/*
* Cancel a running job on a storage daemon. Used by the interactive cancel
* command to cancel a JobId on a Storage Daemon this can be used when the
* Director already removed the Job and thinks it finished but the Storage
* Daemon still thinks its active.
*/
bool cancel_storage_daemon_job(UAContext *ua, STORERES *store, char *JobId)
{
BSOCK *sd;
JCR *control_jcr;
control_jcr = new_control_jcr("*JobCancel*", JT_SYSTEM);
control_jcr->res.wstore = store;
if (!connect_to_storage_daemon(control_jcr, 10, me->SDConnectTimeout, true)) {
ua->error_msg(_("Failed to connect to Storage daemon.\n"));
}
Dmsg0(200, "Connected to storage daemon\n");
sd = control_jcr->store_bsock;
sd->fsend("cancel Job=%s\n", JobId);
while (sd->recv() >= 0) {
ua->send_msg("%s", sd->msg);
}
sd->signal(BNET_TERMINATE);
sd->close();
control_jcr->store_bsock = NULL;
free_jcr(control_jcr);
return true;
}
/*
* Open a connection to a SD.
*/
BSOCK *open_sd_bsock(UAContext *ua)
{
STORERES *store = ua->jcr->res.wstore;
if (!ua->jcr->store_bsock) {
ua->send_msg(_("Connecting to Storage daemon %s at %s:%d ...\n"),
store->name(), store->address, store->SDport);
if (!connect_to_storage_daemon(ua->jcr, 10, me->SDConnectTimeout, true)) {
ua->error_msg(_("Failed to connect to Storage daemon.\n"));
return NULL;
}
}
return ua->jcr->store_bsock;
}
void cancel_storage_daemon_job(JCR *jcr)
{
if (jcr->sd_canceled) {
return; /* cancel only once */
}
UAContext *ua = new_ua_context(jcr);
JCR *control_jcr = new_control_jcr("*JobCancel*", JT_SYSTEM);
BSOCK *sd;
ua->jcr = control_jcr;
if (jcr->store_bsock) {
if (!ua->jcr->wstorage) {
if (jcr->rstorage) {
copy_wstorage(ua->jcr, jcr->rstorage, _("Job resource"));
} else {
copy_wstorage(ua->jcr, jcr->wstorage, _("Job resource"));
}
} else {
USTORE store;
if (jcr->rstorage) {
store.store = jcr->rstore;
} else {
store.store = jcr->wstore;
}
set_wstorage(ua->jcr, &store);
}
if (!connect_to_storage_daemon(ua->jcr, 10, SDConnectTimeout, 1)) {
goto bail_out;
}
Dmsg0(200, "Connected to storage daemon\n");
sd = ua->jcr->store_bsock;
sd->fsend("cancel Job=%s\n", jcr->Job);
while (sd->recv() >= 0) {
}
sd->signal(BNET_TERMINATE);
sd->close();
ua->jcr->store_bsock = NULL;
jcr->sd_canceled = true;
jcr->store_bsock->set_timed_out();
jcr->store_bsock->set_terminated();
sd_msg_thread_send_signal(jcr, TIMEOUT_SIGNAL);
jcr->my_thread_send_signal(TIMEOUT_SIGNAL);
}
bail_out:
free_jcr(control_jcr);
free_ua_context(ua);
}
/*
* Cancel a running job on a storage daemon. The silent flag sets
* if we need to be silent or not e.g. when doing an interactive cancel
* or a system invoked one.
*/
bool cancel_storage_daemon_job(UAContext *ua, JCR *jcr, bool silent)
{
BSOCK *sd;
USTORERES store;
if (!ua->jcr->wstorage) {
if (jcr->rstorage) {
copy_wstorage(ua->jcr, jcr->rstorage, _("Job resource"));
} else {
copy_wstorage(ua->jcr, jcr->wstorage, _("Job resource"));
}
} else {
if (jcr->rstorage) {
store.store = jcr->res.rstore;
} else {
store.store = jcr->res.wstore;
}
set_wstorage(ua->jcr, &store);
}
if (!connect_to_storage_daemon(ua->jcr, 10, me->SDConnectTimeout, true)) {
if (!silent) {
ua->error_msg(_("Failed to connect to Storage daemon.\n"));
}
return false;
}
Dmsg0(200, "Connected to storage daemon\n");
sd = ua->jcr->store_bsock;
sd->fsend(canceljobcmd, jcr->Job);
while (sd->recv() >= 0) {
if (!silent) {
ua->send_msg("%s", sd->msg);
}
}
sd->signal(BNET_TERMINATE);
sd->close();
delete ua->jcr->store_bsock;
ua->jcr->store_bsock = NULL;
if (silent) {
jcr->sd_canceled = true;
}
jcr->store_bsock->set_timed_out();
jcr->store_bsock->set_terminated();
sd_msg_thread_send_signal(jcr, TIMEOUT_SIGNAL);
jcr->my_thread_send_signal(TIMEOUT_SIGNAL);
return true;
}
/*
* Get the status of a remote storage daemon.
*/
void do_native_storage_status(UAContext *ua, STORERES *store, char *cmd)
{
BSOCK *sd;
USTORERES lstore;
lstore.store = store;
pm_strcpy(lstore.store_source, _("unknown source"));
set_wstorage(ua->jcr, &lstore);
/*
* Try connecting for up to 15 seconds
*/
if (!ua->api) {
ua->send_msg(_("Connecting to Storage daemon %s at %s:%d\n"),
store->name(), store->address, store->SDport);
}
if (!connect_to_storage_daemon(ua->jcr, 10, me->SDConnectTimeout, false)) {
ua->send_msg(_("\nFailed to connect to Storage daemon %s.\n====\n"),
store->name());
if (ua->jcr->store_bsock) {
ua->jcr->store_bsock->close();
delete ua->jcr->store_bsock;
ua->jcr->store_bsock = NULL;
}
return;
}
Dmsg0(20, _("Connected to storage daemon\n"));
sd = ua->jcr->store_bsock;
if (cmd) {
sd->fsend(dotstatuscmd, cmd);
} else {
sd->fsend(statuscmd);
}
while (sd->recv() >= 0) {
ua->send_msg("%s", sd->msg);
}
sd->signal( BNET_TERMINATE);
sd->close();
delete ua->jcr->store_bsock;
ua->jcr->store_bsock = NULL;
return;
}
/*
* Do a virtual backup, which consolidates all previous backups into
* a sort of synthetic Full.
*
* Returns: false on failure
* true on success
*/
bool do_native_vbackup(JCR *jcr)
{
char ed1[100];
BSOCK *sd;
char *p;
db_list_ctx jobids;
if (!jcr->res.rstorage) {
Jmsg(jcr, M_FATAL, 0, _("No storage for reading given.\n"));
return false;
}
if (!jcr->res.wstorage) {
Jmsg(jcr, M_FATAL, 0, _("No storage for writing given.\n"));
return false;
}
Dmsg2(100, "rstorage=%p wstorage=%p\n", jcr->res.rstorage, jcr->res.wstorage);
Dmsg2(100, "Read store=%s, write store=%s\n",
((STORERES *)jcr->res.rstorage->first())->name(),
((STORERES *)jcr->res.wstorage->first())->name());
/*
* Print Job Start message
*/
Jmsg(jcr, M_INFO, 0, _("Start Virtual Backup JobId %s, Job=%s\n"),
edit_uint64(jcr->JobId, ed1), jcr->Job);
if (!jcr->accurate) {
Jmsg(jcr, M_WARNING, 0,
_("This Job is not an Accurate backup so is not equivalent to a Full backup.\n"));
}
db_accurate_get_jobids(jcr, jcr->db, &jcr->jr, &jobids);
Dmsg1(10, "Accurate jobids=%s\n", jobids.list);
if (jobids.count == 0) {
Jmsg(jcr, M_FATAL, 0, _("No previous Jobs found.\n"));
return false;
}
/*
* Now we find the last job that ran and store it's info in
* the previous_jr record. We will set our times to the
* values from that job so that anything changed after that
* time will be picked up on the next backup.
*/
p = strrchr(jobids.list, ','); /* find last jobid */
if (p != NULL) {
p++;
} else {
p = jobids.list;
}
memset(&jcr->previous_jr, 0, sizeof(jcr->previous_jr));
jcr->previous_jr.JobId = str_to_int64(p);
Dmsg1(10, "Previous JobId=%s\n", p);
if (!db_get_job_record(jcr, jcr->db, &jcr->previous_jr)) {
Jmsg(jcr, M_FATAL, 0, _("Error getting Job record for previous Job: ERR=%s"),
db_strerror(jcr->db));
return false;
}
if (!create_bootstrap_file(jcr, jobids.list)) {
Jmsg(jcr, M_FATAL, 0, _("Could not get or create the FileSet record.\n"));
return false;
}
/*
* Open a message channel connection with the Storage
* daemon. This is to let him know that our client
* will be contacting him for a backup session.
