Bacula

The following bug has been CLOSED
======================================================================
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000647
======================================================================
Reported By:                aron
Assigned To:                
======================================================================
Project:                    bacula
Bug ID:                     647
Category:                   rpms
Reproducibility:            always
Severity:                   major
Priority:                   normal
Status:                     closed
======================================================================
Date Submitted:             07-29-2006 06:21 PDT
Last Modified:              07-29-2006 07:20 PDT
======================================================================
Summary:                    building rpms from src.rpm
Description: 
I'am trying to build rpms from the src.rpm I downloaded from sf.net:

rpmbuild --rebuild --define "build_fc5 1" --define "build_gconsole 1"
--define "build_postgresql 1" --define "build_x86_64 1" --define
"build_wxconsole 1" bacula-1.38.11-3.src.rpm

and always get this:
Processing files: bacula-debuginfo-1.38.11-3
Requires(rpmlib): rpmlib(CompressedFileNames) <= 3.0.4-1
rpmlib(PayloadFilesHavePrefix) <= 4.0-1


RPM build errors:
    InstallSourcePackage: Header V3 DSA signature: NOKEY, key ID 10a792ad
    user sbarn does not exist - using root
    user sbarn does not exist - using root
    user sbarn does not exist - using root
    user sbarn does not exist - using root
    user sbarn does not exist - using root
    user sbarn does not exist - using root
    user sbarn does not exist - using root
    user sbarn does not exist - using root
    File not found: /var/tmp/bacula-root/etc/bacula/bconsole.conf
    File not found: /var/tmp/bacula-root/usr/sbin/bconsole
    File not found: /var/tmp/bacula-root/etc/bacula/bconsole.conf
    File not found: /var/tmp/bacula-root/usr/sbin/bconsole


======================================================================

----------------------------------------------------------------------
 barninger - 07-29-2006 07:19 PDT 
----------------------------------------------------------------------
There is no --define "build_gconsole 1" switch. Building gnome console is
on by default. To suppress building gnome console --define
"nobuild_gconsole 1".

InstallSourcePackage: Header V3 DSA signature: NOKEY, key ID 10a792ad
User needs to install the bacula gpg key to prevent this warning.

user sbarn does not exist - using root
Normal behavior, not an error.

File not found: /var/tmp/bacula-root/etc/bacula/bconsole.conf
File not found: /var/tmp/bacula-root/usr/sbin/bconsole
This indicates that bconsole had a compile error. Look earlier in the
build and determine the cause of build failure.

----------------------------------------------------------------------
 barninger - 07-29-2006 07:20 PDT 
----------------------------------------------------------------------
See my bugnote.

Bug History
Date Modified  Username       Field                    Change              
======================================================================
07-29-06 06:21 aron           New Bug                                      
07-29-06 06:22 aron           Bug Monitored: aron                          
07-29-06 07:19 barninger      Bugnote Added: 0001805                       
07-29-06 07:20 barninger      Bugnote Added: 0001806                       
07-29-06 07:20 barninger      Status                   new => closed       
======================================================================

A BUGNOTE has been added to this bug.
======================================================================
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000647
======================================================================
Reported By:                aron
Assigned To:                
======================================================================
Project:                    bacula
Bug ID:                     647
Category:                   rpms
Reproducibility:            always
Severity:                   major
Priority:                   normal
Status:                     new
======================================================================
Date Submitted:             07-29-2006 06:21 PDT
Last Modified:              07-29-2006 07:19 PDT
======================================================================
Summary:                    building rpms from src.rpm
Description: 
I'am trying to build rpms from the src.rpm I downloaded from sf.net:

rpmbuild --rebuild --define "build_fc5 1" --define "build_gconsole 1"
--define "build_postgresql 1" --define "build_x86_64 1" --define
"build_wxconsole 1" bacula-1.38.11-3.src.rpm

and always get this:
Processing files: bacula-debuginfo-1.38.11-3
Requires(rpmlib): rpmlib(CompressedFileNames) <= 3.0.4-1
rpmlib(PayloadFilesHavePrefix) <= 4.0-1


RPM build errors:
    InstallSourcePackage: Header V3 DSA signature: NOKEY, key ID 10a792ad
    user sbarn does not exist - using root
    user sbarn does not exist - using root
    user sbarn does not exist - using root
    user sbarn does not exist - using root
    user sbarn does not exist - using root
    user sbarn does not exist - using root
    user sbarn does not exist - using root
    user sbarn does not exist - using root
    File not found: /var/tmp/bacula-root/etc/bacula/bconsole.conf
    File not found: /var/tmp/bacula-root/usr/sbin/bconsole
    File not found: /var/tmp/bacula-root/etc/bacula/bconsole.conf
    File not found: /var/tmp/bacula-root/usr/sbin/bconsole


======================================================================

----------------------------------------------------------------------
 barninger - 07-29-2006 07:19 PDT 
----------------------------------------------------------------------
There is no --define "build_gconsole 1" switch. Building gnome console is
on by default. To suppress building gnome console --define
"nobuild_gconsole 1".

InstallSourcePackage: Header V3 DSA signature: NOKEY, key ID 10a792ad
User needs to install the bacula gpg key to prevent this warning.

user sbarn does not exist - using root
Normal behavior, not an error.

File not found: /var/tmp/bacula-root/etc/bacula/bconsole.conf
File not found: /var/tmp/bacula-root/usr/sbin/bconsole
This indicates that bconsole had a compile error. Look earlier in the
build and determine the cause of build failure.

