2 Bacula® - The Network Backup Solution
4 Copyright (C) 2000-2007 Free Software Foundation Europe e.V.
6 The main author of Bacula is Kern Sibbald, with contributions from
7 many others, a complete list can be found in the file AUTHORS.
8 This program is Free Software; you can redistribute it and/or
9 modify it under the terms of version two of the GNU General Public
10 License as published by the Free Software Foundation and included
13 This program is distributed in the hope that it will be useful, but
14 WITHOUT ANY WARRANTY; without even the implied warranty of
15 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16 General Public License for more details.
18 You should have received a copy of the GNU General Public License
19 along with this program; if not, write to the Free Software
20 Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
23 Bacula® is a registered trademark of John Walker.
24 The licensor of Bacula is the Free Software Foundation Europe
25 (FSFE), Fiduciary Program, Sumatrastrasse 25, 8006 Zürich,
26 Switzerland, email:ftf@fsfeurope.org.
30 * Bacula Director -- backup.c -- responsible for doing backup jobs
32 * Kern Sibbald, March MM
34 * Basic tasks done here:
35 * Open DB and create records for this job.
36 * Open Message Channel with Storage daemon to tell him a job will be starting.
37 * Open connection with File daemon and pass him commands
39 * When the File daemon finishes the job, update the DB.
48 /* Commands sent to File daemon */
49 static char backupcmd[] = "backup\n";
50 static char storaddr[] = "storage address=%s port=%d ssl=%d\n";
52 /* Responses received from File daemon */
53 static char OKbackup[] = "2000 OK backup\n";
54 static char OKstore[] = "2000 OK storage\n";
55 static char EndJob[] = "2800 End Job TermCode=%d JobFiles=%u "
56 "ReadBytes=%llu JobBytes=%llu Errors=%u "
57 "VSS=%d Encrypt=%d\n";
58 /* Pre 1.39.29 (04Dec06) EndJob */
59 static char OldEndJob[] = "2800 End Job TermCode=%d JobFiles=%u "
60 "ReadBytes=%llu JobBytes=%llu Errors=%u\n";
62 * Called here before the job is run to do the job
65 bool do_backup_init(JCR *jcr)
68 free_rstorage(jcr); /* we don't read so release */
70 if (!get_or_create_fileset_record(jcr)) {
75 * Get definitive Job level and since time
77 get_level_since_time(jcr, jcr->since, sizeof(jcr->since));
79 apply_pool_overrides(jcr);
81 jcr->jr.PoolId = get_or_create_pool_record(jcr, jcr->pool->name());
82 if (jcr->jr.PoolId == 0) {
86 /* If pool storage specified, use it instead of job storage */
87 copy_wstorage(jcr, jcr->pool->storage, _("Pool resource"));
90 Jmsg(jcr, M_FATAL, 0, _("No Storage specification found in Job or Pool.\n"));
94 create_clones(jcr); /* run any clone jobs */
100 * Do a backup of the specified FileSet
102 * Returns: false on failure
105 bool do_backup(JCR *jcr)
108 int tls_need = BNET_TLS_NONE;
114 /* Print Job Start message */
115 Jmsg(jcr, M_INFO, 0, _("Start Backup JobId %s, Job=%s\n"),
116 edit_uint64(jcr->JobId, ed1), jcr->Job);
118 set_jcr_job_status(jcr, JS_Running);
119 Dmsg2(100, "JobId=%d JobLevel=%c\n", jcr->jr.JobId, jcr->jr.JobLevel);
120 if (!db_update_job_start_record(jcr, jcr->db, &jcr->jr)) {
121 Jmsg(jcr, M_FATAL, 0, "%s", db_strerror(jcr->db));
126 * Open a message channel connection with the Storage
127 * daemon. This is to let him know that our client
128 * will be contacting him for a backup session.