*
*/
Dmsg0(110, "Open connection with storage daemon\n");
jcr->setJobStatus(JS_WaitSD);
/*
* Start conversation with Storage daemon
*/
if (!connect_to_storage_daemon(jcr, 10, me->SDConnectTimeout, true)) {
return false;
}
sd = jcr->store_bsock;
/*
* Now start a job with the Storage daemon
*/
if (!start_storage_daemon_job(jcr, jcr->res.rstorage, jcr->res.wstorage, /* send_bsr */ true)) {
return false;
}
Dmsg0(100, "Storage daemon connection OK\n");
/*
* We re-update the job start record so that the start
* time is set after the run before job. This avoids
* that any files created by the run before job will
* be saved twice. They will be backed up in the current
* job, but not in the next one unless they are changed.
* Without this, they will be backed up in this job and
* in the next job run because in that case, their date
* is after the start of this run.
*/
jcr->start_time = time(NULL);
jcr->jr.StartTime = jcr->start_time;
jcr->jr.JobTDate = jcr->start_time;
jcr->setJobStatus(JS_Running);
/*
* Update job start record
*/
if (!db_update_job_start_record(jcr, jcr->db, &jcr->jr)) {
Jmsg(jcr, M_FATAL, 0, "%s", db_strerror(jcr->db));
return false;
}
/*
* Declare the job started to start the MaxRunTime check
*/
jcr->setJobStarted();
/*
* Start the job prior to starting the message thread below
* to avoid two threads from using the BSOCK structure at
* the same time.
*/
if (!sd->fsend("run")) {
return false;
}
/*
* Now start a Storage daemon message thread
*/
if (!start_storage_daemon_message_thread(jcr)) {
return false;
}
jcr->setJobStatus(JS_Running);
/*
* Pickup Job termination data
* Note, the SD stores in jcr->JobFiles/ReadBytes/JobBytes/JobErrors
*/
wait_for_storage_daemon_termination(jcr);
jcr->setJobStatus(jcr->SDJobStatus);
db_write_batch_file_records(jcr); /* used by bulk batch file insert */
if (!jcr->is_JobStatus(JS_Terminated)) {
return false;
}
native_vbackup_cleanup(jcr, jcr->JobStatus);
return true;
}
/*
* The bootstrap is stored in a file, so open the file, and loop
* through it processing each storage device in turn. If the
* storage is different from the prior one, we open a new connection
* to the new storage and do a restore for that part.
*
* This permits handling multiple storage daemons for a single
* restore. E.g. your Full is stored on tape, and Incrementals
* on disk.
*/
static inline bool do_ndmp_restore_bootstrap(JCR *jcr)
{
int cnt;
BSOCK *sd;
BSR *bsr;
NIS *nis = NULL;
int32_t current_fi;
bootstrap_info info;
BSR_FINDEX *fileindex;
struct ndm_session ndmp_sess;
struct ndm_job_param ndmp_job;
bool session_initialized = false;
bool retval = false;
int NdmpLoglevel;
if (jcr->res.client->ndmp_loglevel > me->ndmp_loglevel) {
NdmpLoglevel = jcr->res.client->ndmp_loglevel;
} else {
NdmpLoglevel = me->ndmp_loglevel;
}
/*
* We first parse the BSR ourself so we know what to restore.
*/
jcr->bsr = parse_bsr(jcr, jcr->RestoreBootstrap);
if (!jcr->bsr) {
Jmsg(jcr, M_FATAL, 0, _("Error parsing bootstrap file.\n"));
goto bail_out;
}
/*
* Setup all paired read storage.
*/
set_paired_storage(jcr);
if (!jcr->res.pstore) {
Jmsg(jcr, M_FATAL, 0,
_("Read storage %s doesn't point to storage definition with paired storage option.\n"),
jcr->res.rstore->name());
goto bail_out;
}
/*
* Open the bootstrap file
*/
if (!open_bootstrap_file(jcr, info)) {
goto bail_out;
}
nis = (NIS *)malloc(sizeof(NIS));
memset(nis, 0, sizeof(NIS));
/*
* Read the bootstrap file
*/
bsr = jcr->bsr;
while (!feof(info.bs)) {
if (!select_next_rstore(jcr, info)) {
goto cleanup;
}
/*
* Initialize the ndmp restore job. We build the generic job once per storage daemon
* and reuse the job definition for each seperate sub-restore we perform as
* part of the whole job. We only free the env_table between every sub-restore.
*/
if (!ndmp_build_client_job(jcr, jcr->res.client, jcr->res.pstore, NDM_JOB_OP_EXTRACT, &ndmp_job)) {
goto cleanup;
}
/*
* Open a message channel connection with the Storage
* daemon. This is to let him know that our client
* will be contacting him for a backup session.
*
*/
Dmsg0(10, "Open connection with storage daemon\n");
jcr->setJobStatus(JS_WaitSD);
/*
* Start conversation with Storage daemon
*/
if (!connect_to_storage_daemon(jcr, 10, me->SDConnectTimeout, true)) {
goto cleanup;
}
sd = jcr->store_bsock;
/*
* Now start a job with the Storage daemon
*/
if (!start_storage_daemon_job(jcr, jcr->res.rstorage, NULL)) {
goto cleanup;
}
jcr->setJobStatus(JS_Running);
/*
* Send the bootstrap file -- what Volumes/files to restore
*/
if (!send_bootstrap_file(jcr, sd, info) ||
!response(jcr, sd, OKbootstrap, "Bootstrap", DISPLAY_ERROR)) {
goto cleanup;
}
if (!sd->fsend("run")) {
goto cleanup;
}
/*
* Now start a Storage daemon message thread
*/
if (!start_storage_daemon_message_thread(jcr)) {
goto cleanup;
}
Dmsg0(50, "Storage daemon connection OK\n");
/*
* Walk over each bsr record
*/
cnt = 0;
for (bsr = jcr->bsr; bsr; bsr = bsr->next) {
/*
* Walk each fileindex of the current BSR record. Each different fileindex is
* a separate NDMP stream.
*/
for (fileindex = bsr->FileIndex; fileindex; fileindex = fileindex->next) {
for (current_fi = fileindex->findex; current_fi <= fileindex->findex2; current_fi++) {
/*
* See if this is the first Restore NDMP stream or not. For NDMP we can have multiple Backup
* runs as part of the same Job. When we are restoring data from a Native Storage Daemon
* we let it know to expect a next restore session. It will generate a new authorization
* key so we wait for the nextrun_ready conditional variable to be raised by the msg_thread.
*/
if (jcr->store_bsock && cnt > 0) {
jcr->store_bsock->fsend("nextrun");
P(mutex);
pthread_cond_wait(&jcr->nextrun_ready, &mutex);
V(mutex);
}
/*
* Perform the actual NDMP job.
* Initialize a new NDMP session
*/
memset(&ndmp_sess, 0, sizeof(ndmp_sess));
ndmp_sess.conn_snooping = (me->ndmp_snooping) ? 1 : 0;
ndmp_sess.control_agent_enabled = 1;
ndmp_sess.param = (struct ndm_session_param *)malloc(sizeof(struct ndm_session_param));
memset(ndmp_sess.param, 0, sizeof(struct ndm_session_param));
ndmp_sess.param->log.deliver = ndmp_loghandler;
ndmp_sess.param->log_level = native_to_ndmp_loglevel(NdmpLoglevel, debug_level, nis);
nis->jcr = jcr;
ndmp_sess.param->log.ctx = nis;
ndmp_sess.param->log_tag = bstrdup("DIR-NDMP");
/*
* Initialize the session structure.
*/
if (ndma_session_initialize(&ndmp_sess)) {
goto cleanup_ndmp;
}
session_initialized = true;
/*
* Copy the actual job to perform.
*/
jcr->jr.FileIndex = current_fi;
if (bsr->sessid && bsr->sesstime) {
jcr->jr.VolSessionId = bsr->sessid->sessid;
jcr->jr.VolSessionTime = bsr->sesstime->sesstime;
} else {
Jmsg(jcr, M_FATAL, 0, _("Wrong BSR missing sessid and/or sesstime\n"));
goto cleanup_ndmp;
}
memcpy(&ndmp_sess.control_acb->job, &ndmp_job, sizeof(struct ndm_job_param));
if (!fill_restore_environment(jcr,
current_fi,
&ndmp_sess.control_acb->job)) {
Jmsg(jcr, M_ERROR, 0, _("ERROR in fill_restore_environment\n"));
goto cleanup_ndmp;
}
ndma_job_auto_adjust(&ndmp_sess.control_acb->job);
if (!ndmp_validate_job(jcr, &ndmp_sess.control_acb->job)) {
Jmsg(jcr, M_ERROR, 0, _("ERROR in ndmp_validate_job\n"));
goto cleanup_ndmp;
}
/*
* Commission the session for a run.