Bug History
Date Modified  Username       Field                    Change              
======================================================================
07-29-06 06:21 aron           New Bug                                      
07-29-06 06:22 aron           Bug Monitored: aron                          
07-29-06 07:19 barninger      Bugnote Added: 0001805                       
======================================================================

A BUGNOTE has been added to this bug.
======================================================================
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000645
======================================================================
Reported By:                quelrod
Assigned To:                
======================================================================
Project:                    bacula
Bug ID:                     645
Category:                   Director
Reproducibility:            random
Severity:                   crash
Priority:                   normal
Status:                     new
======================================================================
Date Submitted:             07-19-2006 07:40 PDT
Last Modified:              07-19-2006 07:43 PDT
======================================================================
Summary:                    bacula crashes some nights when backups kick off
Description: 
Some nights backups are fine, other nights they aren't.  I ran bacula-dir
in gdb and waited for a crash.  The first night it was running in gdb it
didn't crash.  However, last night it did.  bconsole is hung and it is
sort of still running but it can't do anything and you can't connect to
it.
======================================================================

----------------------------------------------------------------------
 quelrod - 07-19-2006 07:43 PDT 
----------------------------------------------------------------------
I forgot to mention some other strange behavior related to this.  After it
crashes doing a status dir in bconsole doesn't show the last successful
backups.  Before the crash they showed up in the status dir output.

Also, I failed to mention the log shows absolutely nothing.

Bug History
Date Modified  Username       Field                    Change              
======================================================================
07-19-06 07:40 quelrod        New Bug                                      
07-19-06 07:40 quelrod        Bug Monitored: quelrod                       
07-19-06 07:43 quelrod        Bugnote Added: 0001791                       
======================================================================

A BUGNOTE has been added to this bug.
======================================================================
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000645
======================================================================
Reported By:                quelrod
Assigned To:                
======================================================================
Project:                    bacula
Bug ID:                     645
Category:                   Director
Reproducibility:            random
Severity:                   crash
Priority:                   normal
Status:                     new
======================================================================
Date Submitted:             07-19-2006 07:40 PDT
Last Modified:              07-21-2006 07:09 PDT
======================================================================
Summary:                    bacula crashes some nights when backups kick off
Description: 
Some nights backups are fine, other nights they aren't.  I ran bacula-dir
in gdb and waited for a crash.  The first night it was running in gdb it
didn't crash.  However, last night it did.  bconsole is hung and it is
sort of still running but it can't do anything and you can't connect to
it.
======================================================================

----------------------------------------------------------------------
 quelrod - 07-19-2006 07:43 PDT 
----------------------------------------------------------------------
I forgot to mention some other strange behavior related to this.  After it
crashes doing a status dir in bconsole doesn't show the last successful
backups.  Before the crash they showed up in the status dir output.

Also, I failed to mention the log shows absolutely nothing.

----------------------------------------------------------------------
 quelrod - 07-21-2006 07:09 PDT 
----------------------------------------------------------------------
I verified the crash is still present in 1.38.11.

bacula-dir ran in the debugger fine the first night, but the second night
crashed.  Latest gdb bt below:

Program received signal SIGPIPE, Broken pipe.
[Switching to Thread 1077783456 (LWP 28820)]
0xffffe410 in __kernel_vsyscall ()
(gdb)

Thread 4 (Thread 1084386224 (LWP 28840)):
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000000  0xffffe410 in __kernel_vsyscall ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000001  0x4015c3fb in __read_nocancel () from /lib/tls/libpthread.so.0
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000002  0x080963d3 in read_nbytes ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000003  0x080977cd in bnet_recv ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000004  0x0807f846 in handle_UA_client_request ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000005  0x080b4b79 in workq_server ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000006  0x40157297 in start_thread () from /lib/tls/libpthread.so.0
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000007  0x4033037e in clone () from /lib/tls/libc.so.6
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000008  0x40a26bb0 in ?? ()

Thread 3 (Thread 1082284976 (LWP 28826)):
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000000  0xffffe410 in __kernel_vsyscall ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000001  0x40159cec in pthread_cond_timedwait@... ()
   from /lib/tls/libpthread.so.0
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000002  0x080b3e3a in watchdog_thread ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000003  0x40157297 in start_thread () from /lib/tls/libpthread.so.0
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000004  0x4033037e in clone () from /lib/tls/libc.so.6
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000005  0x40825bb0 in ?? ()

Thread 2 (Thread 1080183728 (LWP 28825)):
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000000  0xffffe410 in __kernel_vsyscall ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000001  0x40328f61 in ___newselect_nocancel () from /lib/tls/libc.so.6
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000002  0x08099694 in bnet_thread_server ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000003  0x0807f556 in connect_thread ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000004  0x40157297 in start_thread () from /lib/tls/libpthread.so.0
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000005  0x4033037e in clone () from /lib/tls/libc.so.6
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000006  0x40624bb0 in ?? ()

Thread 1 (Thread 1077783456 (LWP 28820)):
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000000  0xffffe410 in __kernel_vsyscall ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000001  0x4015c37b in __write_nocancel () from /lib/tls/libpthread.so.0
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000002  0x4005d96d in vio_write () from /usr/lib/libmysqlclient_r.so.14
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000003  0x4005e0a8 in net_real_write () from /usr/lib/libmysqlclient_r.so.14
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000004  0x4005e2f8 in net_flush () from /usr/lib/libmysqlclient_r.so.14
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000005  0x4005e855 in net_write_command () from
/usr/lib/libmysqlclient_r.so.14
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000006  0x4005c16b in cli_advanced_command () from
/usr/lib/libmysqlclient_r.so.14
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000007  0x400593bd in mysql_send_query () from
/usr/lib/libmysqlclient_r.so.14
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000008  0x40059460 in mysql_real_query () from
/usr/lib/libmysqlclient_r.so.14
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000009  0x40030cf5 in mysql_query () from /usr/lib/libmysqlclient_r.so.14
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000010 0x08088958 in InsertDB ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000011 0x08089da7 in db_create_job_record ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000012 0x0805d506 in run_job ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000013 0x0804d3c2 in main ()

Here is the stranger part and again I have disappearing backups in status
dir.
Before the crash I had backups from 19-Jul and 20-Jul that ran.  After the
crash all of those backups fail to appear in status dir.

