Jon LaBadie wrote:
On Tue, May 16, 2006 at 12:09:29PM +0200, listrcv wrote:
The counting of SCSI devices seems to increase each time I unload and
reload the modules, so it's now scsi4 and was scsi3 before. Is that normal?
I'm not familiar enough with Linux to know the correct commands.
But in some other OSs devices can be looked for as "look for new",
or "start over again and look for everything".
There's some way to make the OS looking for SCSI devices like 'cat
scsi-add-single-device ... > /proc/scsi/scsi' or so. But I didn't want
to use that.
If your system
isn't doing a startoveragain, then some small change to your
tape could cause it to appear as a new device and get a different
name.
As far as I get it, the available SCSI hosts are numbered starting by 0.
In case there are several SCSI hosts (like several SCSI controllers or
multi-channel controllers or SATA devices that are considered as SCSI
besides real SCSI), each host gets another number, counting upwards from 0.
Since only the modules were unloaded and reloaded and the physical hosts
not changed, I would expect it to get the same number it had before. The
total number of hosts was in no way increased.
It looks pretty weird in the syslog:
May 16 11:52:58 prometheus kernel: st: Unloaded.
May 16 11:54:56 prometheus kernel: PCI: Enabling device 0000:02:02.0
(0016 -> 0017)
May 16 11:54:56 prometheus kernel: scsi3 : Adaptec AIC79XX PCI-X SCSI
HBA DRIVER, Rev 1.3.11
May 16 11:54:56 prometheus kernel: <Adaptec 29320A Ultra320 SCSI
adapter>
May 16 11:54:56 prometheus kernel: aic7901: Ultra320 Wide
Channel A, SCSI Id=7, PCI 33 or 66Mhz, 512 SCBs
May 16 11:54:56 prometheus kernel:
May 16 11:56:46 prometheus kernel: (scsi3:A:0): 80.000MB/s transfers
(40.000MHz, 16bit)
May 16 11:56:46 prometheus kernel: (scsi3:A:1): 80.000MB/s transfers
(40.000MHz, 16bit)
May 16 11:56:49 prometheus kernel: Vendor: EXABYTE Model: VXA 1x10
1U Rev: A102
May 16 11:56:49 prometheus kernel: Type: Medium Changer
ANSI SCSI revision: 04
May 16 11:56:49 prometheus kernel: Vendor: EXABYTE Model: VXA-2
Rev: 100E
May 16 11:56:49 prometheus kernel: Type: Sequential-Access
ANSI SCSI revision: 02
May 16 11:57:49 prometheus kernel: Attached scsi generic sg0 at scsi0,
channel 0, id 0, lun 0, type 0
May 16 11:57:49 prometheus kernel: Attached scsi generic sg1 at scsi3,
channel 0, id 0, lun 0, type 8
May 16 11:57:49 prometheus kernel: Attached scsi generic sg2 at scsi3,
channel 0, id 1, lun 0, type 1
May 16 11:57:53 prometheus kernel: st: Version 20040403, fixed bufsize
32768, s/g segs 256
May 16 11:57:53 prometheus kernel: Attached scsi tape st0 at scsi3,
channel 0, id 1, lun 0
May 16 11:57:53 prometheus kernel: st0: try direct i/o: yes (alignment
512 B), max page reachable by HBA 134217727
May 16 11:59:17 prometheus kernel: st: Unloaded.
May 16 11:59:33 prometheus kernel: PCI: Enabling device 0000:02:02.0
(0016 -> 0017)
May 16 11:59:33 prometheus kernel: scsi4 : Adaptec AIC79XX PCI-X SCSI
HBA DRIVER, Rev 1.3.11
May 16 11:59:33 prometheus kernel: <Adaptec 29320A Ultra320 SCSI
adapter>
May 16 11:59:33 prometheus kernel: aic7901: Ultra320 Wide
Channel A, SCSI Id=7, PCI 33 or 66Mhz, 512 SCBs
May 16 11:59:33 prometheus kernel:
May 16 11:59:48 prometheus kernel: (scsi4:A:0): 80.000MB/s transfers
(40.000MHz, 16bit)
May 16 11:59:49 prometheus kernel: (scsi4:A:1): 80.000MB/s transfers
(40.000MHz, 16bit)
May 16 11:59:52 prometheus kernel: Vendor: EXABYTE Model: VXA 1x10
1U Rev: A102
May 16 11:59:52 prometheus kernel: Type: Medium Changer
ANSI SCSI revision: 04
May 16 11:59:52 prometheus kernel: Vendor: EXABYTE Model: VXA-2
Rev: 100E
May 16 11:59:52 prometheus kernel: Type: Sequential-Access
ANSI SCSI revision: 02
May 16 12:00:15 prometheus kernel: Attached scsi generic sg0 at scsi0,
channel 0, id 0, lun 0, type 0
May 16 12:00:15 prometheus kernel: Attached scsi generic sg1 at scsi4,
channel 0, id 0, lun 0, type 8
May 16 12:00:15 prometheus kernel: Attached scsi generic sg2 at scsi4,
channel 0, id 1, lun 0, type 1
At some time, reloading the module will fail because some maximum number
of SCSI hosts will be reached (32?), and I would be forced to reboot.
That's not good ... Is there some way to renumber the scsi hosts or to
reuse them?
Then whenever
the UUID was detected, regardless of the /dev device automatically
assigned, the names you chose were also created.
Fortunately, devices are assigned as they should :) Where would I find
UIDs on Debian amd64?
Hm, maybe I should better post this to the debian-amd64 list?
GH
|