*/
if (ndma_session_commission(&ndmp_sess)) {
Jmsg(jcr, M_ERROR, 0, _("ERROR in ndma_session_commission\n"));
goto cleanup_ndmp;
}
/*
* Setup the DMA.
*/
if (ndmca_connect_control_agent(&ndmp_sess)) {
Jmsg(jcr, M_ERROR, 0, _("ERROR in ndmca_connect_control_agent\n"));
goto cleanup_ndmp;
}
ndmp_sess.conn_open = 1;
ndmp_sess.conn_authorized = 1;
/*
* Let the DMA perform its magic.
*/
if (ndmca_control_agent(&ndmp_sess) != 0) {
Jmsg(jcr, M_ERROR, 0, _("ERROR in ndmca_control_agent\n"));
goto cleanup_ndmp;
}
/*
* See if there were any errors during the restore.
*/
if (!extract_post_restore_stats(jcr, &ndmp_sess)) {
Jmsg(jcr, M_ERROR, 0, _("ERROR in extract_post_restore_stats\n"));
goto cleanup_ndmp;
}
/*
* Reset the NDMP session states.
*/
ndma_session_decommission(&ndmp_sess);
/*
* Cleanup the job after it has run.
*/
ndma_destroy_env_list(&ndmp_sess.control_acb->job.env_tab);
ndma_destroy_env_list(&ndmp_sess.control_acb->job.result_env_tab);
ndma_destroy_nlist(&ndmp_sess.control_acb->job.nlist_tab);
/*
* Release any tape device name allocated.
*/
if (ndmp_sess.control_acb->job.tape_device) {
free(ndmp_sess.control_acb->job.tape_device);
ndmp_sess.control_acb->job.tape_device = NULL;
}
/*
* Destroy the session.
*/
ndma_session_destroy(&ndmp_sess);
/*
* Free the param block.
*/
free(ndmp_sess.param->log_tag);
free(ndmp_sess.param);
ndmp_sess.param = NULL;
/*
* Reset the initialized state so we don't try to cleanup again.
*/
session_initialized = false;
/*
* Keep track that we finished this part of the restore.
*/
cnt++;
}
}
}
/*
* Tell the storage daemon we are done.
*/
jcr->store_bsock->fsend("finish");
wait_for_storage_daemon_termination(jcr);
}
/*
* Jump to the generic cleanup done for every Job.
*/
retval = true;
goto cleanup;
cleanup_ndmp:
/*
* Only need to cleanup when things are initialized.
*/
if (session_initialized) {
ndma_destroy_env_list(&ndmp_sess.control_acb->job.env_tab);
ndma_destroy_env_list(&ndmp_sess.control_acb->job.result_env_tab);
ndma_destroy_nlist(&ndmp_sess.control_acb->job.nlist_tab);
if (ndmp_sess.control_acb->job.tape_device) {
free(ndmp_sess.control_acb->job.tape_device);
}
/*
* Destroy the session.
*/
ndma_session_destroy(&ndmp_sess);
}
if (ndmp_sess.param) {
free(ndmp_sess.param->log_tag);
free(ndmp_sess.param);
}
cleanup:
if (nis) {
free(nis);
}
free_paired_storage(jcr);
close_bootstrap_file(info);
bail_out:
free_tree(jcr->restore_tree_root);
jcr->restore_tree_root = NULL;
return retval;
}
/**
* The bootstrap is stored in a file, so open the file, and loop
* through it processing each storage device in turn. If the
* storage is different from the prior one, we open a new connection
* to the new storage and do a restore for that part.
*
* This permits handling multiple storage daemons for a single
* restore. E.g. your Full is stored on tape, and Incrementals
* on disk.
*/
static inline bool do_native_restore_bootstrap(JCR *jcr)
{
STORERES *store;
bootstrap_info info;
BSOCK *fd = NULL;
BSOCK *sd = NULL;
bool first_time = true;
POOL_MEM restore_cmd(PM_MESSAGE);
/*
* This command is used for each part
*/
build_restore_command(jcr, restore_cmd);
/*
* Open the bootstrap file
*/
if (!open_bootstrap_file(jcr, info)) {
goto bail_out;
}
/*
* Read the bootstrap file
*/
jcr->passive_client = jcr->res.client->passive;
while (!feof(info.bs)) {
if (!select_next_rstore(jcr, info)) {
goto bail_out;
}
store = jcr->res.rstore;
/**
* Open a message channel connection with the Storage
* daemon. This is to let him know that our client
* will be contacting him for a backup session.
*
*/
Dmsg0(10, "Open connection with storage daemon\n");
jcr->setJobStatus(JS_WaitSD);
/*
* Start conversation with Storage daemon
*/
if (!connect_to_storage_daemon(jcr, 10, me->SDConnectTimeout, true)) {
goto bail_out;
}
sd = jcr->store_bsock;
/*
* Now start a job with the Storage daemon
*/
if (!start_storage_daemon_job(jcr, jcr->rstorage, NULL)) {
goto bail_out;
}
if (first_time) {
/*
* Start conversation with File daemon
*/
jcr->setJobStatus(JS_WaitFD);
jcr->keep_sd_auth_key = true; /* don't clear the sd_auth_key now */
if (!connect_to_file_daemon(jcr, 10, me->FDConnectTimeout, true, true)) {
goto bail_out;
}
fd = jcr->file_bsock;
/*
* Check if the file daemon supports passive client mode.
*/
if (jcr->passive_client && jcr->FDVersion < FD_VERSION_51) {
Jmsg(jcr, M_FATAL, 0,
_("Client \"%s\" doesn't support passive client mode. "
"Please upgrade your client or disable compat mode.\n"),
jcr->res.client->name());
goto bail_out;
}
}
jcr->setJobStatus(JS_Running);
/*
* Send the bootstrap file -- what Volumes/files to restore
*/
if (!send_bootstrap_file(jcr, sd, info) ||
!response(jcr, sd, OKbootstrap, "Bootstrap", DISPLAY_ERROR)) {
goto bail_out;
}
if (!jcr->passive_client) {
int tls_need = BNET_TLS_NONE;
/*
* When the client is not in passive mode we can put the SD in
* listen mode for the FD connection. And ask the FD to connect
* to the SD.
*/
if (!sd->fsend("run")) {
goto bail_out;
}
/*
* Now start a Storage daemon message thread
*/
if (!start_storage_daemon_message_thread(jcr)) {
goto bail_out;
}
Dmsg0(50, "Storage daemon connection OK\n");
/*
* Send Storage daemon address to the File daemon,
* then wait for File daemon to make connection
* with Storage daemon.
*/
if (store->SDDport == 0) {
store->SDDport = store->SDport;
}
/*
* TLS Requirement
*/
if (store->tls_enable) {
if (store->tls_require) {
tls_need = BNET_TLS_REQUIRED;
} else {
tls_need = BNET_TLS_OK;
}
}
fd->fsend(storaddrcmd, store->address,
store->SDDport, tls_need, jcr->sd_auth_key);
memset(jcr->sd_auth_key, 0, strlen(jcr->sd_auth_key));
Dmsg1(6, "dird>filed: %s\n", fd->msg);
if (!response(jcr, fd, OKstore, "Storage", DISPLAY_ERROR)) {
goto bail_out;
}
} else {
int tls_need = BNET_TLS_NONE;
CLIENTRES *client = jcr->res.client;
/*
* In passive mode we tell the FD what authorization key to use
* and the ask the SD to initiate the connection.
*/
fd->fsend(setauthorizationcmd, jcr->sd_auth_key);
memset(jcr->sd_auth_key, 0, strlen(jcr->sd_auth_key));
Dmsg1(6, "dird>filed: %s\n", fd->msg);
if (!response(jcr, fd, OKAuthorization, "Setauthorization", DISPLAY_ERROR)) {
goto bail_out;
}
/*
* TLS Requirement
*/
tls_need = BNET_TLS_NONE;
if (client->tls_enable) {
if (client->tls_require) {
tls_need = BNET_TLS_REQUIRED;
} else {
tls_need = BNET_TLS_OK;
}
}
/*
* Tell the SD to connect to the FD.
*/
sd->fsend(passiveclientcmd, client->address, client->FDport, tls_need);
if (!response(jcr, sd, OKpassiveclient, "Passive client", DISPLAY_ERROR)) {
goto bail_out;
}
/*
* Start the Job in the SD.