Bug History
Date Modified  Username       Field                    Change              
======================================================================
07-19-06 07:40 quelrod        New Bug                                      
07-19-06 07:40 quelrod        Bug Monitored: quelrod                       
07-19-06 07:43 quelrod        Bugnote Added: 0001791                       
07-21-06 07:09 quelrod        Bugnote Added: 0001792                       
======================================================================

The following bug has been CLOSED
======================================================================
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000645
======================================================================
Reported By:                quelrod
Assigned To:                
======================================================================
Project:                    bacula
Bug ID:                     645
Category:                   Director
Reproducibility:            random
Severity:                   crash
Priority:                   normal
Status:                     closed
======================================================================
Date Submitted:             07-19-2006 07:40 PDT
Last Modified:              07-23-2006 01:55 PDT
======================================================================
Summary:                    bacula crashes some nights when backups kick off
Description: 
Some nights backups are fine, other nights they aren't.  I ran bacula-dir
in gdb and waited for a crash.  The first night it was running in gdb it
didn't crash.  However, last night it did.  bconsole is hung and it is
sort of still running but it can't do anything and you can't connect to
it.
======================================================================

----------------------------------------------------------------------
 quelrod - 07-19-2006 07:43 PDT 
----------------------------------------------------------------------
I forgot to mention some other strange behavior related to this.  After it
crashes doing a status dir in bconsole doesn't show the last successful
backups.  Before the crash they showed up in the status dir output.

Also, I failed to mention the log shows absolutely nothing.

----------------------------------------------------------------------
 quelrod - 07-21-2006 07:09 PDT 
----------------------------------------------------------------------
I verified the crash is still present in 1.38.11.

bacula-dir ran in the debugger fine the first night, but the second night
crashed.  Latest gdb bt below:

Program received signal SIGPIPE, Broken pipe.
[Switching to Thread 1077783456 (LWP 28820)]
0xffffe410 in __kernel_vsyscall ()
(gdb)

Thread 4 (Thread 1084386224 (LWP 28840)):
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000000  0xffffe410 in __kernel_vsyscall ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000001  0x4015c3fb in __read_nocancel () from /lib/tls/libpthread.so.0
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000002  0x080963d3 in read_nbytes ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000003  0x080977cd in bnet_recv ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000004  0x0807f846 in handle_UA_client_request ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000005  0x080b4b79 in workq_server ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000006  0x40157297 in start_thread () from /lib/tls/libpthread.so.0
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000007  0x4033037e in clone () from /lib/tls/libc.so.6
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000008  0x40a26bb0 in ?? ()

Thread 3 (Thread 1082284976 (LWP 28826)):
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000000  0xffffe410 in __kernel_vsyscall ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000001  0x40159cec in pthread_cond_timedwait@... ()
   from /lib/tls/libpthread.so.0
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000002  0x080b3e3a in watchdog_thread ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000003  0x40157297 in start_thread () from /lib/tls/libpthread.so.0
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000004  0x4033037e in clone () from /lib/tls/libc.so.6
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000005  0x40825bb0 in ?? ()

Thread 2 (Thread 1080183728 (LWP 28825)):
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000000  0xffffe410 in __kernel_vsyscall ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000001  0x40328f61 in ___newselect_nocancel () from /lib/tls/libc.so.6
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000002  0x08099694 in bnet_thread_server ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000003  0x0807f556 in connect_thread ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000004  0x40157297 in start_thread () from /lib/tls/libpthread.so.0
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000005  0x4033037e in clone () from /lib/tls/libc.so.6
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000006  0x40624bb0 in ?? ()

Thread 1 (Thread 1077783456 (LWP 28820)):
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000000  0xffffe410 in __kernel_vsyscall ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000001  0x4015c37b in __write_nocancel () from /lib/tls/libpthread.so.0
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000002  0x4005d96d in vio_write () from /usr/lib/libmysqlclient_r.so.14
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000003  0x4005e0a8 in net_real_write () from /usr/lib/libmysqlclient_r.so.14
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000004  0x4005e2f8 in net_flush () from /usr/lib/libmysqlclient_r.so.14
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000005  0x4005e855 in net_write_command () from
/usr/lib/libmysqlclient_r.so.14
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000006  0x4005c16b in cli_advanced_command () from
/usr/lib/libmysqlclient_r.so.14
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000007  0x400593bd in mysql_send_query () from
/usr/lib/libmysqlclient_r.so.14
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000008  0x40059460 in mysql_real_query () from
/usr/lib/libmysqlclient_r.so.14
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000009  0x40030cf5 in mysql_query () from /usr/lib/libmysqlclient_r.so.14
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000010 0x08088958 in InsertDB ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000011 0x08089da7 in db_create_job_record ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000012 0x0805d506 in run_job ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000013 0x0804d3c2 in main ()

Here is the stranger part and again I have disappearing backups in status
dir.
Before the crash I had backups from 19-Jul and 20-Jul that ran.  After the
crash all of those backups fail to appear in status dir.

----------------------------------------------------------------------
 kern - 07-23-2006 01:55 PDT 
----------------------------------------------------------------------
I don't believe that this is a bug.  You have not shown any job report
output, and you have shown no evidence that Bacula has "crashed".  What
appears to have happened is a SIGPIPE, which a communications error that
is out of Bacula's control, and does not cause Bacula to crash.

I suggest you submit this as a support request.  See the support page of
the web site:  http://www.bacula.org

Re-open this bug only after getting support and only if you have solid
evidence of a Bacula crash.

Bug History
Date Modified  Username       Field                    Change              
======================================================================
07-19-06 07:40 quelrod        New Bug                                      
07-19-06 07:40 quelrod        Bug Monitored: quelrod                       
07-19-06 07:43 quelrod        Bugnote Added: 0001791                       
07-21-06 07:09 quelrod        Bugnote Added: 0001792                       
07-23-06 01:55 kern           Bugnote Added: 0001793                       
07-23-06 01:55 kern           Resolution               open => not a bug   
07-23-06 01:55 kern           Status                   new => closed       
======================================================================

The following bug has been REOPENED.
======================================================================
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000645
======================================================================
Reported By:                quelrod
Assigned To:                
======================================================================
Project:                    bacula
Bug ID:                     645
Category:                   Director
Reproducibility:            random
Severity:                   crash
Priority:                   normal
Status:                     feedback
======================================================================
Date Submitted:             07-19-2006 07:40 PDT
Last Modified:              07-27-2006 09:39 PDT
======================================================================
Summary:                    bacula crashes some nights when backups kick off
Description: 
Some nights backups are fine, other nights they aren't.  I ran bacula-dir
in gdb and waited for a crash.  The first night it was running in gdb it
didn't crash.  However, last night it did.  bconsole is hung and it is
sort of still running but it can't do anything and you can't connect to
it.
======================================================================

----------------------------------------------------------------------
 quelrod - 07-19-2006 07:43 PDT 
----------------------------------------------------------------------
I forgot to mention some other strange behavior related to this.  After it
crashes doing a status dir in bconsole doesn't show the last successful
backups.  Before the crash they showed up in the status dir output.

