Menu
▾
▴
[Bacula-bugs] [bacula 0000072]: bacula-dir kabooms at run command
|
From: <bac...@li...> - 2004-09-01 19:04:21
|
The following bug has been CLOSED ====================================================================== http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000072 ====================================================================== Reported By: jhreiche Assigned To: ====================================================================== Project: bacula Bug ID: 72 Category: Director Reproducibility: always Severity: crash Priority: normal Status: closed ====================================================================== Date Submitted: 08-06-2004 00:47 PDT Last Modified: 09-01-2004 12:07 PDT ====================================================================== Summary: bacula-dir kabooms at run command Description: I installed Bacula on three systems: bacula-dir Version: 1.34.5 (21 June 2004) i586-pc-linux-gnu suse 9.0 bacula_ts-sd Version: 1.34.5 (21 June 2004) i586-pc-linux-gnu suse 9.0 linux2-fd Version: 1.34.2 (24 Apr 2004) i686-suse-linux-gnu suse 9.0 After setting the fileset and so on, I tried to "run" "linux2ts" wich should backup linux2 to tapes, but the director daemon crashes. You can see the traceback output in additional information. In former times I run bacula on a RedHat 9.0 Linux. Maybe the problem is the result of a different library version. It's maybe a database failure. I'll try to test it today ... ====================================================================== ---------------------------------------------------------------------- kern - 08-06-2004 12:12 PDT ---------------------------------------------------------------------- Thanks for the traceback -- it gets very close to the problem. It looks like something has gone wrong in your Catalog resource. To go further, you will need to run the director under the debugger as described in the Kaboom chapter of the manual, and when it stops, enter the following commands: print *jcr print *jcr->catalog That should help me get a bit farther. ---------------------------------------------------------------------- jhreiche - 08-08-2004 13:13 PDT ---------------------------------------------------------------------- Hi!, I will be on holiday till 18.8.2004. So it will take some time till I can report the output from the debugger. Regards, Jan He. ---------------------------------------------------------------------- jhreiche - 08-26-2004 06:18 PDT ---------------------------------------------------------------------- Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 49156 (LWP 7685)] 0x0805596d in run_job(JCR*) (jcr=0x80e24d8) at job.c:106 106 jcr->db=db_init_database(jcr, jcr->catalog->db_name, jcr->catalog->db_user, Current language: auto; currently c++ ---------------------------------------------------------- (gdb) print *jcr $1 = {next = 0x80db318, prev = 0x0, use_count = 1, my_thread_id = 0, mutex = { __m_reserved = 0, __m_count = 0, __m_owner = 0x0, __m_kind = 0, __m_lock = {__status = 1, __spinlock = 0}}, dir_bsock = 0x0, store_bsock = 0x0, file_bsock = 0x0, daemon_free_jcr = 0x80569f0 <dird_free_jcr(JCR*)>, msg_queue = 0x80dc170, job_end_push = {items = 0x0, num_items = 0, max_items = 0, num_grow = 1, cur_item = 0, own_items = false}, reload_id = 0, dequeuing = false, errmsg = 0x80e1de0 "", Job = "Linux2_ts.2004-08-26_16.14.27", '\0' <repeats 98 times>, JobId = 0, VolSessionId = 0, VolSessionTime = 0, JobFiles = 0, JobErrors = 0, JobBytes = 0, ReadBytes = 0, Errors = 0, JobStatus = 67, JobType = 66, JobLevel = 70, JobPriority = 10, authenticated = 0, sched_time = 1093529662, start_time = 1093529667, run_time = 0, end_time = 0, VolumeName = 0x80e1b80 "", client_name = 0x80e2928 "Linux2", RestoreBootstrap = 0x0, sd_auth_key = 0x0, jcr_msgs = 0x80e2bd8, ClientId = 0, where = 0x0, prefix_links = false, gui = false, cached_pnl = 0, cached_path = 0x0, SD_msg_chan = 0, term_wait = {__c_lock = { __status = 0, __spinlock = 0}, __c_waiting = 0x0, __padding = '\0' <repeats 27 times>, __align = 0}, work_item = 0x0, sd_msg_thread_done = false, ua = 0x0, job = 0x80d6ea8, store = 0x80d80f8, client = 0x80d7e08, pool = 0x80d8a28, full_pool = 0x0, inc_pool = 0x0, dif_pool = 0x0, fileset = 0x80d7748, catalog = 0x0, messages = 0x80d8850, SDJobFiles = 0, SDJobBytes = 0, SDErrors = 0, SDJobStatus = 0, ---Type <return> to continue, or q <return> to quit--- FDJobStatus = 0, ExpectedFiles = 0, db = 0x0, MediaId = 0, PoolId = 0, FileId = 0, FileIndex = 0, fname = 0x0, fn_printed = 0, stime = 0x0, jr = { JobId = 0, Job = "Linux2_ts.2004-08-26_16.14.27", '\0' <repeats 98 times>, Name = "Linux2_ts", '\0' <repeats 118 times>, Type = 66, Level = 70, JobStatus = 67, ClientId = 0, PoolId = 0, FileSetId = 0, SchedTime = 1093529662, StartTime = 1093529667, EndTime = 0, JobTDate = 0, VolSessionId = 0, VolSessionTime = 0, JobFiles = 0, JobErrors = 0, JobMissingFiles = 0, JobBytes = 0, FirstIndex = 0, LastIndex = 0, StartFile = 0, EndFile = 0, StartBlock = 0, EndBlock = 0, cSchedTime = '\0' <repeats 49 times>, cStartTime = '\0' <repeats 49 times>, cEndTime = '\0' <repeats 49 times>, rec_addr = 0}, verify_jr = 0x0, RestoreJobId = 0, client_uname = 0x0, replace = 97, saveMaxConcurrentJobs = 0, acquired_resource_locks = false, NumVols = 0, reschedule_count = 0, spool_data = false} ------------------------------------------------------------------------------ (gdb) print *jcr->catalog Cannot access memory at address 0x0 kind regards, Jan HE. Reiche ---------------------------------------------------------------------- kern - 09-01-2004 12:07 PDT ---------------------------------------------------------------------- Thanks for the traceback. The problem is very clearly that the catalog pointer is NULL, most likely because you forgot to define a catalog for the job you are running. It appears that Bacula does not check for that, and it should, so I will look into it and correct it in 1.35.3. The solution is for you to define a Catalog resource for the job you are trying to run. If you already have or that doesn't fix the job, then I'll need to see your director's conf file and see all the commands you entered before you got the crash. I'm closing the bug, but you can reopen it or submit a new one if my analysis is wrong or you have more info. Bug History Date Modified Username Field Change ====================================================================== 08-06-04 00:47 jhreiche New Bug 08-06-04 12:12 kern Bugnote Added: 0000148 08-06-04 12:12 kern Status new => feedback 08-06-04 12:12 kern version => 1.34.5 08-08-04 13:13 jhreiche Bugnote Added: 0000151 08-26-04 06:18 jhreiche Bugnote Added: 0000187 09-01-04 12:07 kern Bugnote Added: 0000201 09-01-04 12:07 kern Resolution open => Added to TO DO List 09-01-04 12:07 kern Status feedback => closed ====================================================================== |