*/
if (!sd->fsend("run")) {
goto bail_out;
}
/*
* Now start a Storage daemon message thread
*/
if (!start_storage_daemon_message_thread(jcr)) {
goto bail_out;
}
Dmsg0(50, "Storage daemon connection OK\n");
}
/*
* Declare the job started to start the MaxRunTime check
*/
jcr->setJobStarted();
/*
* Only pass "global" commands to the FD once
*/
if (first_time) {
first_time = false;
if (!send_runscripts_commands(jcr)) {
goto bail_out;
}
if (!send_restore_objects(jcr)) {
Dmsg0(000, "FAIL: Send restore objects\n");
goto bail_out;
}
/*
* Only FD version 52 and later understand the sending of plugin options.
*/
if (jcr->FDVersion >= FD_VERSION_52) {
if (!send_plugin_options(jcr)) {
Dmsg0(000, "FAIL: Send plugin options\n");
goto bail_out;
}
} else {
/*
* Plugin options specified and not a FD that understands the new protocol keyword.
*/
if (jcr->plugin_options) {
Jmsg(jcr, M_FATAL, 0,
_("Client \"%s\" doesn't support plugin option passing. "
"Please upgrade your client or disable compat mode.\n"),
jcr->res.client->name());
goto bail_out;
}
}
}
fd->fsend("%s", restore_cmd.c_str());
if (!response(jcr, fd, OKrestore, "Restore", DISPLAY_ERROR)) {
goto bail_out;
}
if (jcr->FDVersion < FD_VERSION_2) { /* Old FD */
break; /* we do only one loop */
} else {
if (!response(jcr, fd, OKstoreend, "Store end", DISPLAY_ERROR)) {
goto bail_out;
}
wait_for_storage_daemon_termination(jcr);
}
} /* the whole boostrap has been send */
if (fd && jcr->FDVersion >= FD_VERSION_2) {
fd->fsend("endrestore");
}
close_bootstrap_file(info);
return true;
bail_out:
if (jcr->file_bsock) {
jcr->file_bsock->signal(BNET_TERMINATE);
jcr->file_bsock->close();
delete jcr->file_bsock;
jcr->file_bsock = NULL;
}
close_bootstrap_file(info);
return false;
}
/*
* Do a verification of the specified files against the Catlaog
*
* Returns: false on failure
* true on success
*/
bool do_verify(JCR *jcr)
{
const char *level;
BSOCK *fd, *sd;
int stat;
char ed1[100];
JOB_DBR jr;
JobId_t verify_jobid = 0;
char *store_address;
uint32_t store_port;
const char *Name;
free_wstorage(jcr); /* we don't write */
memset(&jcr->previous_jr, 0, sizeof(jcr->previous_jr));
/*
* Find JobId of last job that ran. Note, we do this when
* the job actually starts running, not at schedule time,
* so that we find the last job that terminated before
* this job runs rather than before it is scheduled. This
* permits scheduling a Backup and Verify at the same time,
* but with the Verify at a lower priority.
*
* For VERIFY_CATALOG we want the JobId of the last INIT.
* For VERIFY_VOLUME_TO_CATALOG, we want the JobId of the
* last backup Job.
*/
if (jcr->getJobLevel() == L_VERIFY_CATALOG ||
jcr->getJobLevel() == L_VERIFY_VOLUME_TO_CATALOG ||
jcr->getJobLevel() == L_VERIFY_DISK_TO_CATALOG ||
jcr->getJobLevel() == L_VERIFY_DATA) {
memcpy(&jr, &jcr->jr, sizeof(jr));
if (jcr->verify_job &&
(jcr->getJobLevel() == L_VERIFY_VOLUME_TO_CATALOG ||
jcr->getJobLevel() == L_VERIFY_DISK_TO_CATALOG ||
jcr->getJobLevel() == L_VERIFY_DATA)) {
Name = jcr->verify_job->name();
} else {
Name = NULL;
}
Dmsg1(100, "find last jobid for: %s\n", NPRT(Name));
/* see if user supplied a jobid= as run argument or from menu */
if (jcr->RestoreJobId) {
verify_jobid = jcr->RestoreJobId;
Dmsg1(100, "Supplied jobid=%d\n", verify_jobid);
} else {
if (!db_find_last_jobid(jcr, jcr->db, Name, &jr)) {
if (jcr->getJobLevel() == L_VERIFY_CATALOG) {
Jmsg(jcr, M_FATAL, 0, _(
"Unable to find JobId of previous InitCatalog Job.\n"
"Please run a Verify with Level=InitCatalog before\n"
"running the current Job.\n"));
} else {
Jmsg(jcr, M_FATAL, 0, _(
"Unable to find JobId of previous Job for this client.\n"));
}
return false;
}
verify_jobid = jr.JobId;
}
Dmsg1(100, "Last full jobid=%d\n", verify_jobid);
}
/*
* Now get the job record for the previous backup that interests
* us. We use the verify_jobid that we found above.
*/
if (jcr->getJobLevel() == L_VERIFY_CATALOG ||
jcr->getJobLevel() == L_VERIFY_VOLUME_TO_CATALOG ||
jcr->getJobLevel() == L_VERIFY_DISK_TO_CATALOG ||
jcr->getJobLevel() == L_VERIFY_DATA) {
jcr->previous_jr.JobId = verify_jobid;
if (!db_get_job_record(jcr, jcr->db, &jcr->previous_jr)) {
Jmsg(jcr, M_FATAL, 0, _("Could not get job record for previous Job. ERR=%s"),
db_strerror(jcr->db));
return false;
}
if (!(jcr->previous_jr.JobStatus == JS_Terminated ||
jcr->previous_jr.JobStatus == JS_Warnings)) {
Jmsg(jcr, M_FATAL, 0, _("Last Job %d did not terminate normally. JobStatus=%c\n"),
verify_jobid, jcr->previous_jr.JobStatus);
return false;
}
Jmsg(jcr, M_INFO, 0, _("Verifying against JobId=%d Job=%s\n"),
jcr->previous_jr.JobId, jcr->previous_jr.Job);
}
/*
* If we are verifying a Volume, we need the Storage
* daemon, so open a connection, otherwise, just
* create a dummy authorization key (passed to
* File daemon but not used).
*/
if (jcr->getJobLevel() == L_VERIFY_VOLUME_TO_CATALOG || jcr->getJobLevel() == L_VERIFY_DATA) {
int stat;
/*
* Note: negative status is an error, zero status, means
* no files were backed up, so skip calling SD and
* client.
*/
stat = create_restore_bootstrap_file(jcr);
if (stat < 0) { /* error */
return false;
} else if (stat == 0) { /* No files, nothing to do */
verify_cleanup(jcr, JS_Terminated); /* clean up */
return true; /* get out */
}
} else {
jcr->sd_auth_key = bstrdup("dummy"); /* dummy Storage daemon key */
}
/* Pass the original fileset to the client */
if (jcr->getJobLevel() == L_VERIFY_DATA) {
FILESET_DBR fdbr;
memset(&fdbr, 0, sizeof(fdbr));
fdbr.FileSetId = jcr->previous_jr.FileSetId;
if (!db_get_fileset_record(jcr, jcr->db, &fdbr)) {
Jmsg(jcr, M_FATAL, 0,
_("Could not get fileset record from previous Job. ERR=%s"),
db_strerror(jcr->db));
return false;
}
jcr->fileset = (FILESET *)GetResWithName(R_FILESET, fdbr.FileSet);
if (!jcr->fileset) {
if (jcr->verify_job) {
jcr->fileset = jcr->verify_job->fileset;
Jmsg(jcr, M_WARNING, 0,
_("Could not find FileSet resource \"%s\" from previous Job\n"),
fdbr.FileSet);
Jmsg(jcr, M_INFO, 0,
_("Using FileSet \"%\"\n"), jcr->fileset->name());
} else {
Jmsg(jcr, M_FATAL, 0,
_("Could not get FileSet resource for verify Job."));
return false;
}
}
Dmsg1(50, "FileSet = %s\n", jcr->fileset->name());
}
/* Pass the current fileset to the client */
if (jcr->getJobLevel() == L_VERIFY_DISK_TO_CATALOG && jcr->verify_job) {
jcr->fileset = jcr->verify_job->fileset;
}
Dmsg2(100, "ClientId=%u JobLevel=%c\n",
jcr->previous_jr.ClientId, jcr->getJobLevel());
if (!db_update_job_start_record(jcr, jcr->db, &jcr->jr)) {
Jmsg(jcr, M_FATAL, 0, "%s", db_strerror(jcr->db));
return false;
}
/* Print Job Start message */
Jmsg(jcr, M_INFO, 0, _("Start Verify JobId=%s Level=%s Job=%s\n"),
edit_uint64(jcr->JobId, ed1), level_to_str(jcr->getJobLevel()), jcr->Job);
if (jcr->getJobLevel() == L_VERIFY_VOLUME_TO_CATALOG ||
jcr->getJobLevel() == L_VERIFY_DATA)
{
/*
* Start conversation with Storage daemon
*/
jcr->setJobStatus(JS_Blocked);
if (!connect_to_storage_daemon(jcr, 10, SDConnectTimeout, 1)) {
return false;
}
/*
* Now start a job with the Storage daemon
*/
if (!start_storage_daemon_job(jcr, jcr->rstorage, NULL)) {
return false;
}
sd = jcr->store_bsock;
jcr->sd_calls_client = jcr->client->sd_calls_client;
/*
* Send the bootstrap file -- what Volumes/files to restore
*/
if (!send_bootstrap_file(jcr, sd) ||
!response(jcr, sd, OKbootstrap, "Bootstrap", DISPLAY_ERROR)) {
goto bail_out;
}
if (!jcr->sd_calls_client) {
if (!run_storage_and_start_message_thread(jcr, sd)) {
return false;
}
}
}
/*
* OK, now connect to the File daemon
* and ask him for the files.