Also, I failed to mention the log shows absolutely nothing.

----------------------------------------------------------------------
 quelrod - 07-21-2006 07:09 PDT 
----------------------------------------------------------------------
I verified the crash is still present in 1.38.11.

bacula-dir ran in the debugger fine the first night, but the second night
crashed.  Latest gdb bt below:

Program received signal SIGPIPE, Broken pipe.
[Switching to Thread 1077783456 (LWP 28820)]
0xffffe410 in __kernel_vsyscall ()
(gdb)

Thread 4 (Thread 1084386224 (LWP 28840)):
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000000  0xffffe410 in __kernel_vsyscall ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000001  0x4015c3fb in __read_nocancel () from /lib/tls/libpthread.so.0
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000002  0x080963d3 in read_nbytes ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000003  0x080977cd in bnet_recv ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000004  0x0807f846 in handle_UA_client_request ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000005  0x080b4b79 in workq_server ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000006  0x40157297 in start_thread () from /lib/tls/libpthread.so.0
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000007  0x4033037e in clone () from /lib/tls/libc.so.6
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000008  0x40a26bb0 in ?? ()

Thread 3 (Thread 1082284976 (LWP 28826)):
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000000  0xffffe410 in __kernel_vsyscall ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000001  0x40159cec in pthread_cond_timedwait@... ()
   from /lib/tls/libpthread.so.0
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000002  0x080b3e3a in watchdog_thread ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000003  0x40157297 in start_thread () from /lib/tls/libpthread.so.0
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000004  0x4033037e in clone () from /lib/tls/libc.so.6
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000005  0x40825bb0 in ?? ()

Thread 2 (Thread 1080183728 (LWP 28825)):
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000000  0xffffe410 in __kernel_vsyscall ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000001  0x40328f61 in ___newselect_nocancel () from /lib/tls/libc.so.6
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000002  0x08099694 in bnet_thread_server ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000003  0x0807f556 in connect_thread ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000004  0x40157297 in start_thread () from /lib/tls/libpthread.so.0
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000005  0x4033037e in clone () from /lib/tls/libc.so.6
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000006  0x40624bb0 in ?? ()

Thread 1 (Thread 1077783456 (LWP 28820)):
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000000  0xffffe410 in __kernel_vsyscall ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000001  0x4015c37b in __write_nocancel () from /lib/tls/libpthread.so.0
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000002  0x4005d96d in vio_write () from /usr/lib/libmysqlclient_r.so.14
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000003  0x4005e0a8 in net_real_write () from /usr/lib/libmysqlclient_r.so.14
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000004  0x4005e2f8 in net_flush () from /usr/lib/libmysqlclient_r.so.14
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000005  0x4005e855 in net_write_command () from
/usr/lib/libmysqlclient_r.so.14
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000006  0x4005c16b in cli_advanced_command () from
/usr/lib/libmysqlclient_r.so.14
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000007  0x400593bd in mysql_send_query () from
/usr/lib/libmysqlclient_r.so.14
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000008  0x40059460 in mysql_real_query () from
/usr/lib/libmysqlclient_r.so.14
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000009  0x40030cf5 in mysql_query () from /usr/lib/libmysqlclient_r.so.14
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000010 0x08088958 in InsertDB ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000011 0x08089da7 in db_create_job_record ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000012 0x0805d506 in run_job ()
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000013 0x0804d3c2 in main ()

Here is the stranger part and again I have disappearing backups in status
dir.
Before the crash I had backups from 19-Jul and 20-Jul that ran.  After the
crash all of those backups fail to appear in status dir.

----------------------------------------------------------------------
 kern - 07-23-2006 01:55 PDT 
----------------------------------------------------------------------
I don't believe that this is a bug.  You have not shown any job report
output, and you have shown no evidence that Bacula has "crashed".  What
appears to have happened is a SIGPIPE, which a communications error that
is out of Bacula's control, and does not cause Bacula to crash.

I suggest you submit this as a support request.  See the support page of
the web site:  http://www.bacula.org

Re-open this bug only after getting support and only if you have solid
evidence of a Bacula crash.

----------------------------------------------------------------------
 quelrod - 07-27-2006 09:39 PDT 
----------------------------------------------------------------------
I have tried other support channels: irc, mailing lists, etc.  In general
no one seems interested in the problem.

In any case shouldn't bacula handle the SIGPIPE instead of having the
console hang?

After the problem:
treebeard:/usr/bacula/bin # ./bconsole 
Connecting to Director treebeard:9101

and it sits there.  I cannot ctrl-c the thing and have to kill the pid. 
Of course if it couldn't connect it would immediately indicate that.

Also, here is part of my logfile:
26-Jul 08:22 treebeard-dir: Start Backup JobId 249,
Job=NS2_Backup.2006-07-25_20
.00.08
26-Jul 08:22 treebeard-dir: Bacula 1.38.11 (28Jun06): 26-Jul-2006
08:22:30
  JobId:                  249
  Job:                    NS2_Backup.2006-07-25_20.00.08
  Backup Level:           Incremental, since=2006-07-24 10:02:04
  Client:                 "ns2-fd" i386-pc-linux-gnu,debian,3.1
  FileSet:                "NS2" 2006-07-12 10:21:43
  Pool:                   "Inc-Pool"
  Storage:                "Dell132t"
  Scheduled time:         25-Jul-2006 20:00:07
  Start time:             26-Jul-2006 08:22:12
  End time:               26-Jul-2006 08:22:30
  Elapsed time:           18 secs
  Priority:               16
  FD Files Written:       79
  SD Files Written:       79
  FD Bytes Written:       20,488,692 (20.48 MB)
  SD Bytes Written:       20,496,742 (20.49 MB)
  Rate:                   1138.3 KB/s
  Software Compression:   None
  Volume name(s):         000016
  Volume Session Id:      28
  Volume Session Time:    1153320579
  Last Volume Bytes:      86,255,010,163 (86.25 GB)
  Non-fatal FD errors:    0
  SD Errors:              0
  FD termination status:  OK
  SD termination status:  OK
  Termination:            Backup OK

Notice the date Jul 26?

status dir shows the last backup as:
   199  Full     30,492    327,135,920 OK       12-Jul-06 10:30
NS2_Backup

Whether the SIGPIPE can be fully handled the issue of jobs running but
then not showing up in all status output is certainly problematic.

