Type.
242 * Ensure that the device exists and is opened, then store
243 * the media and pool info in the JCR.
245 static int use_device_cmd(JCR *jcr)
247 POOLMEM *dev_name, *media_type, *pool_name, *pool_type;
248 BSOCK *dir = jcr->dir_bsock;
251 if (bnet_recv(dir) <= 0) {
252 Jmsg0(jcr, M_FATAL, 0, _("No Device from Director\n"));
256 Dmsg1(120, "Use device: %s", dir->msg);
257 dev_name = get_memory(dir->msglen);
258 media_type = get_memory(dir->msglen);
259 pool_name = get_memory(dir->msglen);
260 pool_type = get_memory(dir->msglen);
261 if (sscanf(dir->msg, use_device, dev_name, media_type, pool_name, pool_type) == 4) {
262 unbash_spaces(dev_name);
263 unbash_spaces(media_type);
264 unbash_spaces(pool_name);
265 unbash_spaces(pool_type);
268 while ((device=(DEVRES *)GetNextRes(R_DEVICE, (RES *)device))) {
269 /* Find resource, and make sure we were able to open it */
270 if (strcmp(device->hdr.name, dev_name) == 0 && device->dev) {
271 Dmsg1(120, "Found device %s\n", device->hdr.name);
272 jcr->pool_name = get_memory(strlen(pool_name) + 1);
273 strcpy(jcr->pool_name, pool_name);
274 jcr->pool_type = get_memory(strlen(pool_type) + 1);
275 strcpy(jcr->pool_type, pool_type);
276 jcr->media_type = get_memory(strlen(media_type) + 1);
277 strcpy(jcr->media_type, media_type);
278 jcr->dev_name = get_memory(strlen(dev_name) + 1);
279 strcpy(jcr->dev_name, dev_name);
280 jcr->device = device;
281 Dmsg4(120, use_device, dev_name, media_type, pool_name, pool_type);
282 free_memory(dev_name);
283 free_memory(media_type);
284 free_memory(pool_name);
285 free_memory(pool_type);
287 return bnet_fsend(dir, OK_device);
292 unbash_spaces(dir->msg);
293 pm_strcpy(&jcr->errmsg, dir->msg);
294 Jmsg(jcr, M_INFO, 0, _("Failed command: %s\n"), jcr->errmsg);
296 Jmsg(jcr, M_FATAL, 0, _("\n"
297 " Device \"%s\" requested by Dir not found in SD Device resources.\n"),
299 bnet_fsend(dir, NO_device, dev_name);
301 unbash_spaces(dir->msg);
302 pm_strcpy(&jcr->errmsg, dir->msg);
304 Jmsg(jcr, M_INFO, 0, _("Failed command: %s\n"), jcr->errmsg);
306 Jmsg(jcr, M_FATAL, 0, _("Bad Use Device command: %s\n"), jcr->errmsg);
307 bnet_fsend(dir, BAD_use, jcr->errmsg);
310 free_memory(dev_name);
311 free_memory(media_type);
312 free_memory(pool_name);
313 free_memory(pool_type);
314 return 0; /* ERROR return */
318 * Destroy the Job Control Record and associated
319 * resources (sockets).
321 void stored_free_jcr(JCR *jcr)
323 if (jcr->file_bsock) {
324 bnet_close(jcr->file_bsock);
325 jcr->file_bsock = NULL;
327 if (jcr->pool_name) {
328 free_memory(jcr->pool_name);
330 if (jcr->pool_type) {
331 free_memory(jcr->pool_type);
333 if (jcr->media_type) {
334 free_memory(jcr->media_type);
337 free_memory(jcr->dev_name);
340 free_pool_memory(jcr->job_name);
342 if (jcr->client_name) {
343 free_memory(jcr->client_name);
344 jcr->client_name = NULL;
346 if (jcr->fileset_name) {
347 free_memory(jcr->fileset_name);
349 if (jcr->fileset_md5) {
350 free_memory(jcr->fileset_md5);
356 if (jcr->RestoreBootstrap) {
357 unlink(jcr->RestoreBootstrap);
358 free_pool_memory(jcr->RestoreBootstrap);
359 jcr->RestoreBootstrap = NULL;
361 if (jcr->next_dev && jcr->prev_dev) {
362 Emsg0(M_FATAL, 0, _("In free_jcr(), but still attached to device!!!!\n"));
364 pthread_cond_destroy(&jcr->job_start_wait);