Hello,

On 15.11.2005 19:27, Kern Sibbald wrote:

Hello,

I have just released Bacula 1.38.1 source code (compressed tar files) to Source Forge. I have also included the 1.38.1 Win32 binaries. This release is primarily a bug fix release to 1.38.0. There are, however, two new "features":

I assume that I've got (mostly) that version running here:
#version
goblin-dir Version: 1.38.1 (11 November 2005)


1. If you have a single drive autochanger, labelling and other operations on the autochanger will not prompt you for the drive number.

They don't prompt, right. Let me show you some output:

From the console:
#update slots=1 scan storage=HPDAT drive=0
Using default Catalog name=BaculaCat DB=bacula
Connecting to Storage daemon HPDAT at goblin:9103 ...
3306 Issuing autochanger "slots" command.
Device "HP DAT" has 6 slots.
Connecting to Storage daemon HPDAT at goblin:9103 ...

This works, as you see.

Now, the same command, only without drive=0 results in a crash. Unfortunately, still no gdb on that machine :-|

goblin-dir: ua_label.c:729 Sent: readlabel HPDAT Slot=1 drive=0
goblin-dir: ua_label.c:734 Got: 3301 Issuing autochanger "loaded drive 0" 
command.
goblin-dir: ua_label.c:734 Got: 3302 Autochanger "loaded drive 0", result: 
nothing loaded.
goblin-dir: ua_label.c:734 Got: 3304 Issuing autochanger "load slot 1, drive 0" 
command.
goblin-dir: ua_label.c:734 Got: 3305 Autochanger "load slot 1, drive 0", status 
is OK.
goblin-dir: ua_label.c:734 Got: 3301 Issuing autochanger "loaded drive 0" 
command.
goblin-dir: ua_label.c:734 Got: 3302 Autochanger "loaded drive 0", result is 
Slot 1.
goblin-dir: ua_label.c:734 Got: 3001 Volume=DAT-120-0050 Slot=1
goblin-dir: ua_label.c:745 get_vol_name=DAT-120-0050
goblin-dir: ua_label.c:210 Got Vol=DAT-120-0050 from SD for Slot=1
goblin-dir: sql_update.c:420 UPDATE Media SET InChanger=0 WHERE Slot=1 AND 
StorageId=3 AND MediaId!=0
goblin-dir: sql_update.c:297 update_media: FirstWritten=1129011902
goblin-dir: sql_update.c:349 UPDATE Media SET 
VolJobs=11,VolFiles=11,VolBlocks=119141,VolBytes=7685645279,VolMounts=4,VolErrors=0,VolWrites=119143,MaxVolBytes=0,VolStatus='Full',Slot=1,InChanger=1,VolReadTime=0,VolWriteTime=0,VolParts=0,LabelType=0,StorageId=3
 WHERE VolumeName='DAT-120-0050'
goblin-dir: sql_update.c:420 UPDATE Media SET InChanger=0 WHERE Slot=1 AND 
StorageId=3 AND MediaId!=257
goblin-dir: ua_label.c:200 Skipping slot=2
goblin-dir: ua_label.c:200 Skipping slot=3
goblin-dir: ua_label.c:200 Skipping slot=4
goblin-dir: ua_label.c:200 Skipping slot=5
Kaboom! bacula-dir, goblin-dir got signal 11. Attempting traceback.
Kaboom! exepath=/root
goblin-dir: signal.c:131 Working=/var/bacula/
goblin-dir: signal.c:132 btpath=/root/btraceback
goblin-dir: signal.c:133 exepath=/root/bacula-dir
Calling: /root/btraceback /root/bacula-dir 1313
execv: /root/btraceback failed: ERR=No such file or directory
Traceback complete, attempting cleanup ...
goblin-dir: pythonlib.c:216 No startup module.
goblin-dir: jcr.c:150 write_last_jobs seek to 188
Pool   Maxsize  Maxused  Inuse
NoPool      256        0      0
NAME        130        2      2
FNAME       256        9      8
MSG         512        6      5
EMSG       1024        2      2

goblin-dir: util.c:535 edit_job_codes: /sbin/bsmtp -h elf -f "(Bacula) %r" -s 
"Bacula: %t %e of %c %n %l" %r
Segmentation fault
goblin:~ #

The above happens with a _loaded_ but _not_mounted_ tape in the drive.
When there is no tape in the drive, the command works.
I haven't tried other combinations of mounted and loaded state and giving or not giving the drive=x parameter.

2. When you "unmount" an autochanger drive, the drive will be unloaded. Previously the tape was left in the drive.

This works, and this feature is a good thing.

The major bug fixes and changes since version 1.38.0 are:
- A number of fixes to wx-windows (Nicolas Boichat)
- Corrected ACL for Solaris (David Duchscher and Attila Fulop).
- Add bacula_mail_summary.sh to examples directory. It makes
  a single email summary of any number of jobs. Submitted
  by Adrew J. Millar.
- Unmount command now unloads autochanger.
- Fix hang in FD (Martin Simmons)
- Fix Win98 stat() problem in FD (Thorsten Engel)
- Fix update slots which did not clear missing tapes.
- Fix autostart install for FreeBSD (user reported)
- Fix several problems with PostgreSQL scripts (Eric Bollinger)
- Critical: allow restore of the first file of non-portable Win32 backup.
- Important: with muliple concurrent jobs the autochanger could get
  confused because of a missing mutex.

Hmm. Is this the on I'm seeing for quite a while now? If it is, it's not yet fixed, at least not with the version I use...

- Fix accessing last slot in label and update slots.
- Modify configure.in to add execute option to sqlite3 catalog
  scripts.
- Create update_xxx_table_8_to_9 scripts for updatedb
- Move the -lcrypt for PostgreSQL after the PostgreSQL libs in
  autoconf/bacula-macros/db.m4 as suggested by user.  Fixes bug #457.
- Remove @STATIC_CONS@ from tray-monitor Makefile as suggested
  by user. Fixes bug #456.
- AMD64 users see compiler bug warning below in the main ReleaseNotes !!!!

There still remain a few problems with the new autochanger code that will work on in the future. As far as I am aware, there are no problems with 1.36.3 functionality.

Right, we shouldn't forget this.

There *have* been a number of reports from users concerning polling and a few "hangs". To the best of my ability, I have looked into them all, and as far as I can tell, nothing is related to version 1.38 -- all the reported problems that I can reproduce also occur in 1.36.3. A number of the problems seem to be related to users not having a tape in the drive or doing "Offline on Unmount = yes" while running under a 2.6 kernel. I have already discussed this issue in a previous email, and hopefully I will be able to improve how Bacula deals with this in the future ...

Thanks to everyone who contributed to 1.38.0 and to this release. I am really pleased to see so much participation. A number of the names of code contributors are mentioned above, but there are a good number of other users who have contributed in many ways -- financial support, email support, bug reporting/fixes, submittted code in the pipeline, ... Thanks to everyone.

This might be called the "release and run" release because I will be out for a few days (another back operation -- groan), but hopefully I will have Internet access.

Good luck. I hope this goes more quickly than your last hospital stay - not only because I want you to work again soon ;-)

Arno

--
IT-Service Lehmann                    [EMAIL PROTECTED]
Arno Lehmann                  http://www.its-lehmann.de


-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.  Get Certified Today
Register for a JBoss Training Course.  Free Certification Exam
for All Training Attendees Through End of 2005. For more info visit:
http://ads.osdn.com/?ad_id=7628&alloc_id=16845&op=click
_______________________________________________
Bacula-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/bacula-users

Reply via email to