Here is what else is curious the system was running happily for a good
month without this problem.  I wonder if some underlying library is
causing problems now?  All backups are done locally so communications
errors shouldn't be network driven but would be some sort of software
problem.  

Tell me what to do, try, information to provide, etc. as I want to get
this bug solved and get you everything you need to aide in that.

Also, the problem only seems to mainfest during nightly runs and I haven't
found a way to reproduce it.  Also, sometimes it happens every night, other
times every other night.  It's really tempting to think the systems are
cursed given how erratic the behavior is :P

Bug History
Date Modified  Username       Field                    Change              
======================================================================
07-19-06 07:40 quelrod        New Bug                                      
07-19-06 07:40 quelrod        Bug Monitored: quelrod                       
07-19-06 07:43 quelrod        Bugnote Added: 0001791                       
07-21-06 07:09 quelrod        Bugnote Added: 0001792                       
07-23-06 01:55 kern           Bugnote Added: 0001793                       
07-23-06 01:55 kern           Resolution               open => not a bug   
07-23-06 01:55 kern           Status                   new => closed       
07-27-06 09:39 quelrod        Bugnote Added: 0001804                       
07-27-06 09:39 quelrod        Resolution               not a bug => reopened
07-27-06 09:39 quelrod        Status                   closed => feedback  
======================================================================

A BUGNOTE has been added to this bug.
======================================================================
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000627
======================================================================
Reported By:                hus
Assigned To:                
======================================================================
Project:                    bacula
Bug ID:                     627
Category:                   Director
Reproducibility:            always
Severity:                   major
Priority:                   normal
Status:                     feedback
======================================================================
Date Submitted:             06-08-2006 00:42 PDT
Last Modified:              07-23-2006 02:39 PDT
======================================================================
Summary:                    Bacula requests the same tape repeatedly (or sometimes a different one).
Description: 
When a tape becomes full during a job that spans multiple tapes, bacula
alternatingly requests mounting the (correct) next tape and the full one,
all the while complaining that exactly that tape is full and not
acceptable.
======================================================================

----------------------------------------------------------------------
 kern - 06-08-2006 01:25 PDT 
----------------------------------------------------------------------
Switching from one tape to another is a fundamental feature that happens
all the time, and from the output you present, I cannot see that this
is a bug. Bacula wrote to the end of m01, then asked for m02.  There is no
output that I see that shows Bacula alternating requesting different tapes.
 If for some reason it is doing so, it is probably because you do not have
enough tapes and you are requesting Bacula to prune too fast.

Please discuss this on the bacula-users support email list.  If they
believe it is a bug, then resubmit it, but with clear evidence of a bug.

----------------------------------------------------------------------
 hus - 06-08-2006 05:28 PDT 
----------------------------------------------------------------------
Please read the log again:
03-Jun 08:57 mail-sd: Please mount Volume "m01" on Storage Device "DLT"
This looks like a clear indication that bacula requests "m01" which just
became full. The notification mail for that event says:
03-Jun 08:57 mail-sd: Please mount Volume "m01" on Storage Device "DLT"
(/dev/nst0) for Job taos.2006-06-02_14.45.03

I'm fully aware that bacula can use more than one tape. That was not my
question. I clearly stated that I expect bacula to request another tape.

I've already posted this on the mailing list and received no reply.

----------------------------------------------------------------------
 kern - 06-20-2006 14:29 PDT 
----------------------------------------------------------------------
Well, the output looks very ugly, but I don't see any clear evidence of a
bug since I really don't know what happened.  More important than that, I
don't have any idea how to reproduce this and your steps for reproducing
it don't seem to reproduce anything (unless I am missing something).

Can you reproduce it.  If so, how.  If you have lengthy output, please
attach it as an xxx.txt file otherwise it makes the bug report unreadable.

----------------------------------------------------------------------
 kern - 06-20-2006 14:32 PDT 
----------------------------------------------------------------------
Please upload your bacula-dir.conf file.

----------------------------------------------------------------------
 hus - 07-18-2006 09:24 PDT 
----------------------------------------------------------------------
Should I also upload a database dump from the same time as the logs? If so,
which tables?

Before this test I upgraded to 1.38.10.

----------------------------------------------------------------------
 kern - 07-23-2006 02:39 PDT 
----------------------------------------------------------------------
The fact that you set a poll interval and a "Close on Poll" are probably
causing some problems to the algorithm.  Please remove the "Close on Poll"
directive and see if the problem continues.

A database dump will probably not help resolving the problem as it is not
likely to be related.