*/
jcr->setJobStatus(JS_Blocked);
if (!connect_to_file_daemon(jcr, 10, FDConnectTimeout, 1)) {
goto bail_out;
}
jcr->setJobStatus(JS_Running);
fd = jcr->file_bsock;
Dmsg0(30, ">filed: Send include list\n");
if (!send_include_list(jcr)) {
goto bail_out;
}
Dmsg0(30, ">filed: Send exclude list\n");
if (!send_exclude_list(jcr)) {
goto bail_out;
}
/*
* Send Level command to File daemon, as well
* as the Storage address if appropriate.
*/
switch (jcr->getJobLevel()) {
case L_VERIFY_INIT:
level = "init";
break;
case L_VERIFY_CATALOG:
level = "catalog";
break;
case L_VERIFY_DATA:
case L_VERIFY_VOLUME_TO_CATALOG:
if (jcr->sd_calls_client) {
if (jcr->FDVersion < 10) {
Jmsg(jcr, M_FATAL, 0, _("The File daemon does not support SDCallsClient.\n"));
goto bail_out;
}
if (!send_client_addr_to_sd(jcr)) {
goto bail_out;
}
if (!run_storage_and_start_message_thread(jcr, jcr->store_bsock)) {
return false;
}
store_address = jcr->rstore->address; /* dummy */
store_port = 0; /* flag that SD calls FD */
} else {
/*
* send Storage daemon address to the File daemon
*/
if (jcr->rstore->SDDport == 0) {
jcr->rstore->SDDport = jcr->rstore->SDport;
}
store_address = get_storage_address(jcr->client, jcr->rstore);
store_port = jcr->rstore->SDDport;
}
if (!send_store_addr_to_fd(jcr, jcr->rstore, store_address, store_port)) {
goto bail_out;
}
if (!jcr->RestoreBootstrap) {
Jmsg0(jcr, M_FATAL, 0, _("Deprecated feature ... use bootstrap.\n"));
goto bail_out;
}
if (jcr->getJobLevel() == L_VERIFY_VOLUME_TO_CATALOG) {
level = "volume";
} else {
level = "data";
}
break;
case L_VERIFY_DISK_TO_CATALOG:
level="disk_to_catalog";
break;
default:
Jmsg2(jcr, M_FATAL, 0, _("Unimplemented Verify level %d(%c)\n"),
jcr->getJobLevel(),
jcr->getJobLevel());
goto bail_out;
}
if (!send_runscripts_commands(jcr)) {
goto bail_out;
}
/*
* Send verify command/level to File daemon
*/
fd->fsend(verifycmd, level);
if (!response(jcr, fd, OKverify, "Verify", DISPLAY_ERROR)) {
goto bail_out;
}
/*
* Now get data back from File daemon and
* compare it to the catalog or store it in the
* catalog depending on the run type.
*/
/* Compare to catalog */
switch (jcr->getJobLevel()) {
case L_VERIFY_CATALOG:
Dmsg0(10, "Verify level=catalog\n");
jcr->sd_msg_thread_done = true; /* no SD msg thread, so it is done */
jcr->SDJobStatus = JS_Terminated;
get_attributes_and_compare_to_catalog(jcr, jcr->previous_jr.JobId);
break;
case L_VERIFY_VOLUME_TO_CATALOG:
Dmsg0(10, "Verify level=volume\n");
get_attributes_and_compare_to_catalog(jcr, jcr->previous_jr.JobId);
break;
case L_VERIFY_DISK_TO_CATALOG:
Dmsg0(10, "Verify level=disk_to_catalog\n");
jcr->sd_msg_thread_done = true; /* no SD msg thread, so it is done */
jcr->SDJobStatus = JS_Terminated;
get_attributes_and_compare_to_catalog(jcr, jcr->previous_jr.JobId);
break;
case L_VERIFY_INIT:
/* Build catalog */
Dmsg0(10, "Verify level=init\n");
jcr->sd_msg_thread_done = true; /* no SD msg thread, so it is done */
jcr->SDJobStatus = JS_Terminated;
get_attributes_and_put_in_catalog(jcr);
db_end_transaction(jcr, jcr->db); /* terminate any open transaction */
db_write_batch_file_records(jcr);
break;
case L_VERIFY_DATA:
/* Nothing special to do */
bget_dirmsg(fd); /* eat EOD */
break;
default:
Jmsg1(jcr, M_FATAL, 0, _("Unimplemented verify level %d\n"), jcr->getJobLevel());
goto bail_out;
}
stat = wait_for_job_termination(jcr);
verify_cleanup(jcr, stat);
return true;
bail_out:
return false;
}
/*
* Cancel a job -- typically called by the UA (Console program), but may also
* be called by the job watchdog.
*
* Returns: true if cancel appears to be successful
* false on failure. Message sent to ua->jcr.
*/
bool cancel_job(UAContext *ua, JCR *jcr)
{
BSOCK *sd, *fd;
char ed1[50];
int32_t old_status = jcr->JobStatus;
jcr->setJobStatus(JS_Canceled);
switch (old_status) {
case JS_Created:
case JS_WaitJobRes:
case JS_WaitClientRes:
case JS_WaitStoreRes:
case JS_WaitPriority:
case JS_WaitMaxJobs:
case JS_WaitStartTime:
ua->info_msg(_("JobId %s, Job %s marked to be canceled.\n"),
edit_uint64(jcr->JobId, ed1), jcr->Job);
jobq_remove(&job_queue, jcr); /* attempt to remove it from queue */
break;
default:
/* Cancel File daemon */
if (jcr->file_bsock) {
ua->jcr->client = jcr->client;
if (!connect_to_file_daemon(ua->jcr, 10, FDConnectTimeout, 1)) {
ua->error_msg(_("Failed to connect to File daemon.\n"));
return 0;
}
Dmsg0(200, "Connected to file daemon\n");
fd = ua->jcr->file_bsock;
fd->fsend("cancel Job=%s\n", jcr->Job);
while (fd->recv() >= 0) {
ua->send_msg("%s", fd->msg);
}
fd->signal(BNET_TERMINATE);
fd->close();
ua->jcr->file_bsock = NULL;
jcr->file_bsock->set_terminated();
jcr->my_thread_send_signal(TIMEOUT_SIGNAL);
}
/* Cancel Storage daemon */
if (jcr->store_bsock) {
if (!ua->jcr->wstorage) {
if (jcr->rstorage) {
copy_wstorage(ua->jcr, jcr->rstorage, _("Job resource"));
} else {
copy_wstorage(ua->jcr, jcr->wstorage, _("Job resource"));
}
} else {
USTORE store;
if (jcr->rstorage) {
store.store = jcr->rstore;
} else {
store.store = jcr->wstore;
}
set_wstorage(ua->jcr, &store);
}
if (!connect_to_storage_daemon(ua->jcr, 10, SDConnectTimeout, 1)) {
ua->error_msg(_("Failed to connect to Storage daemon.\n"));
return false;
}
Dmsg0(200, "Connected to storage daemon\n");
sd = ua->jcr->store_bsock;
sd->fsend("cancel Job=%s\n", jcr->Job);
while (sd->recv() >= 0) {
ua->send_msg("%s", sd->msg);
}
sd->signal(BNET_TERMINATE);
sd->close();
ua->jcr->store_bsock = NULL;
jcr->store_bsock->set_timed_out();
jcr->store_bsock->set_terminated();
sd_msg_thread_send_signal(jcr, TIMEOUT_SIGNAL);
jcr->my_thread_send_signal(TIMEOUT_SIGNAL);
}
break;
}
return true;
}
/*
* Do a verification of the specified files against the Catlaog
*
* Returns: false on failure
* true on success
*/
bool do_verify(JCR *jcr)
{
const char *level;
BSOCK *fd;
int stat;
char ed1[100];
JOB_DBR jr;
JobId_t verify_jobid = 0;
const char *Name;
free_wstorage(jcr); /* we don't write */
memset(&jcr->previous_jr, 0, sizeof(jcr->previous_jr));
/*
* Find JobId of last job that ran. Note, we do this when
* the job actually starts running, not at schedule time,
* so that we find the last job that terminated before
* this job runs rather than before it is scheduled. This
* permits scheduling a Backup and Verify at the same time,
* but with the Verify at a lower priority.
