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;
75 * Get JobId and permissions from Director
78 Dmsg1(130, "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 fileset_md5 = get_memory(dir->msglen);
84 if (sscanf(dir->msg, jobcmd, &JobId, job, job_name, client_name,
85 &JobType, &level, fileset_name, &no_attributes,
86 &spool_attributes, fileset_md5) != 10) {
87 pm_strcpy(&jcr->errmsg, dir->msg);
88 bnet_fsend(dir, BAD_job, jcr->errmsg);
89 Emsg1(M_FATAL, 0, _("Bad Job Command from Director: %s\n"), jcr->errmsg);
91 free_memory(job_name);
92 free_memory(client_name);
93 free_memory(fileset_name);
94 free_memory(fileset_md5);
95 set_jcr_job_status(jcr, JS_ErrorTerminated);
99 jcr->VolSessionId = newVolSessionId();
100 jcr->VolSessionTime = VolSessionTime;
101 bstrncpy(jcr->Job, job, sizeof(jcr->Job));
102 unbash_spaces(job_name);
103 jcr->job_name = get_pool_memory(PM_NAME);
104 pm_strcpy(&jcr->job_name, job_name);
105 unbash_spaces(client_name);
106 jcr->client_name = get_pool_memory(PM_NAME);
107 pm_strcpy(&jcr->client_name, client_name);
108 unbash_spaces(fileset_name);
109 jcr->fileset_name = get_pool_memory(PM_NAME);
110 pm_strcpy(&jcr->fileset_name, fileset_name);
111 jcr->JobType = JobType;
112 jcr->JobLevel = level;
113 jcr->no_attributes = no_attributes;
114 jcr->spool_attributes = spool_attributes;
115 jcr->fileset_md5 = get_pool_memory(PM_NAME);
116 pm_strcpy(&jcr->fileset_md5, fileset_md5);
118 free_memory(job_name);
119 free_memory(client_name);
120 free_memory(fileset_name);
121 free_memory(fileset_md5);
123 /* Initialize FD start condition variable */
124 if ((errstat = pthread_cond_init(&jcr->job_start_wait, NULL)) != 0) {
125 Jmsg1(jcr, M_FATAL, 0, _("Unable to init job cond variable: ERR=%s\n"), strerror(errstat));
126 set_jcr_job_status(jcr, JS_ErrorTerminated);
129 jcr->authenticated = FALSE;
132 * Pass back an authorization key for the File daemon
134 #ifdef Old_way_not_so_good
135 gettimeofday(&tv, &tz);
136 srandom(tv.tv_usec + tv.tv_sec);
137 sprintf(auth_key, "%ld", (long)random());
139 make_session_key(auth_key, NULL, 1);
140 bnet_fsend(dir, OKjob, jcr->VolSessionId, jcr->VolSessionTime, auth_key);
141 Dmsg1(110, ">dird: %s", dir->msg);
142 jcr->sd_auth_key = bstrdup(auth_key);
143 memset(auth_key, 0, sizeof(auth_key));
146 * Wait for the device, media, and pool information
148 if (!use_device_cmd(jcr)) {
149 set_jcr_job_status(jcr, JS_ErrorTerminated);
150 memset(jcr->sd_auth_key, 0, strlen(jcr->sd_auth_key));
154 set_jcr_job_status(jcr, JS_WaitFD); /* wait for FD to connect */
155 dir_send_job_status(jcr);
157 gettimeofday(&tv, &tz);
158 timeout.tv_nsec = tv.tv_usec * 1000;
159 timeout.tv_sec = tv.tv_sec + 30 * 60; /* wait 30 minutes */
162 Dmsg1(200, "%s waiting on job_start_wait\n", jcr->Job);
163 /* Wait for the File daemon to contact us to start the Job,
164 * when he does, we will be released, unless the 30 minutes
168 for ( ;!job_canceled(jcr); ) {
169 errstat = pthread_cond_timedwait(&jcr->job_start_wait, &jcr->mutex, &timeout);
170 if (errstat == 0 || errstat == ETIMEDOUT) {
176 memset(jcr->sd_auth_key, 0, strlen(jcr->sd_auth_key));
178 if (jcr->authenticated && !job_canceled(jcr)) {
179 run_job(jcr); /* Run the job */
185 * This entry point is only called if we have a separate
186 * Storage Daemon Data port. Otherwise, the connection
187 * is made to the main port, and if it is a File daemon
188 * calling, handle_filed_connection() is called directly.