Bug History
Date Modified  Username       Field                    Change              
======================================================================
06-08-06 00:42 hus            New Bug                                      
06-08-06 01:25 kern           Bugnote Added: 0001739                       
06-08-06 01:25 kern           Resolution               open => not a bug   
06-08-06 01:25 kern           Status                   new => closed       
06-08-06 05:28 hus            Bugnote Added: 0001740                       
06-08-06 05:28 hus            Resolution               not a bug => reopened
06-08-06 05:28 hus            Status                   closed => feedback  
06-20-06 14:29 kern           Bugnote Added: 0001754                       
06-20-06 14:32 kern           Bugnote Added: 0001755                       
06-20-06 14:32 kern           summary                  bacula requests "full" tape => Bacula requests the same tape repeatedly (or sometimes a different one).
07-18-06 08:49 hus            File Added: bacula-dir.conf                    
07-18-06 08:49 hus            File Added: bacula-sd.conf                    
07-18-06 08:50 hus            File Added: mail-dir.conmsg                    
07-18-06 08:51 hus            File Added: bacula.log                       
07-18-06 08:54 hus            File Added: bacula.trace.bz2                    
07-18-06 09:24 hus            Bugnote Added: 0001790                       
07-23-06 02:39 kern           Bugnote Added: 0001800                       
======================================================================

A BUGNOTE has been added to this bug.
======================================================================
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000627
======================================================================
Reported By:                hus
Assigned To:                
======================================================================
Project:                    bacula
Bug ID:                     627
Category:                   Director
Reproducibility:            always
Severity:                   major
Priority:                   normal
Status:                     feedback
======================================================================
Date Submitted:             06-08-2006 00:42 PDT
Last Modified:              07-25-2006 00:10 PDT
======================================================================
Summary:                    Bacula requests the same tape repeatedly (or sometimes a different one).
Description: 
When a tape becomes full during a job that spans multiple tapes, bacula
alternatingly requests mounting the (correct) next tape and the full one,
all the while complaining that exactly that tape is full and not
acceptable.
======================================================================

----------------------------------------------------------------------
 kern - 06-08-2006 01:25 PDT 
----------------------------------------------------------------------
Switching from one tape to another is a fundamental feature that happens
all the time, and from the output you present, I cannot see that this
is a bug. Bacula wrote to the end of m01, then asked for m02.  There is no
output that I see that shows Bacula alternating requesting different tapes.
 If for some reason it is doing so, it is probably because you do not have
enough tapes and you are requesting Bacula to prune too fast.

Please discuss this on the bacula-users support email list.  If they
believe it is a bug, then resubmit it, but with clear evidence of a bug.

----------------------------------------------------------------------
 hus - 06-08-2006 05:28 PDT 
----------------------------------------------------------------------
Please read the log again:
03-Jun 08:57 mail-sd: Please mount Volume "m01" on Storage Device "DLT"
This looks like a clear indication that bacula requests "m01" which just
became full. The notification mail for that event says:
03-Jun 08:57 mail-sd: Please mount Volume "m01" on Storage Device "DLT"
(/dev/nst0) for Job taos.2006-06-02_14.45.03

I'm fully aware that bacula can use more than one tape. That was not my
question. I clearly stated that I expect bacula to request another tape.

I've already posted this on the mailing list and received no reply.

----------------------------------------------------------------------
 kern - 06-20-2006 14:29 PDT 
----------------------------------------------------------------------
Well, the output looks very ugly, but I don't see any clear evidence of a
bug since I really don't know what happened.  More important than that, I
don't have any idea how to reproduce this and your steps for reproducing
it don't seem to reproduce anything (unless I am missing something).

Can you reproduce it.  If so, how.  If you have lengthy output, please
attach it as an xxx.txt file otherwise it makes the bug report unreadable.

----------------------------------------------------------------------
 kern - 06-20-2006 14:32 PDT 
----------------------------------------------------------------------
Please upload your bacula-dir.conf file.

----------------------------------------------------------------------
 hus - 07-18-2006 09:24 PDT 
----------------------------------------------------------------------
Should I also upload a database dump from the same time as the logs? If so,
which tables?

Before this test I upgraded to 1.38.10.

----------------------------------------------------------------------
 kern - 07-23-2006 02:39 PDT 
----------------------------------------------------------------------
The fact that you set a poll interval and a "Close on Poll" are probably
causing some problems to the algorithm.  Please remove the "Close on Poll"
directive and see if the problem continues.

A database dump will probably not help resolving the problem as it is not
likely to be related.

----------------------------------------------------------------------
 hus - 07-25-2006 00:10 PDT 
----------------------------------------------------------------------
"Close on Poll = no" didn't change anything.

Bug History
Date Modified  Username       Field                    Change              
======================================================================
06-08-06 00:42 hus            New Bug                                      
06-08-06 01:25 kern           Bugnote Added: 0001739                       
06-08-06 01:25 kern           Resolution               open => not a bug   
06-08-06 01:25 kern           Status                   new => closed       
06-08-06 05:28 hus            Bugnote Added: 0001740                       
06-08-06 05:28 hus            Resolution               not a bug => reopened
06-08-06 05:28 hus            Status                   closed => feedback  
06-20-06 14:29 kern           Bugnote Added: 0001754                       
06-20-06 14:32 kern           Bugnote Added: 0001755                       
06-20-06 14:32 kern           summary                  bacula requests "full" tape => Bacula requests the same tape repeatedly (or sometimes a different one).
07-18-06 08:49 hus            File Added: bacula-dir.conf                    
07-18-06 08:49 hus            File Added: bacula-sd.conf                    
07-18-06 08:50 hus            File Added: mail-dir.conmsg                    
07-18-06 08:51 hus            File Added: bacula.log                       
07-18-06 08:54 hus            File Added: bacula.trace.bz2                    
07-18-06 09:24 hus            Bugnote Added: 0001790                       
07-23-06 02:39 kern           Bugnote Added: 0001800                       
07-25-06 00:10 hus            Bugnote Added: 0001803                       
======================================================================

======================================================================
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000640
======================================================================
Reported By:                quelrod
Assigned To:                
======================================================================
Project:                    bacula
Bug ID:                     640
Category:                   Director
Reproducibility:            always
Severity:                   crash
Priority:                   normal
Status:                     feedback
======================================================================
Date Submitted:             07-07-2006 09:06 PDT
Last Modified:              07-23-2006 02:36 PDT
======================================================================
Summary:                    SpoolData directive in Schedule resource causes Dir to segfault
Description: 
SpoolData in the schedule/run directive leads to a segfault when the
backups run.  bacula-dir reloads fine either via bconsole reload or
/etc/init.d/bacula-dir restart.  However, in the morning if you left a
bconsole up it is hung until you restart bacula-dir.  It is not until this
restart that the crash is generated.
======================================================================

----------------------------------------------------------------------
 kern - 07-23-2006 02:34 PDT 
----------------------------------------------------------------------
The conditions under which Bacula crashes are not at all clear to me.