*
* For VERIFY_CATALOG we want the JobId of the last INIT.
* For VERIFY_VOLUME_TO_CATALOG, we want the JobId of the
* last backup Job.
*/
if (jcr->get_JobLevel() == L_VERIFY_CATALOG ||
jcr->get_JobLevel() == L_VERIFY_VOLUME_TO_CATALOG ||
jcr->get_JobLevel() == L_VERIFY_DISK_TO_CATALOG) {
memcpy(&jr, &jcr->jr, sizeof(jr));
if (jcr->verify_job &&
(jcr->get_JobLevel() == L_VERIFY_VOLUME_TO_CATALOG ||
jcr->get_JobLevel() == L_VERIFY_DISK_TO_CATALOG)) {
Name = jcr->verify_job->name();
} else {
Name = NULL;
}
Dmsg1(100, "find last jobid for: %s\n", NPRT(Name));
if (!db_find_last_jobid(jcr, jcr->db, Name, &jr)) {
if (jcr->get_JobLevel() == L_VERIFY_CATALOG) {
Jmsg(jcr, M_FATAL, 0, _(
"Unable to find JobId of previous InitCatalog Job.\n"
"Please run a Verify with Level=InitCatalog before\n"
"running the current Job.\n"));
} else {
Jmsg(jcr, M_FATAL, 0, _(
"Unable to find JobId of previous Job for this client.\n"));
}
return false;
}
verify_jobid = jr.JobId;
Dmsg1(100, "Last full jobid=%d\n", verify_jobid);
}
/*
* Now get the job record for the previous backup that interests
* us. We use the verify_jobid that we found above.
*/
if (jcr->get_JobLevel() == L_VERIFY_CATALOG ||
jcr->get_JobLevel() == L_VERIFY_VOLUME_TO_CATALOG ||
jcr->get_JobLevel() == L_VERIFY_DISK_TO_CATALOG) {
jcr->previous_jr.JobId = verify_jobid;
if (!db_get_job_record(jcr, jcr->db, &jcr->previous_jr)) {
Jmsg(jcr, M_FATAL, 0, _("Could not get job record for previous Job. ERR=%s"),
db_strerror(jcr->db));
return false;
}
if (!(jcr->previous_jr.JobStatus == JS_Terminated ||
jcr->previous_jr.JobStatus == JS_Warnings)) {
Jmsg(jcr, M_FATAL, 0, _("Last Job %d did not terminate normally. JobStatus=%c\n"),
verify_jobid, jcr->previous_jr.JobStatus);
return false;
}
Jmsg(jcr, M_INFO, 0, _("Verifying against JobId=%d Job=%s\n"),
jcr->previous_jr.JobId, jcr->previous_jr.Job);
}
/*
* If we are verifying a Volume, we need the Storage
* daemon, so open a connection, otherwise, just
* create a dummy authorization key (passed to
* File daemon but not used).
*/
if (jcr->get_JobLevel() == L_VERIFY_VOLUME_TO_CATALOG) {
if (!create_restore_bootstrap_file(jcr)) {
return false;
}
} else {
jcr->sd_auth_key = bstrdup("dummy"); /* dummy Storage daemon key */
}
if (jcr->get_JobLevel() == L_VERIFY_DISK_TO_CATALOG && jcr->verify_job) {
jcr->fileset = jcr->verify_job->fileset;
}
Dmsg2(100, "ClientId=%u JobLevel=%c\n", jcr->previous_jr.ClientId, jcr->get_JobLevel());
if (!db_update_job_start_record(jcr, jcr->db, &jcr->jr)) {
Jmsg(jcr, M_FATAL, 0, "%s", db_strerror(jcr->db));
return false;
}
/* Print Job Start message */
Jmsg(jcr, M_INFO, 0, _("Start Verify JobId=%s Level=%s Job=%s\n"),
edit_uint64(jcr->JobId, ed1), level_to_str(jcr->get_JobLevel()), jcr->Job);
if (jcr->get_JobLevel() == L_VERIFY_VOLUME_TO_CATALOG) {
/*
* Start conversation with Storage daemon
*/
set_jcr_job_status(jcr, JS_Blocked);
if (!connect_to_storage_daemon(jcr, 10, SDConnectTimeout, 1)) {
return false;
}
/*
* Now start a job with the Storage daemon
*/
if (!start_storage_daemon_job(jcr, jcr->rstorage, NULL)) {
return false;
}
if (!jcr->store_bsock->fsend("run")) {
return false;
}
/*
* Now start a Storage daemon message thread
*/
if (!start_storage_daemon_message_thread(jcr)) {
return false;
}
Dmsg0(50, "Storage daemon connection OK\n");
}
/*
* OK, now connect to the File daemon
* and ask him for the files.
*/
set_jcr_job_status(jcr, JS_Blocked);
if (!connect_to_file_daemon(jcr, 10, FDConnectTimeout, 1)) {
goto bail_out;
}
set_jcr_job_status(jcr, JS_Running);
fd = jcr->file_bsock;
Dmsg0(30, ">filed: Send include list\n");
if (!send_include_list(jcr)) {
goto bail_out;
}
Dmsg0(30, ">filed: Send exclude list\n");
if (!send_exclude_list(jcr)) {
goto bail_out;
}
/*
* Send Level command to File daemon, as well
* as the Storage address if appropriate.
*/
switch (jcr->get_JobLevel()) {
case L_VERIFY_INIT:
level = "init";
break;
case L_VERIFY_CATALOG:
level = "catalog";
break;
case L_VERIFY_VOLUME_TO_CATALOG:
/*
* send Storage daemon address to the File daemon
*/
if (jcr->rstore->SDDport == 0) {
jcr->rstore->SDDport = jcr->rstore->SDport;
}
bnet_fsend(fd, storaddr, jcr->rstore->address, jcr->rstore->SDDport);
if (!response(jcr, fd, OKstore, "Storage", DISPLAY_ERROR)) {
goto bail_out;
}
/*
* Send the bootstrap file -- what Volumes/files to restore
*/
if (!send_bootstrap_file(jcr, fd) ||
!response(jcr, fd, OKbootstrap, "Bootstrap", DISPLAY_ERROR)) {
goto bail_out;
}
if (!jcr->RestoreBootstrap) {
Jmsg0(jcr, M_FATAL, 0, _("Deprecated feature ... use bootstrap.\n"));
goto bail_out;
}
level = "volume";
break;
case L_VERIFY_DATA:
level = "data";
break;
case L_VERIFY_DISK_TO_CATALOG:
level="disk_to_catalog";
break;
default:
Jmsg2(jcr, M_FATAL, 0, _("Unimplemented Verify level %d(%c)\n"), jcr->get_JobLevel(),
jcr->get_JobLevel());
goto bail_out;
}
if (!send_runscripts_commands(jcr)) {
goto bail_out;
}
/*
* Send verify command/level to File daemon
*/
fd->fsend(verifycmd, level);
if (!response(jcr, fd, OKverify, "Verify", DISPLAY_ERROR)) {
goto bail_out;
}
/*
* Now get data back from File daemon and
* compare it to the catalog or store it in the
* catalog depending on the run type.