131 Dmsg0(110, "Open connection with storage daemon\n");
132 set_jcr_job_status(jcr, JS_WaitSD);
134 * Start conversation with Storage daemon
136 if (!connect_to_storage_daemon(jcr, 10, SDConnectTimeout, 1)) {
140 * Now start a job with the Storage daemon
142 if (!start_storage_daemon_job(jcr, NULL, jcr->wstorage)) {
147 * Start the job prior to starting the message thread below
148 * to avoid two threads from using the BSOCK structure at
151 if (!bnet_fsend(jcr->store_bsock, "run")) {
156 * Now start a Storage daemon message thread. Note,
157 * this thread is used to provide the catalog services
158 * for the backup job, including inserting the attributes
159 * into the catalog. See catalog_update() in catreq.c
161 if (!start_storage_daemon_message_thread(jcr)) {
164 Dmsg0(150, "Storage daemon connection OK\n");
166 set_jcr_job_status(jcr, JS_WaitFD);
167 if (!connect_to_file_daemon(jcr, 10, FDConnectTimeout, 1)) {
171 set_jcr_job_status(jcr, JS_Running);
172 fd = jcr->file_bsock;
174 if (!send_include_list(jcr)) {
178 if (!send_exclude_list(jcr)) {
182 if (!send_level_command(jcr)) {
187 * send Storage daemon address to the File daemon
190 if (store->SDDport == 0) {
191 store->SDDport = store->SDport;
194 /* TLS Requirement */
195 if (store->tls_enable) {
196 if (store->tls_require) {
197 tls_need = BNET_TLS_REQUIRED;
199 tls_need = BNET_TLS_OK;
203 fd->fsend(storaddr, store->address, store->SDDport, tls_need);
204 if (!response(jcr, fd, OKstore, "Storage", DISPLAY_ERROR)) {
208 if (!send_runscripts_commands(jcr)) {
213 * We re-update the job start record so that the start
214 * time is set after the run before job. This avoids
215 * that any files created by the run before job will
216 * be saved twice. They will be backed up in the current
217 * job, but not in the next one unless they are changed.
218 * Without this, they will be backed up in this job and
219 * in the next job run because in that case, their date
220 * is after the start of this run.
222 jcr->start_time = time(NULL);
223 jcr->jr.StartTime = jcr->start_time;
224 if (!db_update_job_start_record(jcr, jcr->db, &jcr->jr)) {
225 Jmsg(jcr, M_FATAL, 0, "%s", db_strerror(jcr->db));
228 /* Send backup command */
229 fd->fsend(backupcmd);
230 if (!response(jcr, fd, OKbackup, "backup", DISPLAY_ERROR)) {
234 /* Pickup Job termination data */
235 stat = wait_for_job_termination(jcr);
236 db_write_batch_file_records(jcr); /* used by bulk batch file insert */
237 if (stat == JS_Terminated) {
238 backup_cleanup(jcr, stat);
243 /* Come here only after starting SD thread */
245 set_jcr_job_status(jcr, JS_ErrorTerminated);
246 Dmsg1(400, "wait for sd. use=%d\n", jcr->use_count());
248 cancel_storage_daemon_job(jcr);
249 wait_for_storage_daemon_termination(jcr);
250 Dmsg1(400, "after wait for sd. use=%d\n", jcr->use_count());
256 * Here we wait for the File daemon to signal termination,
257 * then we wait for the Storage daemon. When both
258 * are done, we return the job status.