Does it crash any time there is a SpoolData directive in the run command?
Or only after a reload?

----------------------------------------------------------------------
 kern - 07-23-2006 02:36 PDT 
----------------------------------------------------------------------
If this is a crash only after doing a "reload", then it is very likely that
it has been fixed in 1.38.11

Bug History
Date Modified  Username       Field                    Change              
======================================================================
07-07-06 09:06 quelrod        New Bug                                      
07-19-06 07:40 quelrod        Bug Monitored: quelrod                       
07-23-06 02:34 kern           Bugnote Added: 0001798                       
07-23-06 02:34 kern           Status                   new => feedback     
07-23-06 02:36 kern           Bugnote Added: 0001799                       
======================================================================

A BUGNOTE has been added to this bug.
======================================================================
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000640
======================================================================
Reported By:                quelrod
Assigned To:                
======================================================================
Project:                    bacula
Bug ID:                     640
Category:                   Director
Reproducibility:            always
Severity:                   crash
Priority:                   normal
Status:                     feedback
======================================================================
Date Submitted:             07-07-2006 09:06 PDT
Last Modified:              07-24-2006 07:01 PDT
======================================================================
Summary:                    SpoolData directive in Schedule resource causes Dir to segfault
Description: 
SpoolData in the schedule/run directive leads to a segfault when the
backups run.  bacula-dir reloads fine either via bconsole reload or
/etc/init.d/bacula-dir restart.  However, in the morning if you left a
bconsole up it is hung until you restart bacula-dir.  It is not until this
restart that the crash is generated.
======================================================================

----------------------------------------------------------------------
 kern - 07-23-2006 02:34 PDT 
----------------------------------------------------------------------
The conditions under which Bacula crashes are not at all clear to me.

Does it crash any time there is a SpoolData directive in the run command?
Or only after a reload?

----------------------------------------------------------------------
 kern - 07-23-2006 02:36 PDT 
----------------------------------------------------------------------
If this is a crash only after doing a "reload", then it is very likely that
it has been fixed in 1.38.11

----------------------------------------------------------------------
 quelrod - 07-24-2006 07:01 PDT 
----------------------------------------------------------------------
I am now on 1.38.11 but due to bug http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000645 it is nearly impossible for me
to know if this bug is now fixed as well.  Well, rather I already have
semi-unreliable backups I don't want to make them more unreliable until I
see the other problem fixed.

It is somewhat perplexing to me as well.  After a reload to utilize the
new config it is fine.  Backups run by hand are fine.  When scheduled
backups go to run at night access to the director hangs.  Pre-existing
bconsole's are hung and new ones can't even complete the startup routine
to the director.  However, the director port is still open as the system
monitoring software doesn't notice it as down.  Now, at this point in the
hung state it isn't technically crashed, just hung and unusable.  If a
restart is done then the director segfaults.

You may be correct that this is fixed in 1.38.11 as the changelog seems to
discuss my bug but is too light on details for me to be sure.

Bug History
Date Modified  Username       Field                    Change              
======================================================================
07-07-06 09:06 quelrod        New Bug                                      
07-19-06 07:40 quelrod        Bug Monitored: quelrod                       
07-23-06 02:34 kern           Bugnote Added: 0001798                       
07-23-06 02:34 kern           Status                   new => feedback     
07-23-06 02:36 kern           Bugnote Added: 0001799                       
07-24-06 07:01 quelrod        Bugnote Added: 0001802                       
======================================================================

The following bug has been CLOSED
======================================================================
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000643
======================================================================
Reported By:                anicka
Assigned To:                
======================================================================
Project:                    bacula
Bug ID:                     643
Category:                   bconsole
Reproducibility:            always
Severity:                   major
Priority:                   normal
Status:                     closed
======================================================================
Date Submitted:             07-19-2006 05:55 PDT
Last Modified:              07-23-2006 02:07 PDT
======================================================================
Summary:                    bconsole does not compile with glibc 2.4.90
Description: 
Compilation of bconsole with glibc 2.4.90 fails with following error
message:

console.c:78: error: 'bool tee' redeclared as different kind of symbol
/usr/include/bits/fcntl.h:232: error: previous declaration of 'int
tee(int, int, size_t, unsigned int)'

The easiest fix is probably renaming the variable tee in console.c, see
attached patch.
======================================================================

----------------------------------------------------------------------
 kern - 07-23-2006 02:07 PDT 
----------------------------------------------------------------------
This is a bug in glibc, and you should report it. 
/user/include/bits/fcntl.h does not have the right to export a name such
as tee().  You can most likely simply delete the offending line from the
file.  They could possibly break hundreds of programs with this change.

Thanks for the patch. I'll look at applying it, but am not very inclined
to do so since it is a glibc bug that should be fixed.

Bug History
Date Modified  Username       Field                    Change              
======================================================================
07-19-06 05:55 anicka         New Bug                                      
07-19-06 05:55 anicka         File Added: bacula-1.38.11-tee_fcntl.diff                    
07-23-06 02:07 kern           Bugnote Added: 0001796                       
07-23-06 02:07 kern           Resolution               open => not a bug   
07-23-06 02:07 kern           Status                   new => closed       
======================================================================

The following bug has been CLOSED
======================================================================
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000644
======================================================================
Reported By:                mdevogelaere
Assigned To:                
======================================================================
Project:                    bacula
Bug ID:                     644
Category:                   Director
Reproducibility:            always
Severity:                   block
Priority:                   normal
Status:                     closed
======================================================================
Date Submitted:             07-19-2006 06:28 PDT
Last Modified:              07-23-2006 02:02 PDT
======================================================================
Summary:                    Director wants wrong volume after updating slots on a storage daemon when 2 daemons are defined.
Description: 
I need to backup some servers on the local network and on another network.
Therefore i defined 2 storage daemons:
- Tape listening on my local network (10.0.0.60) 
- SSH-Tape listening on 127.0.0.1
The servers on the other network send their data through an SSH-tunnel.
This has been working fine since bacula 1.36.
Recently i modified the setup to use concurrent jobs and now the director
always wants tapes that are not present in the loader: 
- update slots Storage=Tape: director requests the correct tapes for all
jobs on the local network and wants incorrect tapes for all jobs on the
other network.
- update slots Storage=SSH-Tape: director request the correct tapes for
all jobs on the other network and wants incorrect tapes for all jobs on
the local network.
======================================================================

----------------------------------------------------------------------
 kern - 07-23-2006 02:02 PDT 
----------------------------------------------------------------------
Bacula does not support two Storage daemons accessing the same storage
device since version 1.38.0.  This kludge worked in previous versions of
Bacula because Bacula only supported one Autochanger and never checked. 
Now it supports multiple autochangers, and it knows which changer is on
which SD.
The manual is out of date on this subject, and I will fix that.

