On Tuesday 24 May 2005 06:47, [EMAIL PROTECTED] wrote:
> Hi kern,
>
> When i do the fill test, this is what i get back. This is for the single
> tape mode. The 'test' command works fine.
>
> Unusual System Events
> =-=-=-=-=-=-=-=-=-=-=
> May 23 22:00:49 gto kernel: st0: Error with sense data: Info fld=0x0,
> Deferred st09:00: sense key Medium Error
> May 23 22:00:49 gto kernel: Additional sense indicates Excessive write
> errors May 23 22:00:49 gto btape: btape: btape Error: block.c:552 Write
> error at 61:3287 on device /dev/nst0. ERR=Input/output error.
> May 23 22:00:49 gto kernel: st0: Error with sense data: Info fld=0x0,
> Deferred st09:00: sense key Medium Error
> May 23 22:00:49 gto kernel: Additional sense indicates Excessive write
> errors May 23 22:00:49 gto btape: btape: btape Error: Error writing final
> EOF to tape. This tape may not be readable. dev.c:1207 ioctl MTWEOF error
> on /dev/nst0. ERR=Input/output error.
> May 23 22:00:49 gto btape: btape: btape Error: block.c:552 Write error at
> 61:3287 on device /dev/nst0. ERR=Input/output error.
> May 23 22:00:49 gto btape: btape: btape Error: Error writing final EOF to
> tape. This tape may not be readable. dev.c:1207 ioctl MTWEOF error on
> /dev/nst0. ERR=Input/output error.
>
> Where do you think i am going wrong?
Well, this is pretty clearly where the problem is, and I'm surprised that you
didn't see the same kinds of errors in your job output.
I don't know exactly what is going on. You can probably tell if you know the
size of your medium. Either it is writing off the end of the tape because
the driver is faulty, or it is writing off the end of the tape because the
tape is lacking an end of tape marker, or your tape is bad, or your drive is
dirty.
Try changing the medium (different tape). If the same thing happens, try
limiting the size that Bacula will write (I forget the directive). In any
case, you must resolve this problem so that Bacula writes and switches tapes
without error. Once you do, your restores will go much better.
Congratulations for having tested! A lot of people skip this step and get
into trouble when they need their files back.
> Thanks,
> Soumik
>
> Quoting Kern Sibbald <[EMAIL PROTECTED]>:
> > Hello,
> >
> > Thanks. I suspect that your tape drive (or the OS driver) is not properly
> > indicating an end of file before the physical end of the tape is reached.
> > If
> >
> > this is the case, the last block written to the tape will be lost.
> >
> > You can test this by using the btape "fill" command. However, you should
> > run
> >
> > the "test" command first to ensure that all the basic stuff is working.
> > I'd
> >
> > be interested to see the output from the fill command. Since you have an
> > autochanger, it is best to run it using two tapes (i.e. the m option).
> >
> > On Monday 23 May 2005 21:21, Soumik Ukil wrote:
> > > Here are some of the details:
> > >
> > > I am running the director on Debian Linux. The director version is
> > > 1.36.1 (26 November 2004). The client is running SuSe Linux 9.2.
> > > The backup media i am using is Exabyte Mammoth-2 tape storage using
> >
> > Exabyte
> >
> > > 8mm data cartridges. The restore was initialized from bconsole using
> > > restore. Then i marked the files and started the job. As for how i know
> > > that the file is the last in the volume, i am including the relevant
> > > part of the job report. The actual restore was thousands of files so i
> > > am just sending the relevant part:
> > >
> > > max-fd: -rw-r--r-- 1 18477 13101 74257 2005-04-27 11:10:03
> >
> > /raid0/slice1/raid0/slice2/home21/awahle/mozilla-hppa2.0w-hp-hpux11.00-1.
> >7.