259 * Also used by restore.c
261 int wait_for_job_termination(JCR *jcr)
264 BSOCK *fd = jcr->file_bsock;
266 uint32_t JobFiles, Errors;
267 uint64_t ReadBytes = 0;
268 uint64_t JobBytes = 0;
272 set_jcr_job_status(jcr, JS_Running);
273 /* Wait for Client to terminate */
274 while ((n = bget_dirmsg(fd)) >= 0) {
276 (sscanf(fd->msg, EndJob, &jcr->FDJobStatus, &JobFiles,
277 &ReadBytes, &JobBytes, &Errors, &VSS, &Encrypt) == 7 ||
278 sscanf(fd->msg, OldEndJob, &jcr->FDJobStatus, &JobFiles,
279 &ReadBytes, &JobBytes, &Errors) == 5)) {
281 set_jcr_job_status(jcr, jcr->FDJobStatus);
282 Dmsg1(100, "FDStatus=%c\n", (char)jcr->JobStatus);
284 Jmsg(jcr, M_WARNING, 0, _("Unexpected Client Job message: %s\n"),
287 if (job_canceled(jcr)) {
292 if (is_bnet_error(fd)) {
293 Jmsg(jcr, M_FATAL, 0, _("Network error with FD during %s: ERR=%s\n"),
294 job_type_to_str(jcr->JobType), fd->bstrerror());
296 fd->signal(BNET_TERMINATE); /* tell Client we are terminating */
298 /* Force cancel in SD if failing */
299 if (job_canceled(jcr) || !fd_ok) {
300 cancel_storage_daemon_job(jcr);
303 /* Note, the SD stores in jcr->JobFiles/ReadBytes/JobBytes/Errors */
304 wait_for_storage_daemon_termination(jcr);
307 /* Return values from FD */
309 jcr->JobFiles = JobFiles;
310 jcr->Errors = Errors;
311 jcr->ReadBytes = ReadBytes;
312 jcr->JobBytes = JobBytes;
314 jcr->Encrypt = Encrypt;
316 Jmsg(jcr, M_FATAL, 0, _("No Job status returned from FD.\n"));
319 // Dmsg4(100, "fd_ok=%d FDJS=%d JS=%d SDJS=%d\n", fd_ok, jcr->FDJobStatus,
320 // jcr->JobStatus, jcr->SDJobStatus);
322 /* Return the first error status we find Dir, FD, or SD */
323 if (!fd_ok || is_bnet_error(fd)) {
324 jcr->FDJobStatus = JS_ErrorTerminated;
326 if (jcr->JobStatus != JS_Terminated) {
327 return jcr->JobStatus;
329 if (jcr->FDJobStatus != JS_Terminated) {
330 return jcr->FDJobStatus;
332 return jcr->SDJobStatus;
336 * Release resources allocated during backup.
338 void backup_cleanup(JCR *jcr, int TermCode)
340 char sdt[50], edt[50], schedt[50];
341 char ec1[30], ec2[30], ec3[30], ec4[30], ec5[30], compress[50];
342 char ec6[30], ec7[30], ec8[30], elapsed[50];
343 char term_code[100], fd_term_msg[100], sd_term_msg[100];
344 const char *term_msg;
345 int msg_type = M_INFO;
348 double kbps, compression;
351 Dmsg2(100, "Enter backup_cleanup %d %c\n", TermCode, TermCode);
352 memset(&mr, 0, sizeof(mr));
353 memset(&cr, 0, sizeof(cr));
355 update_job_end(jcr, TermCode);
357 if (!db_get_job_record(jcr, jcr->db, &jcr->jr)) {
358 Jmsg(jcr, M_WARNING, 0, _("Error getting Job record for Job report: ERR=%s"),
359 db_strerror(jcr->db));
360 set_jcr_job_status(jcr, JS_ErrorTerminated);
363 bstrncpy(cr.Name, jcr->client->name(), sizeof(cr.Name));
364 if (!db_get_client_record(jcr, jcr->db, &cr)) {
365 Jmsg(jcr, M_WARNING, 0, _("Error getting Client record for Job report: ERR=%s"),
