From: Mantis B. T. <no...@bu...> - 2012-05-26 09:53:46
|
The following issue has been SUBMITTED. ====================================================================== http://bugs.bacula.org/view.php?id=1875 ====================================================================== Reported By: hip Assigned To: ====================================================================== Project: bacula Issue ID: 1875 Category: Storage Daemon Reproducibility: always Severity: crash Priority: normal Status: new ====================================================================== Date Submitted: 2012-05-26 10:53 BST Last Modified: 2012-05-26 10:53 BST ====================================================================== Summary: Storage Deamon crashes (coredump/traceback) when trying to run a job on Solaris 11 Sparc Description: I was trying to compile Bacula 5.2.6 on a T4-1 Sparc. I've tried with gcc4.5 which comes along with Solaris 11 as well as 3.4.6 (that one I used when I compiled bacula 3.0.3 back on solaris 10). Compiling works fine, there are some minor issues with conio, so I used ./configure --with-mysql --disable-conio When I try to run a job, i get almost immediatelly a failure back 26-May 11:21 node4-dir JobId 14855: Start Backup JobId 14855, Job=Backup_dionysus2.2012-05-26_11.21.53_08 26-May 11:21 node4-dir JobId 14855: Fatal error: <stored: bad response to Job command: No data available 26-May 11:21 node4-dir JobId 14855: Error: Bacula node4-dir 5.2.6 (21Feb12): and after that the sd is no longer there. (Traceback see "Additional Information") I've started sd in Debug200 and could see that it failes in job.c It still comes to line 84: (SOURCE) Dmsg1(100, "<dird: %s", dir->msg); and also has the right content in there: (DEBUG OUTPUT) node4-sd: job.c:84-0 <dird: JobId=14853 job=Backup_dionysus2.2012-05-26_11.13.11_06 job_name=Backup_dionysus2 client_name=dionysus2-fd type=66 level=73 FileSet=FullSetdionysus2 NoAttr=0 SpoolAttr=0 FileSetMD5=aXl2X9ZDY6cQx/QkA9+8mC SpoolData=0 WritePartAfterJob=1 PreferMountedVols=1 SpoolSize=0 rerunning=0 VolSessionId=0 VolSessionTime=0 It also still makes line 85: (SOURCE) bstrncpy(spool_size, "0", sizeof(spool_size)); (I've added some additional debugs to verify) but then when it comes to line 86 it fails: (SOURCE) stat = sscanf(dir->msg, jobcmd, &JobId, job.c_str(), job_name.c_str(), client_name.c_str(), &JobType, &level, fileset_name.c_str(), &no_attributes, &spool_attributes, fileset_md5.c_str(), &spool_data, &write_part_after_job, &PreferMountedVols, spool_size, &jcr->rerunning, &jcr->VolSessionId, &jcr->VolSessionTime);( (DEBUG OUTPUT) Bacula interrupted by signal 10: Bus Error Kaboom! bacula-sd, node4-sd got signal 10 - Bus Error. Attempting traceback. Kaboom! exepath=/sbin/ Calling: /sbin/btraceback /sbin/bacula-sd 194 /var/bacula/working gcore: /var/bacula/working/bacula-sd.194 dumped It looks like the traceback worked ... as this is a standard c-command and some variables are just defined inside the funktion, I expect it to have something to do with the POOL_MEM-type-ish vars. Steps to Reproduce: Take Solaris 11 Sparc download current source configure it using ./configure --with-mysql --disable-conio run a job. Additional Information: Subject: Bacula DBX traceback of bacula-sd on node4 To: ro...@no... Date: Sat, 26 May 2012 11:21:55 +0200 (CEST) mdb: failed to dereference symbol: unknown symbol name C++ symbol demangling enabled ******** RUNNING LWPS/THREADS: 1 3 4 ******** STACK TRACE OF CURRENT LWP: ffbfed60 libc.so.1`__pollsys+4(ffbfee20, 1, 0, 0, 0, 0) ffbfedc0 libc.so.1`pselect+0x1e4(1, ffbfef70, ffbfee20, fed90d38, 0, 0) ffbfee60 libc.so.1`select+0xb0(6, ffbfef70, 0, 0, 0, 20) ffbfeec8 libbac-5.2.6.so`bnet_thread_server+0x304(1b0, ffbffa78, 1, 1bc, ff2f7900, ff312ee8) ffbffb28 main+0x710(6d800, 1, ffbffc38, 6d800, 6e000, 0) ffbffbb8 _start+0x5c(0, 0, 0, 0, 0, 0) ******** VARIABLES DUMP OF CURRENT LWP: ******** STACK TRACE OF LWPS: stack pointer for thread 1: ffbfed60 [ ffbfed60 libc.so.1`__pollsys+4() ] ffbfedc0 libc.so.1`pselect+0x1e4() ffbfee60 libc.so.1`select+0xb0() ffbfeec8 libbac-5.2.6.so`bnet_thread_server+0x304() ffbffb28 main+0x710() ffbffbb8 _start+0x5c() stack pointer for thread 3: fea7bc68 [ fea7bc68 libc.so.1`__lwp_park+0x10() ] fea7bd00 libc.so.1`cond_wait_queue+0x4c() fea7bd60 libc.so.1`cond_wait_common+0x2e0() fea7bdd0 libc.so.1`__cond_timedwait+0x5c() fea7be38 libc.so.1`cond_timedwait+0x14() fea7be98 libc.so.1`pthread_cond_timedwait+0xc() fea7bef8 libbac-5.2.6.so`watchdog_thread+0x260() fea7bfa0 libc.so.1`_lwp_start() stack pointer for thread 4: feb9aff8 [ feb9aff8 libc.so.1`__waitid+8() ] feb9b058 libc.so.1`waitpid+0x64() feb9b138 libbac-5.2.6.so`signal_handler+0x304() feb9b560 libc.so.1`__sighndlr+0xc() feb9b5c0 libc.so.1`call_user_handler+0x370() feb9b648 libc.so.1`sigacthandler+0x58() feb9b9c0 libbac-5.2.6.so`bsscanf+0x528() feb9ba48 job_cmd+0x13c() feb9bc28 handle_connection_request+0x3c4() feb9bf08 libbac-5.2.6.so`workq_server+0x228() feb9bfa0 libc.so.1`_lwp_start() ******** VARIABLES DUMP OF LWPS: ====================================================================== Issue History Date Modified Username Field Change ====================================================================== 2012-05-26 10:53 hip New Issue ====================================================================== |