2 * Job control and execution for Storage Daemon
10 Copyright (C) 2000-2005 Kern Sibbald
12 This program is free software; you can redistribute it and/or
13 modify it under the terms of the GNU General Public License as
14 published by the Free Software Foundation; either version 2 of
15 the License, or (at your option) any later version.
17 This program is distributed in the hope that it will be useful,
18 but WITHOUT ANY WARRANTY; without even the implied warranty of
19 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
20 General Public License for more details.
22 You should have received a copy of the GNU General Public
23 License along with this program; if not, write to the Free
24 Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
32 /* Imported variables */
33 extern uint32_t VolSessionTime;
35 /* Imported functions */
36 extern uint32_t newVolSessionId();
38 /* Forward referenced functions */
39 static bool use_device_cmd(JCR *jcr);
41 /* Requests from the Director daemon */
42 static char jobcmd[] = "JobId=%d job=%127s job_name=%127s client_name=%127s "
43 "type=%d level=%d FileSet=%127s NoAttr=%d SpoolAttr=%d FileSetMD5=%127s "
44 "SpoolData=%d WritePartAfterJob=%d";
45 static char use_device[] = "use device=%127s media_type=%127s pool_name=%127s pool_type=%127s\n";
46 static char use_devices[] = "use devices=%127s media_type=%127s pool_name=%127s pool_type=%127s\n";
48 /* Responses sent to Director daemon */
49 static char OKjob[] = "3000 OK Job SDid=%u SDtime=%u Authorization=%s\n";
50 static char OK_device[] = "3000 OK use device\n";
51 static char NO_device[] = "3914 Device \"%s\" not in SD Device resources.\n";
52 static char BAD_use[] = "3913 Bad use command: %s\n";
53 static char BAD_job[] = "3915 Bad Job command: %s\n";
56 * Director requests us to start a job
57 * Basic tasks done here:
58 * - We pickup the JobId to be run from the Director.
59 * - We pickup the device, media, and pool from the Director
60 * - Wait for a connection from the File Daemon (FD)
61 * - Accept commands from the FD (i.e. run the job)
62 * - Return when the connection is terminated or
65 bool job_cmd(JCR *jcr)
69 BSOCK *dir = jcr->dir_bsock;
70 POOL_MEM job_name, client_name, job, fileset_name, fileset_md5;
71 int JobType, level, spool_attributes, no_attributes, spool_data, write_part_after_job;
74 struct timespec timeout;
78 * Get JobId and permissions from Director
80 Dmsg1(100, "Job_cmd: %s\n", dir->msg);
81 if (sscanf(dir->msg, jobcmd, &JobId, job.c_str(), job_name.c_str(),
83 &JobType, &level, fileset_name.c_str(), &no_attributes,
84 &spool_attributes, fileset_md5.c_str(), &spool_data, &write_part_after_job) != 12) {
85 pm_strcpy(jcr->errmsg, dir->msg);
86 bnet_fsend(dir, BAD_job, jcr->errmsg);
87 Emsg1(M_FATAL, 0, _("Bad Job Command from Director: %s\n"), jcr->errmsg);
88 set_jcr_job_status(jcr, JS_ErrorTerminated);
92 * Since this job could be rescheduled, we
93 * check to see if we have it already. If so
94 * free the old jcr and use the new one.
