Re: [Bacula-users] Bacula tapes marked FULL too early - VolBytes too low

[<lst...@kw...>]

Zitat von Dan Langille <da...@la...>:

> On Nov 22, 2012, at 9:40 AM, Jacky Carimalo wrote:
>
>> Le 20/11/2012 16:39, Tilman Schmidt a écrit :
>>> Am 20.11.2012 17:14, schrieb Jacky Carimalo:
>>>> Le 20/11/2012 15:32, John Drescher a écrit :
>>>>> Also you should avoid /dev/st0 and use /dev/nst0 since /dev/st0 can
>>>>> cause corruption in your tapes.
>>>>>
>>>> But I have no choice :
>>>>
>>>> bacula2-64:~/bacula/bin# lsscsi -g
>>>> [0:0:0:0]    disk    IFT      A16F-R2221       347C  /dev/sdb /dev/sg2
>>>> [0:0:1:0]    tape    HP       Ultrium 3-SCSI   G24H  /dev/st0 /dev/sg3
>>>> [0:0:1:1]    tape    HP       Ultrium 3-SCSI   G24H  /dev/st1 /dev/sg4
>>>> [0:0:1:2]    mediumx OVERLAND NEO Series       0604  /dev/sch0 /dev/sg5
>>>> [0:0:1:3]    process Over     NEO VIA FCO      2.22  - /dev/sg6
>>>> ...
>>> I think you have. Just because /dev/nst0 isn't listed by lsscsi
>>> doesn't mean it isn't there. Look:
>>>
>>> [ts@backup ~]$ lsscsi -g
>>> [0:0:0:0]    cd/dvd  HL-DT-ST DVDRAM GSA-H12N  UL01  /dev/sr0   /dev/sg0
>>> [2:0:0:0]    disk    ATA      ST3250620NS      3.AE  /dev/sda   /dev/sg1
>>> [3:0:0:0]    disk    ATA      ST3250620NS      3.AE  /dev/sdb   /dev/sg2
>>> [5:0:5:0]    tape    HP       Ultrium 3-SCSI   G54W  /dev/st0   /dev/sg3
>>> [5:0:5:1]    mediumx HP       1x8 autoloader   1.50  /dev/sch0  /dev/sg4
>>> [ts@backup ~]$ ls -l /dev/*st0*
>>> crw-rw---- 1 root tape 9, 128 19. Okt 16:59 /dev/nst0
>>> crw-rw---- 1 root tape 9, 224 19. Okt 16:59 /dev/nst0a
>>> crw-rw---- 1 root tape 9, 160 19. Okt 16:59 /dev/nst0l
>>> crw-rw---- 1 root tape 9, 192 19. Okt 16:59 /dev/nst0m
>>> crw-rw---- 1 root tape 9,   0 19. Okt 16:59 /dev/st0
>>> crw-rw---- 1 root tape 9,  96 19. Okt 16:59 /dev/st0a
>>> crw-rw---- 1 root tape 9,  32 19. Okt 16:59 /dev/st0l
>>> crw-rw---- 1 root tape 9,  64 19. Okt 16:59 /dev/st0m
>>>
>>>
>>
>> So, I used as you well indicates : /dev/nst0 and /dev/nst1.
>>
>> Autochanger {
>>   Name = OVL_LTO-3
>>   Device = OVL_LTO-3_Drive-1, OVL_LTO-3_Drive-2
>>   Changer Device = /dev/changerovllto3
>>   Changer Command = "/root/bacula/bin/mtx-changer %c %o %S %a %d"
>> }
>>
>> Device {
>>   Name = OVL_LTO-3_Drive-1
>>   Drive Index = 0
>>   Media Type = LTO-3
>>   AutoChanger = yes
>>   Archive Device = /dev/nst1
>>   AutomaticMount = yes;               # when device opened, read it
>>   AlwaysOpen = yes;
>>   LabelMedia = yes;
>>   RandomAccess = yes;
>>   Alert Command = "sh -c 'tapeinfo -f %c |grep TapeAlert|cat'"
>> }
>>
>> Device {
>>   Name = OVL_LTO-3_Drive-2
>>   Drive Index = 1
>>   Media Type = LTO-3
>>   AutoChanger = yes
>>   Archive Device = /dev/nst0
>>   AutomaticMount = yes;               # when device opened, read it
>>   AlwaysOpen = no;
>>    LabelMedia = yes;
>>   RandomAccess = yes;
>>   Alert Command = "sh -c 'tapeinfo -f %c |grep TapeAlert|cat'"
>> }
>>
>> -----------------------------------------------------------------------------------------------
>> But there are still errors :
>>
>> dmesg :
>> Nov 22 12:33:22 singleton.u06.univ-nantes.prive kernel:  
>> [72917.185264] lpfc 0000:08:00.0: 0:(0):0713 SCSI layer issued  
>> Device Reset (1, 1) return x2002
>> Nov 22 12:33:22 singleton.u06.univ-nantes.prive kernel:  
>> [72917.190481] st1: Sense Key : Unit Attention [current]
>> Nov 22 12:33:22 singleton.u06.univ-nantes.prive kernel:  
>> [72917.190487] st1: Add. Sense: Bus device reset function occurred
>> ...
>> -----------------------------------------------------------------------------------------------
>>
>> with errors in bacula job :
>> 2012-11-22 12:33:22 bacula.u11.univ-nantes.prive-
>> Error: block.c:590 Write error at 44:14976 on device  
>> "OVL_LTO-3_Drive-1" (/dev/nst1). ERR=Input/output error.
>> Error: Error writing final EOF to tape. This Volume may not be readable.
>> dev.c:1566 ioctl MTWEOF error on "OVL_LTO-3_Drive-1" (/dev/nst1).  
>> ERR=Input/output error.
>> ...
>> -----------------------------------------------------------------------------------------------
>>
>> I will try cleaning procedures as last solution for me ?
>
> Have we already determined that the SCSI chain is properly terminated?


[0:0:1:3]    process Over     NEO VIA FCO      2.22  - /dev/sg6

Looks like FibreChannel to me...
But for sure the OP should check cabeling and to use the recommended  
Firmware on both ends.

Regards

Andreas