Is there any scsi list ? (MSL6000 problem)
Javi Barroso
I'm having trouble with a MSL6000 tape cabine.
I would like to ask to help, but I don't known which is the appropiate
list.
So, I'm asking you about the problem, and I'll be appreciate any
pointer.
My problem:
I have a -old (debian sarge)- bacula system with a MSL6000 attached
with fiber channel. This system was working until 2 weeks ago.
The logs are (a lot of times):
Bacula:
28-Oct 01:47 backup-sd: BackupCatalog.2009-10-27_20.00.22 Fatal error:
3992 Bad autochanger "load slot 5, drive 0": ERR=Child died from
signal 15: Termination.
28-Oct 01:47 backup-fd: BackupCatalog.2009-10-27_20.00.22 Fatal error:
job.c:1617 Bad response to Append Data command. Wanted 3000 OK data
, got 3903 Error append data
/var/log/messages:
Oct 23 15:31:04 backup kernel: st0: Error 70000 (sugg. bt 0x0, driver
bt 0x0, host bt 0x7).
Command line:
# mtx -f /dev/autochanger1 status
cannot open SCSI device '/dev/autochanger1' - No such file or
directory
# ls -l /dev/autochanger1
lrwxrwxrwx 1 root root 3 2009-10-30 09:47 /dev/autochanger1 -> sg4
The device MSL6000 seems to be present in the system:
# lsscsi -g
[0:0:0:0] storage HP HSV200 5110 - /dev/
sg0
[0:0:1:0] storage HP HSV200 5110 - /dev/
sg1
[0:0:2:0] mediumx HP MSL6000 Series 0520 /dev/sch0 /dev/
sg4
[0:0:2:1] tape HP Ultrium 3-SCSI G63W /dev/st2 /dev/
sg5
[0:0:2:2] tape HP Ultrium 3-SCSI G63W /dev/st3 /dev/
sg6
[0:0:2:3] storage HP NS E1200-320 571f - /dev/
sg7
[1:0:3:0] tape HP Ultrium 3-SCSI G63W /dev/st0 /dev/
sg2
[1:0:4:0] tape HP Ultrium 3-SCSI G54W /dev/st1 /dev/
sg3
When I try to remove and add, it dissappear:
# echo "scsi remove-single-device 0 0 2 0" > /proc/scsi/scsi
# echo "scsi add-single-device 0 0 2 0" > /proc/scsi/scsi
# dmesg
lpfc 0000:03:01.0: 0:0713 SCSI layer issued LUN reset (2, 0) Data:
x2002 x3 x2
lpfc 0000:03:01.0: 0:0714 SCSI layer issued Bus Reset Data: x2002
scsi 0:0:2:0: scsi: Device offlined - not ready after error recovery
# lsscsi -g
[0:0:0:0] storage HP HSV200 5110 - /dev/
sg0
[0:0:1:0] storage HP HSV200 5110 - /dev/
sg1
[0:0:2:1] tape HP Ultrium 3-SCSI G63W /dev/st2 /dev/
sg5
[0:0:2:2] tape HP Ultrium 3-SCSI G63W /dev/st3 /dev/
sg6
[0:0:2:3] storage HP NS E1200-320 571f - /dev/
sg7
[1:0:3:0] tape HP Ultrium 3-SCSI G63W /dev/st0 /dev/
sg2
[1:0:4:0] tape HP Ultrium 3-SCSI G54W /dev/st1 /dev/
sg3
If then I reboot the system, 2 tapes , NS E1200-320 and MSL6000
dissappear (all connected by fiber channel):
# lsscsi
[0:0:0:0] storage HP HSV200 5110 -
[0:0:1:0] storage HP HSV200 5110 -
[1:0:3:0] tape HP Ultrium 3-SCSI G63W /dev/st0
[1:0:4:0] tape HP Ultrium 3-SCSI G54W /dev/st1
Two tapes are connected by scsi wire.
If I power off - power on MSL6000 linux get the device:
# dmesg
Vendor: HP Model: MSL6000 Series Rev: 0520
Type: Medium Changer ANSI SCSI revision: 02
Vendor: HP Model: Ultrium 3-SCSI Rev: G63W
Type: Sequential-Access ANSI SCSI revision: 03
st 0:0:2:1: Attached scsi tape st2
st2: try direct i/o: yes (alignment 512 B)
Vendor: HP Model: Ultrium 3-SCSI Rev: G63W
Type: Sequential-Access ANSI SCSI revision: 03
st 0:0:2:2: Attached scsi tape st3
st3: try direct i/o: yes (alignment 512 B)
Vendor: HP Model: NS E1200-320 Rev: 571f
Type: RAID ANSI SCSI revision: 04
SCSI Media Changer driver v0.25
ch0: type #1 (mt): 0x0+1 [medium transport]
ch0: type #2 (st): 0x20+58 [storage]
ch0: type #3 (ie): 0x1c0+2 [import/export]
ch0: type #4 (dt): 0x1e0+4 [data transfer]
ch0: dt 0x1e0: READ ELEMENT STATUS failed
ch0: dt 0x1e1: READ ELEMENT STATUS failed
ch0: dt 0x1e2: READ ELEMENT STATUS failed
ch0: dt 0x1e3: READ ELEMENT STATUS failed
ch0: INITIALIZE ELEMENT STATUS, may take some time ...
ch0: ... finished
ch 0:0:2:0: Attached scsi changer ch0
st 1:0:3:0: Attached scsi generic sg0 type 1
st 1:0:4:0: Attached scsi generic sg1 type 1
scsi 0:0:0:0: Attached scsi generic sg2 type 12
scsi 0:0:1:0: Attached scsi generic sg3 type 12
ch 0:0:2:0: Attached scsi generic sg4 type 8
st 0:0:2:1: Attached scsi generic sg5 type 1
st 0:0:2:2: Attached scsi generic sg6 type 1
Then I can remove and add the device and MSL6000 rescan the labels
# echo "scsi remove-single-device 0 0 2 0" > /proc/scsi/scsi
# echo "scsi add-single-device 0 0 2 0" > /proc/scsi/scsi
scsi 0:0:2:3: Attached scsi generic sg7 type 12
Vendor: HP Model: MSL6000 Series Rev: 0520
Type: Medium Changer ANSI SCSI revision: 02
ch0: type #1 (mt): 0x0+1 [medium transport]
ch0: type #2 (st): 0x20+58 [storage]
ch0: type #3 (ie): 0x1c0+2 [import/export]
ch0: type #4 (dt): 0x1e0+4 [data transfer]
ch0: dt 0x1e0: ID 1, LUN 0, name: HP HSV200 5110
ch0: dt 0x1e1: ID 2, LUN 0, name: HP MSL6000 Series 0520
ch0: dt 0x1e2: ID 3, LUN 0, Huh? device not found!
ch0: dt 0x1e3: ID 4, LUN 0, Huh? device not found!
ch0: INITIALIZE ELEMENT STATUS, may take some time ...
ch0: ... finished
ch 0:0:2:0: Attached scsi changer ch0
ch 0:0:2:0: Attached scsi generic sg4 type 8
Now I'm going to program a cron job executing
mtx -f /dev/autochanger1 status
every 5 minutes, so I can know when it fails again :(
Any help is appreciated
Thanks!
PD: Comp.unix.question users, please cc to me