96 ojcr = get_jcr_by_full_name(job.c_str());
97 if (ojcr && !ojcr->authenticated) {
98 Dmsg2(100, "Found ojcr=0x%x Job %s\n", (unsigned)(long)ojcr, job.c_str());
102 jcr->VolSessionId = newVolSessionId();
103 jcr->VolSessionTime = VolSessionTime;
104 bstrncpy(jcr->Job, job, sizeof(jcr->Job));
105 unbash_spaces(job_name);
106 jcr->job_name = get_pool_memory(PM_NAME);
107 pm_strcpy(jcr->job_name, job_name);
108 unbash_spaces(client_name);
109 jcr->client_name = get_pool_memory(PM_NAME);
110 pm_strcpy(jcr->client_name, client_name);
111 unbash_spaces(fileset_name);
112 jcr->fileset_name = get_pool_memory(PM_NAME);
113 pm_strcpy(jcr->fileset_name, fileset_name);
114 jcr->JobType = JobType;
115 jcr->JobLevel = level;
116 jcr->no_attributes = no_attributes;
117 jcr->spool_attributes = spool_attributes;
118 jcr->spool_data = spool_data;
119 jcr->write_part_after_job = write_part_after_job;
120 jcr->fileset_md5 = get_pool_memory(PM_NAME);
121 pm_strcpy(jcr->fileset_md5, fileset_md5);
123 jcr->authenticated = false;
126 * Pass back an authorization key for the File daemon
128 make_session_key(auth_key, NULL, 1);
129 bnet_fsend(dir, OKjob, jcr->VolSessionId, jcr->VolSessionTime, auth_key);
130 Dmsg1(110, ">dird: %s", dir->msg);
131 jcr->sd_auth_key = bstrdup(auth_key);
132 memset(auth_key, 0, sizeof(auth_key));
135 * Wait for the device, media, and pool information
137 if (!use_device_cmd(jcr)) {
138 set_jcr_job_status(jcr, JS_ErrorTerminated);
139 memset(jcr->sd_auth_key, 0, strlen(jcr->sd_auth_key));
143 /* The following jobs don't need the FD */
144 switch (jcr->JobType) {
148 jcr->authenticated = true;
153 set_jcr_job_status(jcr, JS_WaitFD); /* wait for FD to connect */
154 dir_send_job_status(jcr);
156 gettimeofday(&tv, &tz);
157 timeout.tv_nsec = tv.tv_usec * 1000;
158 timeout.tv_sec = tv.tv_sec + 30 * 60; /* wait 30 minutes */
160 Dmsg1(100, "%s waiting on FD to contact SD\n", jcr->Job);
162 * Wait for the File daemon to contact us to start the Job,
163 * when he does, we will be released, unless the 30 minutes
167 for ( ;!job_canceled(jcr); ) {
168 errstat = pthread_cond_timedwait(&jcr->job_start_wait, &jcr->mutex, &timeout);
169 if (errstat == 0 || errstat == ETIMEDOUT) {
175 memset(jcr->sd_auth_key, 0, strlen(jcr->sd_auth_key));
177 if (jcr->authenticated && !job_canceled(jcr)) {
178 Dmsg1(100, "Running job %s\n", jcr->Job);
179 run_job(jcr); /* Run the job */
185 * After receiving a connection (in job.c) if it is
186 * from the File daemon, this routine is called.
188 void handle_filed_connection(BSOCK *fd, char *job_name)
192 bmicrosleep(0, 50000); /* wait 50 millisecs */
193 if (!(jcr=get_jcr_by_full_name(job_name))) {
194 Jmsg1(NULL, M_FATAL, 0, _("Job name not found: %s\n"), job_name);
195 Dmsg1(100, "Job name not found: %s\n", job_name);
199 jcr->file_bsock = fd;
200 jcr->file_bsock->jcr = jcr;
202 Dmsg1(110, "Found Job %s\n", job_name);
204 if (jcr->authenticated) {
205 Jmsg2(jcr, M_FATAL, 0, "Hey!!!! JobId %u Job %s already authenticated.\n",
206 jcr->JobId, jcr->Job);
212 * Authenticate the File daemon
214 if (jcr->authenticated || !authenticate_filed(jcr)) {
215 Dmsg1(100, "Authentication failed Job %s\n", jcr->Job);
216 Jmsg(jcr, M_FATAL, 0, _("Unable to authenticate File daemon\n"));
218 jcr->authenticated = true;
219 Dmsg1(110, "OK Authentication Job %s\n", jcr->Job);
223 if (!jcr->authenticated) {
224 set_jcr_job_status(jcr, JS_ErrorTerminated);
226 pthread_cond_signal(&jcr->job_start_wait); /* wake waiting job */
234 * Use Device command from Director
235 * He tells is what Device Name to use, the Media Type,
236 * the Pool Name, and the Pool Type.