*/
/* Compare to catalog */
switch (jcr->get_JobLevel()) {
case L_VERIFY_CATALOG:
Dmsg0(10, "Verify level=catalog\n");
jcr->sd_msg_thread_done = true; /* no SD msg thread, so it is done */
jcr->SDJobStatus = JS_Terminated;
get_attributes_and_compare_to_catalog(jcr, jcr->previous_jr.JobId);
break;
case L_VERIFY_VOLUME_TO_CATALOG:
Dmsg0(10, "Verify level=volume\n");
get_attributes_and_compare_to_catalog(jcr, jcr->previous_jr.JobId);
break;
case L_VERIFY_DISK_TO_CATALOG:
Dmsg0(10, "Verify level=disk_to_catalog\n");
jcr->sd_msg_thread_done = true; /* no SD msg thread, so it is done */
jcr->SDJobStatus = JS_Terminated;
get_attributes_and_compare_to_catalog(jcr, jcr->previous_jr.JobId);
break;
case L_VERIFY_INIT:
/* Build catalog */
Dmsg0(10, "Verify level=init\n");
jcr->sd_msg_thread_done = true; /* no SD msg thread, so it is done */
jcr->SDJobStatus = JS_Terminated;
get_attributes_and_put_in_catalog(jcr);
db_end_transaction(jcr, jcr->db); /* terminate any open transaction */
db_write_batch_file_records(jcr);
break;
default:
Jmsg1(jcr, M_FATAL, 0, _("Unimplemented verify level %d\n"), jcr->get_JobLevel());
goto bail_out;
}
stat = wait_for_job_termination(jcr);
verify_cleanup(jcr, stat);
return true;
bail_out:
return false;
}
/*
* Do a verification of the specified files against the Catlaog
*
* Returns: false on failure
* true on success
*/
bool do_verify(JCR *jcr)
{
int JobLevel;
const char *level;
BSOCK *fd = NULL;
BSOCK *sd = NULL;
int status;
char ed1[100];
JOB_DBR jr;
JobId_t verify_jobid = 0;
const char *Name;
free_wstorage(jcr); /* we don't write */
memset(&jcr->previous_jr, 0, sizeof(jcr->previous_jr));
/*
* Find JobId of last job that ran. Note, we do this when
* the job actually starts running, not at schedule time,
* so that we find the last job that terminated before
* this job runs rather than before it is scheduled. This
* permits scheduling a Backup and Verify at the same time,
* but with the Verify at a lower priority.
*
* For VERIFY_CATALOG we want the JobId of the last INIT.
* For VERIFY_VOLUME_TO_CATALOG, we want the JobId of the
* last backup Job.
*/
JobLevel = jcr->getJobLevel();
switch (JobLevel) {
case L_VERIFY_CATALOG:
case L_VERIFY_VOLUME_TO_CATALOG:
case L_VERIFY_DISK_TO_CATALOG:
memcpy(&jr, &jcr->jr, sizeof(jr));
if (jcr->res.verify_job &&
(JobLevel == L_VERIFY_VOLUME_TO_CATALOG ||
JobLevel == L_VERIFY_DISK_TO_CATALOG)) {
Name = jcr->res.verify_job->name();
} else {
Name = NULL;
}
Dmsg1(100, "find last jobid for: %s\n", NPRT(Name));
/*
* See if user supplied a jobid= as run argument or from menu
*/
if (jcr->VerifyJobId) {
verify_jobid = jcr->VerifyJobId;
Dmsg1(100, "Supplied jobid=%d\n", verify_jobid);
} else {
if (!db_find_last_jobid(jcr, jcr->db, Name, &jr)) {
if (JobLevel == L_VERIFY_CATALOG) {
Jmsg(jcr, M_FATAL, 0, _(
"Unable to find JobId of previous InitCatalog Job.\n"
"Please run a Verify with Level=InitCatalog before\n"
"running the current Job.\n"));
} else {
Jmsg(jcr, M_FATAL, 0, _(
"Unable to find JobId of previous Job for this client.\n"));
}
return false;
}
verify_jobid = jr.JobId;
}
Dmsg1(100, "Last full jobid=%d\n", verify_jobid);
/*
* Now get the job record for the previous backup that interests
* us. We use the verify_jobid that we found above.
*/
jcr->previous_jr.JobId = verify_jobid;
if (!db_get_job_record(jcr, jcr->db, &jcr->previous_jr)) {
Jmsg(jcr, M_FATAL, 0, _("Could not get job record for previous Job. ERR=%s"),
db_strerror(jcr->db));
return false;
}
if (!(jcr->previous_jr.JobStatus == JS_Terminated ||
jcr->previous_jr.JobStatus == JS_Warnings)) {
Jmsg(jcr, M_FATAL, 0, _("Last Job %d did not terminate normally. JobStatus=%c\n"),
verify_jobid, jcr->previous_jr.JobStatus);
return false;
}
Jmsg(jcr, M_INFO, 0, _("Verifying against JobId=%d Job=%s\n"),
jcr->previous_jr.JobId, jcr->previous_jr.Job);
}
/*
* If we are verifying a Volume, we need the Storage
* daemon, so open a connection, otherwise, just
* create a dummy authorization key (passed to
* File daemon but not used).
*/
switch (JobLevel) {
case L_VERIFY_VOLUME_TO_CATALOG:
/*
* Note: negative status is an error, zero status, means
* no files were backed up, so skip calling SD and
* client.
*/
status = create_restore_bootstrap_file(jcr);
if (status < 0) { /* error */
return false;
} else if (status == 0) { /* No files, nothing to do */
verify_cleanup(jcr, JS_Terminated); /* clean up */
return true; /* get out */
}
if (jcr->res.verify_job) {
jcr->res.fileset = jcr->res.verify_job->fileset;
}
break;
default:
jcr->sd_auth_key = bstrdup("dummy"); /* dummy Storage daemon key */
break;
}
Dmsg2(100, "ClientId=%u JobLevel=%c\n", jcr->previous_jr.ClientId, JobLevel);
if (!db_update_job_start_record(jcr, jcr->db, &jcr->jr)) {
Jmsg(jcr, M_FATAL, 0, "%s", db_strerror(jcr->db));
return false;
}
/*
* Print Job Start message
*/
Jmsg(jcr, M_INFO, 0, _("Start Verify JobId=%s Level=%s Job=%s\n"),
edit_uint64(jcr->JobId, ed1), level_to_str(JobLevel), jcr->Job);
switch (JobLevel) {
case L_VERIFY_VOLUME_TO_CATALOG:
/*
* Start conversation with Storage daemon
*/
jcr->setJobStatus(JS_Blocked);
if (!connect_to_storage_daemon(jcr, 10, me->SDConnectTimeout, true)) {
return false;
}
sd = jcr->store_bsock;
/*
* Now start a job with the Storage daemon
*/
if (!start_storage_daemon_job(jcr, jcr->res.rstorage, NULL, /* send_bsr */ true)) {
return false;
}
jcr->passive_client = jcr->res.client->passive;
if (!jcr->passive_client) {
/*
* Start the Job in the SD.
*/
if (!sd->fsend("run")) {
return false;
}
/*
* Now start a Storage daemon message thread
*/
if (!start_storage_daemon_message_thread(jcr)) {
return false;
}
Dmsg0(50, "Storage daemon connection OK\n");
}
/*
* OK, now connect to the File daemon and ask him for the files.
*/
jcr->setJobStatus(JS_Blocked);
if (!connect_to_file_daemon(jcr, 10, me->FDConnectTimeout, true)) {
goto bail_out;
}
send_job_info(jcr);
fd = jcr->file_bsock;
/*
* Check if the file daemon supports passive client mode.
*/
if (jcr->passive_client && jcr->FDVersion < FD_VERSION_51) {
Jmsg(jcr, M_FATAL, 0,
_("Client \"%s\" doesn't support passive client mode. "
"Please upgrade your client or disable compat mode.\n"),
jcr->res.client->name());
goto bail_out;
}
break;
default:
/*
* OK, now connect to the File daemon and ask him for the files.
*/
jcr->setJobStatus(JS_Blocked);
if (!connect_to_file_daemon(jcr, 10, me->FDConnectTimeout, true)) {
goto bail_out;
}
send_job_info(jcr);
fd = jcr->file_bsock;
break;
}
jcr->setJobStatus(JS_Running);
Dmsg0(30, ">filed: Send include list\n");
if (!send_include_list(jcr)) {
goto bail_out;
}
Dmsg0(30, ">filed: Send exclude list\n");
if (!send_exclude_list(jcr)) {
goto bail_out;
}
/*
* Send Level command to File daemon, as well as the Storage address if appropriate.