You can most likely resolve your problem by simply defining a name for
your storage daemon's address (i.e. Address = SD1 rather than Address =
<IP address>).  On your director machine (possibly in /etc/hosts), you can
resolve that name to one local address, and in your Client machine, you can
resolve it to a different address providing they both access the same SD. 
In the Director's Storage definition, you will have only *one* Storage
definition.  In that case, it will probably work correctly.

Bug History
Date Modified  Username       Field                    Change              
======================================================================
07-19-06 06:28 mdevogelaere   New Bug                                      
07-23-06 02:02 kern           Bugnote Added: 0001795                       
07-23-06 02:02 kern           Resolution               open => not a bug   
07-23-06 02:02 kern           Status                   new => closed       
======================================================================

The following bug has been CLOSED
======================================================================
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000646
======================================================================
Reported By:                ifmpeter
Assigned To:                
======================================================================
Project:                    bacula
Bug ID:                     646
Category:                   configure/build process
Reproducibility:            always
Severity:                   minor
Priority:                   normal
Status:                     closed
======================================================================
Date Submitted:             07-20-2006 02:18 PDT
Last Modified:              07-23-2006 01:56 PDT
======================================================================
Summary:                    Build/Configure process ignores "--prefix" for manual pages
Description: 
The makefiles always tries to install the manual
pages into /usr/share/man even if I have specified another 
prefix with "./configure --prefix".


======================================================================

----------------------------------------------------------------------
 kern - 07-23-2006 01:56 PDT 
----------------------------------------------------------------------
Bacula does not use prefix, but has a special configuration option for
setting where the man pages go.  Please read the installation document.

Bug History
Date Modified  Username       Field                    Change              
======================================================================
07-20-06 02:18 ifmpeter       New Bug                                      
07-23-06 01:56 kern           Bugnote Added: 0001794                       
07-23-06 01:56 kern           Resolution               open => not a bug   
07-23-06 01:56 kern           Status                   new => closed       
======================================================================

A BUGNOTE has been added to this bug.
======================================================================
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000642
======================================================================
Reported By:                jrhett
Assigned To:                
======================================================================
Project:                    bacula
Bug ID:                     642
Category:                   Director
Reproducibility:            always
Severity:                   crash
Priority:                   normal
Status:                     new
======================================================================
Date Submitted:             07-13-2006 18:01 PDT
Last Modified:              07-13-2006 18:10 PDT
======================================================================
Summary:                    director crashes when backing up one system
Description: 
13-Jul 17:54 triceratops-fd: triceratops.2006-07-13_17.53.45 Fatal error:
backup.c:500 Network send error to SD. ERR=Broken pipe

And Director kisses off...
======================================================================

----------------------------------------------------------------------
 jrhett - 07-13-2006 18:10 PDT 
----------------------------------------------------------------------
I just noted that this is similar to bug 641.  Interesting parts are:

1. These are both FreeBSD machines.  They are both running the same
version -- in fact, both are built with the same script, and running the
same compiled binary package.

2. There's no firewall between then

3. I have successfully rsynced 2.7TB of data between the machines with no
packet loss

4. No other systems are experiencing problems.  All are likewise FreeBSD
6.0.

Bug History
Date Modified  Username       Field                    Change              
======================================================================
07-13-06 18:01 jrhett         New Bug                                      
07-13-06 18:10 jrhett         Bugnote Added: 0001788                       
======================================================================

The following bug has been CLOSED
======================================================================
http://bugs.bacula.org/bug_view_advanced_page.php?bug_id=0000642
======================================================================
Reported By:                jrhett
Assigned To:                
======================================================================
Project:                    bacula
Bug ID:                     642
Category:                   Director
Reproducibility:            always
Severity:                   crash
Priority:                   normal
Status:                     closed
======================================================================
Date Submitted:             07-13-2006 18:01 PDT
Last Modified:              07-14-2006 02:44 PDT
======================================================================
Summary:                    director crashes when backing up one system
Description: 
13-Jul 17:54 triceratops-fd: triceratops.2006-07-13_17.53.45 Fatal error:
backup.c:500 Network send error to SD. ERR=Broken pipe

And Director kisses off...
======================================================================

----------------------------------------------------------------------
 jrhett - 07-13-2006 18:10 PDT 
----------------------------------------------------------------------
I just noted that this is similar to bug 641.  Interesting parts are:

1. These are both FreeBSD machines.  They are both running the same
version -- in fact, both are built with the same script, and running the
same compiled binary package.

2. There's no firewall between then

3. I have successfully rsynced 2.7TB of data between the machines with no
packet loss

4. No other systems are experiencing problems.  All are likewise FreeBSD
6.0.

----------------------------------------------------------------------
 kern - 07-14-2006 02:44 PDT 
----------------------------------------------------------------------
This is a configuration error.  Please ask for help on the bacula-users
list and send them the information you have provided here.

Bug History
Date Modified  Username       Field                    Change              
======================================================================
07-13-06 18:01 jrhett         New Bug                                      
07-13-06 18:10 jrhett         Bugnote Added: 0001788                       
07-14-06 02:44 kern           Bugnote Added: 0001789                       
07-14-06 02:44 kern           Resolution               open => not a bug   
07-14-06 02:44 kern           Status                   new => closed       
======================================================================