> >
> > >tvf max-fd: -rwxr--r-- 1 18477 13101 76715827 2005-04-27 12:10:42
> > > /raid0/slice1/raid0/slice2/home21/awahle/fusion-backup-27apr05.tar
> > > 15-May 11:54 gto-sd: End of Volume at file 70 on device /dev/nst0,
> > > Volume "Full-0001"
> > > 15-May 11:56 gto-sd: RestoreFiles.2005-05-15_07.21.39 Warning: Wrong
> >
> > Volume
> >
> > > mounted on device /dev/nst0: Wanted Inc-0002 have Full-0001
> > > 15-May 11:56 gto-sd: 3301 Issuing autochanger "loaded drive 0" command.
> > > 15-May 11:57 gto-sd: 3302 Autochanger "loaded drive 0", result is Slot
> > > 1. 15-May 11:57 gto-sd: 3303 Issuing autochanger "unload slot 1, drive
> > > 0" command.
> > > 15-May 11:57 gto-sd: 3304 Issuing autochanger "load slot 5, drive 0"
> > > command. 15-May 11:57 gto-sd: 3305 Autochanger "load slot 5, drive 0",
> > > status is OK. 15-May 11:58 gto-sd: Ready to read from volume "Inc-0002"
> > > on device /dev/nst0. 15-May 11:58 gto-sd: Forward spacing to file:block
> > > 13:0. 15-May 11:58 max-fd: RestoreFiles.2005-05-15_07.21.39 Error:
> > > attribs.c:339 File size of restored file
> > > /raid0/slice1/raid0/slice2/home21/awahle/fusion-ba ckup-27apr05.tar not
> > > correct. Original 7671582720, restored 3907616768.
> > >
> > >
> > >
> > > Soumik
> > >
> > > On Monday 23 May 2005 13:57, Kern Sibbald wrote:
> > > > On Sunday 22 May 2005 23:32, Soumik Ukil wrote:
> > > > > I have been trying test restores of a disk spanning Full
> > > > > Incremental and differential backups, and both times it runs ok,
> > > > > almost. Both times, out of the thousands of files , only one file
> > > > > gets restored incorrectly - that is the size it restores is lesser
> > > > > than the original backup.This is the exact error message in the log
> > > > > file:
> > > > >
> > > > > 21-May 20:50 max-fd: RestoreFiles.2005-05-21_14.49.48 Error:
> > > > > attribs.c:339 File size of restored file
> > > > > /raid0/slice1/raid0/slice2/home21/awahle/And reas-06may05/Slides
> > > > > 2005/ECEjan2005/ECE-Seminar-27jan05.ppt not correct. Original
> >
> > 20966400,
> >
> > > > > restored 3047424.
> > > > >
> > > > > Both times the particular file which had a problem getting backed
> > > > > up was at the end of a Full volume, the very last file. I am pretty
> > > > > clueless right now- does any body have an idea?
> > > >
> > > > It is clear that a file that spans a volume is more at risk of
> > > > getting put back incorrectly. However, I have tested this thousands
> > > > of times and never had any problems.
> > > >
> > > > I'd suggest you furnish a bit more information as with what you have
> > > > given, it is not possible to make any assement -- e.g. what OS are
> > > > you running your Director on? What version Director and FD are you
> > > > using? What OS is your FD (client) running on? What backup media are
> > > > you using?
> > > >
> > > > How are you initiating the restore? Please supply a full listing of
> > > > the output (files restored as well as the job report).
> > > >
> > > > How do you know that it is the last file on a Volume?
> >
> > --
> > Best regards,
> >
> > Kern
> >
> > (">
> > /\
> > V_V
--
Best regards,
Kern
(">
/\
V_V
-------------------------------------------------------
This SF.Net email is sponsored by Oracle Space Sweepstakes
Want to be the first software developer in space?
Enter now for the Oracle Space Sweepstakes!
http://ads.osdn.com/?ad_id=7412&alloc_id=16344&op=click
_______________________________________________
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