366 db_strerror(jcr->db));
369 bstrncpy(mr.VolumeName, jcr->VolumeName, sizeof(mr.VolumeName));
370 if (!db_get_media_record(jcr, jcr->db, &mr)) {
371 Jmsg(jcr, M_WARNING, 0, _("Error getting Media record for Volume \"%s\": ERR=%s"),
372 mr.VolumeName, db_strerror(jcr->db));
373 set_jcr_job_status(jcr, JS_ErrorTerminated);
376 update_bootstrap_file(jcr);
378 switch (jcr->JobStatus) {
380 if (jcr->Errors || jcr->SDErrors) {
381 term_msg = _("Backup OK -- with warnings");
383 term_msg = _("Backup OK");
387 case JS_ErrorTerminated:
388 term_msg = _("*** Backup Error ***");
389 msg_type = M_ERROR; /* Generate error message */
390 if (jcr->store_bsock) {
391 jcr->store_bsock->signal(BNET_TERMINATE);
392 if (jcr->SD_msg_chan) {
393 pthread_cancel(jcr->SD_msg_chan);
398 term_msg = _("Backup Canceled");
399 if (jcr->store_bsock) {
400 jcr->store_bsock->signal(BNET_TERMINATE);
401 if (jcr->SD_msg_chan) {
402 pthread_cancel(jcr->SD_msg_chan);
407 term_msg = term_code;
408 sprintf(term_code, _("Inappropriate term code: %c\n"), jcr->JobStatus);
411 bstrftimes(schedt, sizeof(schedt), jcr->jr.SchedTime);
412 bstrftimes(sdt, sizeof(sdt), jcr->jr.StartTime);
413 bstrftimes(edt, sizeof(edt), jcr->jr.EndTime);
414 RunTime = jcr->jr.EndTime - jcr->jr.StartTime;
418 kbps = (double)jcr->jr.JobBytes / (1000 * RunTime);
420 if (!db_get_job_volume_names(jcr, jcr->db, jcr->jr.JobId, &jcr->VolumeName)) {
422 * Note, if the job has erred, most likely it did not write any
423 * tape, so suppress this "error" message since in that case
424 * it is normal. Or look at it the other way, only for a
425 * normal exit should we complain about this error.
427 if (jcr->JobStatus == JS_Terminated && jcr->jr.JobBytes) {
428 Jmsg(jcr, M_ERROR, 0, "%s", db_strerror(jcr->db));
430 jcr->VolumeName[0] = 0; /* none */
433 if (jcr->ReadBytes == 0) {
434 bstrncpy(compress, "None", sizeof(compress));
436 compression = (double)100 - 100.0 * ((double)jcr->JobBytes / (double)jcr->ReadBytes);
437 if (compression < 0.5) {
438 bstrncpy(compress, "None", sizeof(compress));
440 bsnprintf(compress, sizeof(compress), "%.1f %%", (float)compression);
443 jobstatus_to_ascii(jcr->FDJobStatus, fd_term_msg, sizeof(fd_term_msg));
444 jobstatus_to_ascii(jcr->SDJobStatus, sd_term_msg, sizeof(sd_term_msg));
446 // bmicrosleep(15, 0); /* for debugging SIGHUP */
448 Jmsg(jcr, msg_type, 0, _("Bacula %s %s (%s): %s\n"
449 " Build OS: %s %s %s\n"
452 " Backup Level: %s%s\n"
453 " Client: \"%s\" %s\n"
454 " FileSet: \"%s\" %s\n"
455 " Pool: \"%s\" (From %s)\n"
456 " Storage: \"%s\" (From %s)\n"
457 " Scheduled time: %s\n"
460 " Elapsed time: %s\n"
462 " FD Files Written: %s\n"
463 " SD Files Written: %s\n"
464 " FD Bytes Written: %s (%sB)\n"
465 " SD Bytes Written: %s (%sB)\n"
467 " Software Compression: %s\n"
470 " Volume name(s): %s\n"