238 * Ensure that the device exists and is opened, then store
239 * the media and pool info in the JCR.
241 static bool use_device_cmd(JCR *jcr)
243 POOL_MEM dev_name, media_type, pool_name, pool_type;
244 BSOCK *dir = jcr->dir_bsock;
250 if (bnet_recv(dir) <= 0) {
251 Jmsg0(jcr, M_FATAL, 0, _("No Device from Director\n"));
255 Dmsg1(120, "Use device: %s", dir->msg);
257 * If there are multiple devices, the director sends us
258 * use_devices (note plurel) until the last one, at which
259 * time, it sends us a use_device command (note singlular)
260 * so we stop looking after getting the use_device.
262 ok = sscanf(dir->msg, use_device, dev_name.c_str(), media_type.c_str(),
263 pool_name.c_str(), pool_type.c_str()) == 4;
265 quit = true; /* got last device */
267 ok = sscanf(dir->msg, use_devices, dev_name.c_str(), media_type.c_str(),
268 pool_name.c_str(), pool_type.c_str()) == 4;
271 unbash_spaces(dev_name);
272 unbash_spaces(media_type);
273 unbash_spaces(pool_name);
274 unbash_spaces(pool_type);
276 foreach_res(device, R_DEVICE) {
277 /* Find resource, and make sure we were able to open it */
278 if (fnmatch(dev_name.c_str(), device->hdr.name, 0) == 0 &&
279 device->dev && strcmp(device->media_type, media_type.c_str()) == 0) {
280 const int name_len = MAX_NAME_LENGTH;
283 dcr = new_dcr(jcr, device->dev);
287 Dmsg1(120, "Found device %s\n", device->hdr.name);
288 bstrncpy(dcr->pool_name, pool_name, name_len);
289 bstrncpy(dcr->pool_type, pool_type, name_len);
290 bstrncpy(dcr->media_type, media_type, name_len);
291 bstrncpy(dcr->dev_name, dev_name, name_len);
292 jcr->device = device;
293 Dmsg1(220, "Got: %s", dir->msg);
294 return bnet_fsend(dir, OK_device);
299 unbash_spaces(dir->msg);
300 pm_strcpy(jcr->errmsg, dir->msg);
301 Jmsg(jcr, M_INFO, 0, _("Failed command: %s\n"), jcr->errmsg);
303 Jmsg(jcr, M_FATAL, 0, _("\n"
304 " Device \"%s\" with MediaType \"%s\" requested by Dir not found in SD Device resources.\n"),
305 dev_name.c_str(), media_type.c_str());
306 bnet_fsend(dir, NO_device, dev_name.c_str());
308 unbash_spaces(dir->msg);
309 pm_strcpy(jcr->errmsg, dir->msg);
311 Jmsg(jcr, M_INFO, 0, _("Failed command: %s\n"), jcr->errmsg);
313 Jmsg(jcr, M_FATAL, 0, _("Bad Use Device command: %s\n"), jcr->errmsg);
314 bnet_fsend(dir, BAD_use, jcr->errmsg);
318 return false; /* ERROR return */
322 * Destroy the Job Control Record and associated
323 * resources (sockets).
325 void stored_free_jcr(JCR *jcr)
327 if (jcr->file_bsock) {
328 bnet_close(jcr->file_bsock);
329 jcr->file_bsock = NULL;
332 free_pool_memory(jcr->job_name);
334 if (jcr->client_name) {
335 free_memory(jcr->client_name);
336 jcr->client_name = NULL;
338 if (jcr->fileset_name) {
339 free_memory(jcr->fileset_name);
341 if (jcr->fileset_md5) {
342 free_memory(jcr->fileset_md5);
348 if (jcr->RestoreBootstrap) {
349 unlink(jcr->RestoreBootstrap);
350 free_pool_memory(jcr->RestoreBootstrap);
351 jcr->RestoreBootstrap = NULL;
353 if (jcr->next_dev || jcr->prev_dev) {
354 Emsg0(M_FATAL, 0, _("In free_jcr(), but still attached to device!!!!\n"));
356 pthread_cond_destroy(&jcr->job_start_wait);