*/
switch (JobLevel) {
case L_VERIFY_INIT:
level = "init";
break;
case L_VERIFY_CATALOG:
level = "catalog";
break;
case L_VERIFY_VOLUME_TO_CATALOG:
if (!jcr->RestoreBootstrap) {
Jmsg0(jcr, M_FATAL, 0, _("Deprecated feature ... use bootstrap.\n"));
goto bail_out;
}
if (!jcr->passive_client) {
int tls_need = BNET_TLS_NONE;
STORERES *store = jcr->res.rstore;
/*
* Send Storage daemon address to the File daemon
*/
if (store->SDDport == 0) {
store->SDDport = store->SDport;
}
/*
* TLS Requirement
*/
if (store->tls.enable) {
if (store->tls.require) {
tls_need = BNET_TLS_REQUIRED;
} else {
tls_need = BNET_TLS_OK;
}
}
fd->fsend(storaddrcmd, store->address, store->SDDport, tls_need, jcr->sd_auth_key);
if (!response(jcr, fd, OKstore, "Storage", DISPLAY_ERROR)) {
goto bail_out;
}
} else {
int tls_need = BNET_TLS_NONE;
CLIENTRES *client = jcr->res.client;
/*
* TLS Requirement
*/
if (client->tls.enable) {
if (client->tls.require) {
tls_need = BNET_TLS_REQUIRED;
} else {
tls_need = BNET_TLS_OK;
}
}
/*
* Tell the SD to connect to the FD.
*/
sd->fsend(passiveclientcmd, client->address, client->FDport, tls_need);
if (!response(jcr, sd, OKpassiveclient, "Passive client", DISPLAY_ERROR)) {
goto bail_out;
}
/*
* Start the Job in the SD.
*/
if (!sd->fsend("run")) {
goto bail_out;
}
/*
* Now start a Storage daemon message thread
*/
if (!start_storage_daemon_message_thread(jcr)) {
goto bail_out;
}
Dmsg0(50, "Storage daemon connection OK\n");
}
level = "volume";
break;
case L_VERIFY_DATA:
level = "data";
break;
case L_VERIFY_DISK_TO_CATALOG:
level="disk_to_catalog";
break;
default:
Jmsg2(jcr, M_FATAL, 0, _("Unimplemented Verify level %d(%c)\n"), JobLevel, JobLevel);
goto bail_out;
}
if (!send_runscripts_commands(jcr)) {
goto bail_out;
}
/*
* Send verify command/level to File daemon
*/
fd->fsend(verifycmd, level);
if (!response(jcr, fd, OKverify, "Verify", DISPLAY_ERROR)) {
goto bail_out;
}
/*
* Now get data back from File daemon and
* compare it to the catalog or store it in the
* catalog depending on the run type.
*/
switch (JobLevel) {
case L_VERIFY_CATALOG:
/*
* Verify from catalog
*/
Dmsg0(10, "Verify level=catalog\n");
jcr->sd_msg_thread_done = true; /* no SD msg thread, so it is done */
jcr->SDJobStatus = JS_Terminated;
get_attributes_and_compare_to_catalog(jcr, jcr->previous_jr.JobId);
break;
case L_VERIFY_VOLUME_TO_CATALOG:
/*
* Verify Volume to catalog entries
*/
Dmsg0(10, "Verify level=volume\n");
get_attributes_and_compare_to_catalog(jcr, jcr->previous_jr.JobId);
break;
case L_VERIFY_DISK_TO_CATALOG:
/*
* Verify Disk attributes to catalog
*/
Dmsg0(10, "Verify level=disk_to_catalog\n");
jcr->sd_msg_thread_done = true; /* no SD msg thread, so it is done */
jcr->SDJobStatus = JS_Terminated;
get_attributes_and_compare_to_catalog(jcr, jcr->previous_jr.JobId);
break;
case L_VERIFY_INIT:
/*
* Build catalog
*/
Dmsg0(10, "Verify level=init\n");
jcr->sd_msg_thread_done = true; /* no SD msg thread, so it is done */
jcr->SDJobStatus = JS_Terminated;
get_attributes_and_put_in_catalog(jcr);
db_end_transaction(jcr, jcr->db); /* terminate any open transaction */
db_write_batch_file_records(jcr);
break;
default:
Jmsg1(jcr, M_FATAL, 0, _("Unimplemented verify level %d\n"), JobLevel);
goto bail_out;
}
status = wait_for_job_termination(jcr);
verify_cleanup(jcr, status);
return true;
bail_out:
if (jcr->file_bsock) {
jcr->file_bsock->signal(BNET_TERMINATE);
jcr->file_bsock->close();
delete jcr->file_bsock;
jcr->file_bsock = NULL;
}
return false;
}
/*
* Do a backup of the specified FileSet
*
* Returns: false on failure
* true on success
*/
bool do_backup(JCR *jcr)
{
int stat;
int tls_need = BNET_TLS_NONE;
BSOCK *fd;
STORE *store;
char ed1[100];
if (jcr->get_JobLevel() == L_VIRTUAL_FULL) {
return do_vbackup(jcr);
}
/* Print Job Start message */
Jmsg(jcr, M_INFO, 0, _("Start Backup JobId %s, Job=%s\n"),
edit_uint64(jcr->JobId, ed1), jcr->Job);
set_jcr_job_status(jcr, JS_Running);
Dmsg2(100, "JobId=%d JobLevel=%c\n", jcr->jr.JobId, jcr->jr.JobLevel);
if (!db_update_job_start_record(jcr, jcr->db, &jcr->jr)) {
Jmsg(jcr, M_FATAL, 0, "%s", db_strerror(jcr->db));
return false;
}
/*
* Open a message channel connection with the Storage
* daemon. This is to let him know that our client
* will be contacting him for a backup session.
*
*/
Dmsg0(110, "Open connection with storage daemon\n");
set_jcr_job_status(jcr, JS_WaitSD);
/*
* Start conversation with Storage daemon
*/
if (!connect_to_storage_daemon(jcr, 10, SDConnectTimeout, 1)) {
return false;
}
/*
* Now start a job with the Storage daemon
*/
if (!start_storage_daemon_job(jcr, NULL, jcr->wstorage)) {
return false;
}
/*
* Start the job prior to starting the message thread below
* to avoid two threads from using the BSOCK structure at
* the same time.
*/
if (!bnet_fsend(jcr->store_bsock, "run")) {
return false;
}
/*
* Now start a Storage daemon message thread. Note,
* this thread is used to provide the catalog services
* for the backup job, including inserting the attributes
* into the catalog. See catalog_update() in catreq.c
*/
if (!start_storage_daemon_message_thread(jcr)) {
return false;
}
Dmsg0(150, "Storage daemon connection OK\n");
set_jcr_job_status(jcr, JS_WaitFD);
if (!connect_to_file_daemon(jcr, 10, FDConnectTimeout, 1)) {
goto bail_out;
}
set_jcr_job_status(jcr, JS_Running);
fd = jcr->file_bsock;
if (!send_include_list(jcr)) {
goto bail_out;
}
if (!send_exclude_list(jcr)) {
goto bail_out;
}
if (!send_level_command(jcr)) {
goto bail_out;
}
/*
* send Storage daemon address to the File daemon
*/
store = jcr->wstore;
if (store->SDDport == 0) {
store->SDDport = store->SDport;
}
/* TLS Requirement */
if (store->tls_enable) {
if (store->tls_require) {
tls_need = BNET_TLS_REQUIRED;
} else {
tls_need = BNET_TLS_OK;
}
}
fd->fsend(storaddr, store->address, store->SDDport, tls_need);
if (!response(jcr, fd, OKstore, "Storage", DISPLAY_ERROR)) {
goto bail_out;
}
if (!send_runscripts_commands(jcr)) {
goto bail_out;
}
/*
* We re-update the job start record so that the start
* time is set after the run before job. This avoids
* that any files created by the run before job will
* be saved twice. They will be backed up in the current
* job, but not in the next one unless they are changed.
* Without this, they will be backed up in this job and
* in the next job run because in that case, their date
* is after the start of this run.
*/
jcr->start_time = time(NULL);
jcr->jr.StartTime = jcr->start_time;
if (!db_update_job_start_record(jcr, jcr->db, &jcr->jr)) {
Jmsg(jcr, M_FATAL, 0, "%s", db_strerror(jcr->db));
}
/*
* If backup is in accurate mode, we send the list of
* all files to FD.
*/
if (!send_accurate_current_files(jcr)) {
goto bail_out;
}
/* Send backup command */
fd->fsend(backupcmd);
if (!response(jcr, fd, OKbackup, "backup", DISPLAY_ERROR)) {
goto bail_out;
}
/* Pickup Job termination data */
stat = wait_for_job_termination(jcr);
db_write_batch_file_records(jcr); /* used by bulk batch file insert */
if (stat == JS_Terminated) {
backup_cleanup(jcr, stat);
return true;
}
return false;
/* Come here only after starting SD thread */
bail_out:
set_jcr_job_status(jcr, JS_ErrorTerminated);
Dmsg1(400, "wait for sd. use=%d\n", jcr->use_count());
/* Cancel SD */
wait_for_job_termination(jcr, FDConnectTimeout);
Dmsg1(400, "after wait for sd. use=%d\n", jcr->use_count());
return false;
}