471 " Volume Session Id: %d\n"
472 " Volume Session Time: %d\n"
473 " Last Volume Bytes: %s (%sB)\n"
474 " Non-fatal FD errors: %d\n"
476 " FD termination status: %s\n"
477 " SD termination status: %s\n"
478 " Termination: %s\n\n"),
479 my_name, VERSION, LSMDATE, edt,
480 HOST_OS, DISTNAME, DISTVER,
483 level_to_str(jcr->JobLevel), jcr->since,
484 jcr->client->name(), cr.Uname,
485 jcr->fileset->name(), jcr->FSCreateTime,
486 jcr->pool->name(), jcr->pool_source,
487 jcr->wstore->name(), jcr->wstore_source,
491 edit_utime(RunTime, elapsed, sizeof(elapsed)),
493 edit_uint64_with_commas(jcr->jr.JobFiles, ec1),
494 edit_uint64_with_commas(jcr->SDJobFiles, ec2),
495 edit_uint64_with_commas(jcr->jr.JobBytes, ec3),
496 edit_uint64_with_suffix(jcr->jr.JobBytes, ec4),
497 edit_uint64_with_commas(jcr->SDJobBytes, ec5),
498 edit_uint64_with_suffix(jcr->SDJobBytes, ec6),
502 jcr->Encrypt?"yes":"no",
506 edit_uint64_with_commas(mr.VolBytes, ec7),
507 edit_uint64_with_suffix(mr.VolBytes, ec8),
514 Dmsg0(100, "Leave backup_cleanup()\n");
517 void update_bootstrap_file(JCR *jcr)
519 /* Now update the bootstrap file if any */
520 if (jcr->JobStatus == JS_Terminated && jcr->jr.JobBytes &&
521 jcr->job->WriteBootstrap) {
525 POOLMEM *fname = get_pool_memory(PM_FNAME);
526 fname = edit_job_codes(jcr, fname, jcr->job->WriteBootstrap, "");
528 VOL_PARAMS *VolParams = NULL;
534 bpipe = open_bpipe(fname+1, 0, "w"); /* skip first char "|" */
535 fd = bpipe ? bpipe->wfd : NULL;
537 /* ***FIXME*** handle BASE */
538 fd = fopen(fname, jcr->JobLevel==L_FULL?"w+b":"a+b");
541 VolCount = db_get_job_volume_parameters(jcr, jcr->db, jcr->JobId,
544 Jmsg(jcr, M_ERROR, 0, _("Could not get Job Volume Parameters to "
545 "update Bootstrap file. ERR=%s\n"), db_strerror(jcr->db));
546 if (jcr->SDJobFiles != 0) {
547 set_jcr_job_status(jcr, JS_ErrorTerminated);
551 /* Start output with when and who wrote it */
552 bstrftimes(edt, sizeof(edt), time(NULL));
553 fprintf(fd, "# %s - %s - %s%s\n", edt, jcr->jr.Job,
554 level_to_str(jcr->JobLevel), jcr->since);
555 for (int i=0; i < VolCount; i++) {
556 /* Write the record */
557 fprintf(fd, "Volume=\"%s\"\n", VolParams[i].VolumeName);
558 fprintf(fd, "MediaType=\"%s\"\n", VolParams[i].MediaType);
559 fprintf(fd, "VolSessionId=%u\n", jcr->VolSessionId);
560 fprintf(fd, "VolSessionTime=%u\n", jcr->VolSessionTime);
561 fprintf(fd, "VolFile=%u-%u\n", VolParams[i].StartFile,
562 VolParams[i].EndFile);
563 fprintf(fd, "VolBlock=%u-%u\n", VolParams[i].StartBlock,
564 VolParams[i].EndBlock);
565 fprintf(fd, "FileIndex=%d-%d\n", VolParams[i].FirstIndex,
566 VolParams[i].LastIndex);
578 Jmsg(jcr, M_ERROR, 0, _("Could not open WriteBootstrap file:\n"
579 "%s: ERR=%s\n"), fname, be.bstrerror());
580 set_jcr_job_status(jcr, JS_ErrorTerminated);
582 free_pool_memory(fname);