A long time ago I used to be able to do:
echo "scsi add-single-device 0 0 11 0" > /proc/scsi/scsi
echo "scsi remove-single-device 0 0 11 0" > /proc/scsi/scsi
When I wanted to unplug a SCA scsi drive for replacement. I tried this
recently on my 2.4.20 kernel and nothing happened. No errors, no change
to /proc/scsi/scsi, no entry in dmsg, it just ignored it. Has this been
deprecated for a new way of removing hotswap drives?
Thanks,
Robert
:wq!
---------------------------------------------------------------------------
Robert L. Harris | PGP Key ID: E344DA3B
@ x-hkp://pgp.mit.edu
DISCLAIMER:
These are MY OPINIONS ALONE. I speak for no-one else.
Diagnosis: witzelsucht
IPv6 = [email protected] http://ipv6.rdlg.net
IPv4 = [email protected] http://www.rdlg.net
Robert L. Harris wrote:
> A long time ago I used to be able to do:
>
> echo "scsi add-single-device 0 0 11 0" > /proc/scsi/scsi
> echo "scsi remove-single-device 0 0 11 0" > /proc/scsi/scsi
>
> When I wanted to unplug a SCA scsi drive for replacement. I tried this
> recently on my 2.4.20 kernel and nothing happened. No errors, no change
> to /proc/scsi/scsi, no entry in dmsg, it just ignored it. Has this been
> deprecated for a new way of removing hotswap drives?
Robert,
It is not deprecated (and is still present in the lk 2.5
development series since we still have no other way of
doing this from the user space).
The parsing of that expression is very rigid: no tabs
or redundant spaces.
Doug Gilbert
Does it work on scsi-emulation devices such as 3Ware raid controllers?
I'm trying to swap out a failed IDE drive on a 3Ware card. I went into
tw_cli and executed "maint remove c1 p1" and it shows the drive offline.
The disk is in what appears to be a hot-swap connection similar to SCA.
Do I need to do the line below to remove it since the 3ware card
registers the drive offline?
[email protected]:~# tw_cli info c1 p1
Controller 1, Port 1
----------------------
Status: OFFLINE JBOD
Model: Maxtor 4G160J8
Size: 163.92 GB (320173056 blocks)
Serial #: G805DE1E
FW: GAK819K0
Unit: 1
/proc/scsi/scsi does still show the device:
Host: scsi1 Channel: 00 Id: 01 Lun: 00
Vendor: 3ware Model: 3w-xxxx Rev: 1.0
Type: Direct-Access ANSI SCSI revision: ffffffff
and the echo, remove below doesn't remove it. It does happily though
work on some other systems with SCA interfaces.
Thanks,
Robert
Thus spake Douglas Gilbert ([email protected]):
> Robert L. Harris wrote:
> > A long time ago I used to be able to do:
> >
> > echo "scsi add-single-device 0 0 11 0" > /proc/scsi/scsi
> > echo "scsi remove-single-device 0 0 11 0" > /proc/scsi/scsi
> >
> > When I wanted to unplug a SCA scsi drive for replacement. I tried this
> > recently on my 2.4.20 kernel and nothing happened. No errors, no change
> > to /proc/scsi/scsi, no entry in dmsg, it just ignored it. Has this been
> > deprecated for a new way of removing hotswap drives?
>
> Robert,
> It is not deprecated (and is still present in the lk 2.5
> development series since we still have no other way of
> doing this from the user space).
>
> The parsing of that expression is very rigid: no tabs
> or redundant spaces.
>
> Doug Gilbert
Robert -
On Mon, May 12, 2003 at 10:12:55AM -0400, Robert L. Harris wrote:
> /proc/scsi/scsi does still show the device:
>
> Host: scsi1 Channel: 00 Id: 01 Lun: 00
> Vendor: 3ware Model: 3w-xxxx Rev: 1.0
> Type: Direct-Access ANSI SCSI revision: ffffffff
>
> and the echo, remove below doesn't remove it. It does happily though
> work on some other systems with SCA interfaces.
> > > echo "scsi add-single-device 0 0 11 0" > /proc/scsi/scsi
> > > echo "scsi remove-single-device 0 0 11 0" > /proc/scsi/scsi
If this is for the host/channel/id/lun as per your cat /proc/scsi/scsi,
you are specifying id (target) 11, where you should have used 1.
You can also check the result of the write to see if it worked, not sure
if it does anything on 2.4 when the device does not exist, on 2.5 trying to
remove a non-existent device gives me:
[root@elm3b79 root]# echo "scsi remove-single-device 1 2 3 4" > /proc/scsi/scsi
[root@elm3b79 root]# echo $?
1
-- Patrick Mansfield
Just had it hit me again.
# cat /proc/scsi/scsi
Attached devices:
Host: scsi0 Channel: 00 Id: 00 Lun: 00
Vendor: SEAGATE Model: ST318404LC Rev: 0006
Type: Direct-Access ANSI SCSI revision: 03
Host: scsi0 Channel: 00 Id: 01 Lun: 00
Vendor: SEAGATE Model: ST318404LC Rev: 0006
Type: Direct-Access ANSI SCSI revision: 03
Host: scsi0 Channel: 00 Id: 02 Lun: 00
Vendor: SEAGATE Model: ST318404LC Rev: 0006
Type: Direct-Access ANSI SCSI revision: 03
Host: scsi0 Channel: 00 Id: 03 Lun: 00
Vendor: SEAGATE Model: ST318404LC Rev: 0006
Type: Direct-Access ANSI SCSI revision: 03
# echo "scsi remove-single-device 0 0 2 0" > /proc/scsi/scsi
# cat /proc/scsi/scsi
Attached devices:
Host: scsi0 Channel: 00 Id: 00 Lun: 00
Vendor: SEAGATE Model: ST318404LC Rev: 0006
Type: Direct-Access ANSI SCSI revision: 03
Host: scsi0 Channel: 00 Id: 01 Lun: 00
Vendor: SEAGATE Model: ST318404LC Rev: 0006
Type: Direct-Access ANSI SCSI revision: 03
Host: scsi0 Channel: 00 Id: 02 Lun: 00
Vendor: SEAGATE Model: ST318404LC Rev: 0006
Type: Direct-Access ANSI SCSI revision: 03
Host: scsi0 Channel: 00 Id: 03 Lun: 00
Vendor: SEAGATE Model: ST318404LC Rev: 0006
Type: Direct-Access ANSI SCSI revision: 03
Kernel is 2.4.16. Unfortunately I didn't build the kernel so I don't
know what options it was compiled with. Is it possible the previous
admin left out a piece needed to enable this?
Thus spake Douglas Gilbert ([email protected]):
> Robert L. Harris wrote:
> > A long time ago I used to be able to do:
> >
> > echo "scsi add-single-device 0 0 11 0" > /proc/scsi/scsi
> > echo "scsi remove-single-device 0 0 11 0" > /proc/scsi/scsi
> >
> > When I wanted to unplug a SCA scsi drive for replacement. I tried this
> > recently on my 2.4.20 kernel and nothing happened. No errors, no change
> > to /proc/scsi/scsi, no entry in dmsg, it just ignored it. Has this been
> > deprecated for a new way of removing hotswap drives?
>
> Robert,
> It is not deprecated (and is still present in the lk 2.5
> development series since we still have no other way of
> doing this from the user space).
>
> The parsing of that expression is very rigid: no tabs
> or redundant spaces.
>
> Doug Gilbert