190 void connection_from_filed(void *arg)
192 BSOCK *fd = (BSOCK *)arg;
193 char job_name[MAX_NAME_LENGTH];
195 Dmsg0(110, "enter connection_from_filed\n");
196 if (bnet_recv(fd) <= 0) {
197 Emsg0(M_FATAL, 0, _("Unable to authenticate Client.\n"));
200 Dmsg1(100, "got: %s\n", fd->msg);
202 if (fd->msglen < 17 || fd->msglen > 17+127 ||
203 sscanf(fd->msg, "Hello Start Job %127s\n", job_name) != 1) {
204 Emsg1(M_FATAL, 0, _("Bad Hello from FD: %s\n"), fd->msg);
207 handle_filed_connection(fd, job_name);
211 void handle_filed_connection(BSOCK *fd, char *job_name)
215 if (!(jcr=get_jcr_by_full_name(job_name))) {
216 Jmsg1(NULL, M_FATAL, 0, _("Job name not found: %s\n"), job_name);
220 jcr->file_bsock = fd;
221 jcr->file_bsock->jcr = jcr;
223 Dmsg1(110, "Found Job %s\n", job_name);
225 if (jcr->authenticated) {
226 Pmsg2(000, "Hey!!!! JobId %u Job %s already authenticated.\n",
227 jcr->JobId, jcr->Job);
231 * Authenticate the File daemon
233 if (jcr->authenticated || !authenticate_filed(jcr)) {
234 Dmsg1(100, "Authentication failed Job %s\n", jcr->Job);
235 Jmsg(jcr, M_FATAL, 0, _("Unable to authenticate File daemon\n"));
237 jcr->authenticated = TRUE;
238 Dmsg1(110, "OK Authentication Job %s\n", jcr->Job);
242 if (!jcr->authenticated) {
243 set_jcr_job_status(jcr, JS_ErrorTerminated);
245 pthread_cond_signal(&jcr->job_start_wait); /* wake waiting job */
253 * Use Device command from Director
254 * He tells is what Device Name to use, the Media Type,
255 * the Pool Name, and the Pool Type.
257 * Ensure that the device exists and is opened, then store
258 * the media and pool info in the JCR.
260 static int use_device_cmd(JCR *jcr)
262 POOLMEM *dev_name, *media_type, *pool_name, *pool_type;
263 BSOCK *dir = jcr->dir_bsock;
266 if (bnet_recv(dir) <= 0) {
267 Jmsg0(jcr, M_FATAL, 0, _("No Device from Director\n"));
271 Dmsg1(120, "Use device: %s", dir->msg);
272 dev_name = get_memory(dir->msglen);
273 media_type = get_memory(dir->msglen);
274 pool_name = get_memory(dir->msglen);
275 pool_type = get_memory(dir->msglen);
276 if (sscanf(dir->msg, use_device, dev_name, media_type, pool_name, pool_type) == 4) {
277 unbash_spaces(dev_name);
278 unbash_spaces(media_type);
279 unbash_spaces(pool_name);
280 unbash_spaces(pool_type);
283 while ((device=(DEVRES *)GetNextRes(R_DEVICE, (RES *)device))) {
284 /* Find resource, and make sure we were able to open it */
285 if (strcmp(device->hdr.name, dev_name) == 0 && device->dev) {
286 Dmsg1(120, "Found device %s\n", device->hdr.name);
287 jcr->pool_name = get_memory(strlen(pool_name) + 1);
288 strcpy(jcr->pool_name, pool_name);
289 jcr->pool_type = get_memory(strlen(pool_type) + 1);
290 strcpy(jcr->pool_type, pool_type);
291 jcr->media_type = get_memory(strlen(media_type) + 1);
292 strcpy(jcr->media_type, media_type);
293 jcr->dev_name = get_memory(strlen(dev_name) + 1);
294 strcpy(jcr->dev_name, dev_name);
295 jcr->device = device;
296 Dmsg4(120, use_device, dev_name, media_type, pool_name, pool_type);
297 free_memory(dev_name);
298 free_memory(media_type);
299 free_memory(pool_name);
300 free_memory(pool_type);
302 return bnet_fsend(dir, OK_device);
307 unbash_spaces(dir->msg);
308 pm_strcpy(&jcr->errmsg, dir->msg);
309 Jmsg(jcr, M_INFO, 0, _("Failed command: %s\n"), jcr->errmsg);
311 Jmsg(jcr, M_FATAL, 0, _("\n"
312 " Device \"%s\" requested by Dir not found in SD Device resources.\n"),
314 bnet_fsend(dir, NO_device, dev_name);
316 unbash_spaces(dir->msg);
317 pm_strcpy(&jcr->errmsg, dir->msg);
319 Jmsg(jcr, M_INFO, 0, _("Failed command: %s\n"), jcr->errmsg);
321 Jmsg(jcr, M_FATAL, 0, _("Bad Use Device command: %s\n"), jcr->errmsg);
322 bnet_fsend(dir, BAD_use, jcr->errmsg);
325 free_memory(dev_name);
326 free_memory(media_type);
327 free_memory(pool_name);
328 free_memory(pool_type);
329 return 0; /* ERROR return */
333 * Destroy the Job Control Record and associated
334 * resources (sockets).
336 void stored_free_jcr(JCR *jcr)
338 if (jcr->file_bsock) {
339 bnet_close(jcr->file_bsock);
340 jcr->file_bsock = NULL;
342 if (jcr->pool_name) {
343 free_memory(jcr->pool_name);
345 if (jcr->pool_type) {
346 free_memory(jcr->pool_type);
348 if (jcr->media_type) {
349 free_memory(jcr->media_type);
352 free_memory(jcr->dev_name);
355 free_pool_memory(jcr->job_name);
357 if (jcr->client_name) {
358 free_memory(jcr->client_name);
359 jcr->client_name = NULL;
361 if (jcr->fileset_name) {
362 free_memory(jcr->fileset_name);
364 if (jcr->fileset_md5) {
365 free_memory(jcr->fileset_md5);
371 if (jcr->RestoreBootstrap) {
372 unlink(jcr->RestoreBootstrap);
373 free_pool_memory(jcr->RestoreBootstrap);
374 jcr->RestoreBootstrap = NULL;
376 if (jcr->next_dev && jcr->prev_dev) {
377 Emsg0(M_FATAL, 0, _("In free_jcr(), but still attached to device!!!!\n"));